Search Results

Search found 29 results on 2 pages for '10gbe'.

Page 1/2 | 1 2  | Next Page >

  • Bridging 10GbE with 12.04 - bridging works but the bridging computer has no internet access

    - by Donal
    I have been trying to get 12.04 bridging working with two 10GbE cards. I have 2 10GbE cards in a linux box being used only for this bridge, 1 with 2 10GbaseT ports and another with a single CX4 port. I have 2 client computers connected with 10GbaseT cards and the CX4 card connects to a procurve switch. I can get the bridging happening mostly the way that I want, The clients receive dhcp information from the dhcp server (not the bridging machine) and can connect to and properly see the rest of the network. Speeds are ok, not amazing but working on that is another matter. My problem is that the bridging machine has no internet access ... meaning I can't update anything or apt-get anything It can ping all other machines on the local network. I've tried the helpful hints from: https://help.ubuntu.com/community/NetworkConnectionBridge "Enabling Internet Use on the Bridging Computer" and get the following RTNETLINK answers: File exists but dhclient br0 does nothing for me :( I think if it is anything it a multiple route problem as both br0 and eth4 have ipaddresses ... even though I have only set it up so that br0 has one ... Bridge setup details: /etc/network/interface auto br0 iface br0 inet static address 192.168.0.246 netmask 255.255.255.0 gateway 192.168.0.1 broadcast 192.168.0.255 dns-nameservers 192.168.0.1 dns-search example.com dns-domain example.com #(eth2 & eth3 are the 10GbaseT) #(eth4 is the CX4 connection) pre-up ip link set eth2 down pre-up ip link set eth3 down pre-up ip link set eth4 down pre-up brctl addbr br0 pre-up brctl addif br0 eth4 eth3 eth2 pre-up ip addr flush dev eth3 pre-up ip addr flush dev eth2 pre-up ip addr flush dev eth4 post-down ip link set eth4 down post-down ip link set eth2 down post-down ip link set eth3 down post-down ip link set br0 down post-down brctl delif br0 eth2 eth3 eth4 post-down brctl delbr br0 ifconfig -a br0 Link encap:Ethernet HWaddr 00:15:17:22:20:34 inet addr:192.168.0.102 Bcast:192.168.0.255 Mask:255.255.255.0 inet6 addr: fe80::215:17ff:fe22:2034/64 Scope:Link UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:4957 errors:0 dropped:0 overruns:0 frame:0 TX packets:1077 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:0 RX bytes:596320 (596.3 KB) TX bytes:139952 (139.9 KB) eth4 Link encap:Ethernet HWaddr 00:60:dd:47:7c:05 inet addr:192.168.0.57 Bcast:192.168.0.255 Mask:255.255.255.0 inet6 addr: fe80::260:ddff:fe47:7c05/64 Scope:Link UP BROADCAST RUNNING MULTICAST MTU:9000 Metric:1 RX packets:15391 errors:0 dropped:51 overruns:0 frame:0 TX packets:1207 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:5916769 (5.9 MB) TX bytes:154312 (154.3 KB) Interrupt:70 route -n Kernel IP routing table Destination Gateway Genmask Flags Metric Ref Use Iface 0.0.0.0 192.168.0.1 0.0.0.0 UG 0 0 0 eth4 0.0.0.0 192.168.0.1 0.0.0.0 UG 100 0 0 br0 169.254.0.0 0.0.0.0 255.255.0.0 U 1000 0 0 br0 192.168.0.0 0.0.0.0 255.255.255.0 U 0 0 0 br0 192.168.0.0 0.0.0.0 255.255.255.0 U 1 0 0 eth4

    Read the article

  • Very low throughput on 10GbE network

    - by aix
    I have two Linux machines, each equipped with a Solarflare SFN5122F 10GbE NIC. The two NICs are connected together with an SFP+ Direct Attach cable. I am using netperf to measure TCP throughput between the two machines. On one box, I run: netserver and on the other: netperf -t TCP_STREAM -H 192.168.x.x -- -m 32768 I get: MIGRATED TCP STREAM TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to 192.168.x.x (192.168.x.x) port 0 AF_INET Recv Send Send Socket Socket Message Elapsed Size Size Size Time Throughput bytes bytes bytes secs. 10^6bits/sec 87380 16384 32768 10.02 1321.34 The measured throughput is 1.3Gb/s. This is 7.5x below the theoretical maximum, and only 30% faster than 1GbE. What steps can I take to troubleshoot this?

    Read the article

  • 10Gbe sfp+ Cross Over Cable required? Is there such a thing?

    - by dc-patos
    To preface, this is my first experience with 10GBe networking and I have encountered an issue which research does not seem to document a solution for... I have two servers (older DL580G5 and DL380G5), each with a HP NC522SFP 10Gbe dual sfp+ port adapter. I have purchased copper "passive" direct connect adapter cables (which look like twinax), which seem to work well when I connect them to the sfp+ ports on my Dell 5524 switch. However, if I directly connect the two servers with the same cable, the link doesn't come up. I am running WS2012 standard on each server. My intention is to use one of these servers as a home brew SAN and I would like to enable mutiple 10Gbe paths for iSCSI traffic. My question(s): Can I connect the two adapters to each other, such as I would with other less speedy generations of ethernet? If I can, do I require a crossover cable, or some type of other sfp+ cable solution to do this? My 10Gbe sfp+ switch ports are premium, but server to server connections are doable in small numbers for me and I would really like the multiple paths this would give me. Is there a simple solution?

    Read the article

  • Configure spanning tree from HP to Cisco hardware

    - by Tim Brigham
    I have three switches I'd like to configure in a loop - a Cisco stack (3750s) and two HP 2900 series. Each is connected to the next with a 10 gig backplane of one form or another. How do I configure the spanning tree on these systems to make this function correctly? From the documents I've looked at it looks like I need to set both sets of hardware to use MST mode but I'm not sure past that point. The trunking, etc is all set up as needed. HP Switch 1 A4 connected to Cisco 1/0/1. HP Switch 2 B2 connected to Cisco 2/0/1. HP Switch 1 A2 connected to HP Switch 2 A1. HP Switch 1 show spanning-tree Multiple Spanning Tree (MST) Information STP Enabled : Yes Force Version : MSTP-operation IST Mapped VLANs : 1-4094 Switch MAC Address : 0021f7-126580 Switch Priority : 32768 Max Age : 20 Max Hops : 20 Forward Delay : 15 Topology Change Count : 352,485 Time Since Last Change : 2 secs CST Root MAC Address : 0018ba-c74268 CST Root Priority : 1 CST Root Path Cost : 200000 CST Root Port : 1 IST Regional Root MAC Address : 0021f7-126580 IST Regional Root Priority : 32768 IST Regional Root Path Cost : 0 IST Remaining Hops : 20 Root Guard Ports : TCN Guard Ports : BPDU Protected Ports : BPDU Filtered Ports : PVST Protected Ports : PVST Filtered Ports : | Prio | Designated Hello Port Type | Cost rity State | Bridge Time PtP Edge ----- --------- + --------- ---- ---------- + ------------- ---- --- ---- ... A1 | Auto 128 Disabled | A2 10GbE-CX4 | 2000 128 Forwarding | 0021f7-126580 2 Yes No A3 10GbE-CX4 | Auto 128 Disabled | A4 10GbE-SR | 2000 128 Forwarding | 0021f7-126580 2 Yes No HP Switch 2 show spanning-tree Multiple Spanning Tree (MST) Information STP Enabled : Yes Force Version : MSTP-operation IST Mapped VLANs : 1-4094 Switch MAC Address : 0024a8-cd6000 Switch Priority : 32768 Max Age : 20 Max Hops : 20 Forward Delay : 15 Topology Change Count : 19,623 Time Since Last Change : 32 secs CST Root MAC Address : 0018ba-c74268 CST Root Priority : 1 CST Root Path Cost : 202000 CST Root Port : A1 IST Regional Root MAC Address : 0024a8-cd6000 IST Regional Root Priority : 32768 IST Regional Root Path Cost : 0 IST Remaining Hops : 20 Root Guard Ports : TCN Guard Ports : BPDU Protected Ports : BPDU Filtered Ports : PVST Protected Ports : PVST Filtered Ports : | Prio | Designated Hello Port Type | Cost rity State | Bridge Time PtP Edge ----- --------- + --------- ---- ---------- + ------------- ---- --- ---- ... A1 10GbE-CX4 | 2000 128 Forwarding | 0021f7-126580 2 Yes No A2 10GbE-CX4 | Auto 128 Disabled | B1 SFP+SR | 2000 128 Blocking | a44c11-a67c80 2 Yes No B2 | Auto 128 Disabled | Cisco Stack 1 show spanning-tree ... (additional VLANs) VLAN0100 Spanning tree enabled protocol ieee Root ID Priority 1 Address 0018.bac7.426e Cost 2 Port 107 (TenGigabitEthernet2/1/1) Hello Time 2 sec Max Age 20 sec Forward Delay 15 sec Bridge ID Priority 32868 (priority 32768 sys-id-ext 100) Address a44c.11a6.7c80 Hello Time 2 sec Max Age 20 sec Forward Delay 15 sec Aging Time 300 sec Interface Role Sts Cost Prio.Nbr Type ------------------- ---- --- --------- -------- -------------------------------- Te1/1/1 Desg FWD 2 128.53 P2p Te2/1/1 Root FWD 2 128.107 P2p

    Read the article

  • ProCurve network expansion

    - by Blue Warrior NFB
    I've hit a bit of a wall with our network scale-out. As it stands right now: We have five ProCurve 2910al switches connected as above, but with 10GbE connections (two CX4, two fiber). This fully populates the central switch above, there will be no more 10GbE Ethernet connections from that device. This group of switches is not stacked (no stack directive). Sometime in the next two or three months I'll need to add a sixth, and I'm not sure how deep of a hole I'm in. Ideally I'd replace the core switch with something more capable and has more 10GbE ports. However, that's a major outage and that requires special scheduling. The two edge switches connected via fiber have dual-port 10GbE cards in them, so I could physically put another switch on the far end of one of those. I don't know how much of a good or bad idea that would be though. Is that too many segments between end-points? Some config-excerpts: Running configuration: ; J9147A Configuration Editor; Created on release #W.14.49 hostname "REDACTED-SW01" time timezone 120 module 1 type J9147A module 2 type J9008A module 3 type J9149A no stack trunk B1 Trk3 Trunk trunk B2 Trk4 Trunk trunk A1 Trk11 Trunk trunk A2 Trk12 Trunk vlan 15 name "VM-MGMT" untagged Trk2,Trk5,Trk7 ip helper-address 10.1.10.4 ip address 10.1.11.1 255.255.255.0 tagged 37-40,Trk3-Trk4,Trk11-Trk12 jumbo ip proxy-arp exit

    Read the article

  • HP to Cisco spanning tree root flapping

    - by Tim Brigham
    Per a recent question I recently configured both my HP (2x 2900) and Cisco (1x 3750) hardware to use MSTP for interoperability. I thought this was functional until I applied the change to the third device (HP switch 1 below) at which time the spanning tree root started flapping causing performance issues (5% packet loss) between my two HP switches. I'm not sure why. HP Switch 1 A4 connected to Cisco 1/0/1. HP Switch 2 B2 connected to Cisco 2/0/1. HP Switch 1 A2 connected to HP Switch 2 A1. I'd prefer the Cisco stack to act as the root. EDIT: There is one specific line - 'spanning-tree 1 path-cost 500000' in the HP switch 2 that I didn't add and was preexisting. I'm not sure if it could have the kind of impact that I'm describing. I'm more a security and monitoring guy then networking. EDIT 2: I'm starting to believe the problem lies in the fact that the value for my MST 0 instance on the Cisco is still at the default 32768. I worked up a diagram: This is based on every show command I could find for STP. I'll make this change after hours and see if it helps. Cisco 3750 Config: version 12.2 spanning-tree mode mst spanning-tree extend system-id spanning-tree mst configuration name mstp revision 1 instance 1 vlan 1, 40, 70, 100, 250 spanning-tree mst 1 priority 0 vlan internal allocation policy ascending interface TenGigabitEthernet1/1/1 switchport trunk encapsulation dot1q switchport mode trunk ! interface TenGigabitEthernet2/1/1 switchport trunk encapsulation dot1q switchport mode trunk ! interface Vlan1 no ip address ! interface Vlan100 ip address 192.168.100.253 255.255.255.0 ! Cisco 3750 show spanning tree: show spanning-tree MST0 Spanning tree enabled protocol mstp Root ID Priority 32768 Address 0004.ea84.5f80 Cost 200000 Port 53 (TenGigabitEthernet1/1/1) Hello Time 2 sec Max Age 20 sec Forward Delay 15 sec Bridge ID Priority 32768 (priority 32768 sys-id-ext 0) Address a44c.11a6.7c80 Hello Time 2 sec Max Age 20 sec Forward Delay 15 sec Interface Role Sts Cost Prio.Nbr Type ------------------- ---- --- --------- -------- -------------------------------- Te1/1/1 Root FWD 2000 128.53 P2p MST1 Spanning tree enabled protocol mstp Root ID Priority 1 Address a44c.11a6.7c80 This bridge is the root Hello Time 2 sec Max Age 20 sec Forward Delay 15 sec Bridge ID Priority 1 (priority 0 sys-id-ext 1) Address a44c.11a6.7c80 Hello Time 2 sec Max Age 20 sec Forward Delay 15 sec Interface Role Sts Cost Prio.Nbr Type ------------------- ---- --- --------- -------- -------------------------------- Te1/1/1 Desg FWD 2000 128.53 P2p Cisco 3750 show logging: %LINEPROTO-5-UPDOWN: Line protocol on Interface Vlan1, changed state to down %LINEPROTO-5-UPDOWN: Line protocol on Interface Vlan100, changed state to down %LINEPROTO-5-UPDOWN: Line protocol on Interface Vlan1, changed state to up %LINEPROTO-5-UPDOWN: Line protocol on Interface Vlan100, changed state to up %LINEPROTO-5-UPDOWN: Line protocol on Interface Vlan1, changed state to down %LINEPROTO-5-UPDOWN: Line protocol on Interface Vlan1, changed state to up HP Switch 1: ; J9049A Configuration Editor; Created on release #T.13.71 vlan 1 name "DEFAULT_VLAN" untagged 1-8,10,13-16,18-23,A1-A4 ip address 100.100.100.17 255.255.255.0 no untagged 9,11-12,17,24 exit vlan 100 name "192.168.100" untagged 9,11-12,17,24 tagged 1-8,10,13-16,18-23,A1-A4 no ip address exit vlan 21 name "Users_2" tagged 1,A1-A4 no ip address exit vlan 40 name "Cafe" tagged 1,4,7,A1-A4 no ip address exit vlan 250 name "Firewall" tagged 1,4,7,A1-A4 no ip address exit vlan 70 name "DMZ" tagged 1,4,7-8,13,A1-A4 no ip address exit spanning-tree spanning-tree config-name "mstp" spanning-tree config-revision 1 spanning-tree instance 1 vlan 1 40 70 100 250 password manager password operator HP Switch 1 show spanning tree: show spanning-tree Multiple Spanning Tree (MST) Information STP Enabled : Yes Force Version : MSTP-operation IST Mapped VLANs : 2-39,41-69,71-99,101-249,251-4094 Switch MAC Address : 0021f7-126580 Switch Priority : 32768 Max Age : 20 Max Hops : 20 Forward Delay : 15 Topology Change Count : 363,490 Time Since Last Change : 14 hours CST Root MAC Address : 0004ea-845f80 CST Root Priority : 32768 CST Root Path Cost : 200000 CST Root Port : 1 IST Regional Root MAC Address : 0021f7-126580 IST Regional Root Priority : 32768 IST Regional Root Path Cost : 0 IST Remaining Hops : 20 Root Guard Ports : TCN Guard Ports : BPDU Protected Ports : BPDU Filtered Ports : PVST Protected Ports : PVST Filtered Ports : | Prio | Designated Hello Port Type | Cost rity State | Bridge Time PtP Edge ----- --------- + --------- ---- ---------- + ------------- ---- --- ---- A1 | Auto 128 Disabled | A2 10GbE-CX4 | 2000 128 Forwarding | 0021f7-126580 2 Yes No A3 10GbE-CX4 | Auto 128 Disabled | A4 10GbE-SR | Auto 128 Disabled | HP Switch 1 Logging: I removed the date / time fields since they are inaccurate (no NTP configured on these switches) 00839 stp: MSTI 1 Root changed from 0:a44c11-a67c80 to 32768:0021f7-126580 00839 stp: MSTI 1 Root changed from 32768:0021f7-126580 to 0:a44c11-a67c80 00842 stp: MSTI 1 starved for an MSTI Msg Rx on port A4 from 0:a44c11-a67c80 00839 stp: MSTI 1 Root changed from 0:a44c11-a67c80 to 32768:0021f7-126580 00839 stp: MSTI 1 Root changed from 32768:0021f7-126580 to 0:a44c11-a67c80 00839 stp: MSTI 1 Root changed from 0:a44c11-a67c80 to ... HP Switch 2 Configuration: ; J9146A Configuration Editor; Created on release #W.14.49 vlan 1 name "DEFAULT_VLAN" untagged 1,3-17,21-24,A1-A2,B2 ip address 100.100.100.36 255.255.255.0 no untagged 2,18-20,B1 exit vlan 100 name "192.168.100" untagged 2,18-20 tagged 1,3-17,21-24,A1-A2,B1-B2 no ip address exit vlan 21 name "Users_2" tagged 1,A1-A2,B2 no ip address exit vlan 40 name "Cafe" tagged 1,13-14,16,A1-A2,B2 no ip address exit vlan 250 name "Firewall" tagged 1,13-14,16,A1-A2,B2 no ip address exit vlan 70 name "DMZ" tagged 1,13-14,16,A1-A2,B2 no ip address exit logging 192.168.100.18 spanning-tree spanning-tree 1 path-cost 500000 spanning-tree config-name "mstp" spanning-tree config-revision 1 spanning-tree instance 1 vlan 1 40 70 100 250 HP Switch 2 Spanning Tree: show spanning-tree Multiple Spanning Tree (MST) Information STP Enabled : Yes Force Version : MSTP-operation IST Mapped VLANs : 2-39,41-69,71-99,101-249,251-4094 Switch MAC Address : 0024a8-cd6000 Switch Priority : 32768 Max Age : 20 Max Hops : 20 Forward Delay : 15 Topology Change Count : 21,793 Time Since Last Change : 14 hours CST Root MAC Address : 0004ea-845f80 CST Root Priority : 32768 CST Root Path Cost : 200000 CST Root Port : A1 IST Regional Root MAC Address : 0021f7-126580 IST Regional Root Priority : 32768 IST Regional Root Path Cost : 2000 IST Remaining Hops : 19 Root Guard Ports : TCN Guard Ports : BPDU Protected Ports : BPDU Filtered Ports : PVST Protected Ports : PVST Filtered Ports : | Prio | Designated Hello Port Type | Cost rity State | Bridge Time PtP Edge ----- --------- + --------- ---- ---------- + ------------- ---- --- ---- A1 10GbE-CX4 | 2000 128 Forwarding | 0021f7-126580 2 Yes No A2 10GbE-CX4 | Auto 128 Disabled | B1 SFP+SR | 2000 128 Forwarding | 0024a8-cd6000 2 Yes No B2 | Auto 128 Disabled | HP Switch 2 Logging: I removed the date / time fields since they are inaccurate (no NTP configured on these switches) 00839 stp: CST Root changed from 32768:0021f7-126580 to 32768:0004ea-845f80 00839 stp: IST Root changed from 32768:0021f7-126580 to 32768:0024a8-cd6000 00839 stp: CST Root changed from 32768:0004ea-845f80 to 32768:0024a8-cd6000 00839 stp: CST Root changed from 32768:0024a8-cd6000 to 32768:0004ea-845f80 00839 stp: CST Root changed from 32768:0004ea-845f80 to 32768:0024a8-cd6000 00435 ports: port B1 is Blocked by STP 00839 stp: CST Root changed from 32768:0024a8-cd6000 to 32768:0021f7-126580 00839 stp: IST Root changed from 32768:0024a8-cd6000 to 32768:0021f7-126580 00839 stp: CST Root changed from 32768:0021f7-126580 to 32768:0004ea-845f80

    Read the article

  • Sun ZFS Backup Appliance ??!

    - by user13138569
    ???Sun ZFS Backup Appliance ????????? Exadata ? D2D ??????????????? High Capacity ? High Performance ? 2 ????????????????SunRack 1242 ?????????? ??????? ZFS Storage Appliance ???????????????? Web ??????????????Analytics ??????????????? ????????????????????????? High CapacityHigh Performance ?????? ??? 2 (??????) 2 (??????) CPU Intel Xeon E7-4820 (8-core, 2GHz)x4 Intel Xeon E7-4820 (8-core, 2GHz)x4 ??? 256GB 256GB ???????? QDR Infiniband, 10Gbe, 1Gbe(???? 4 ???????) QDR Infiniband, 10Gbe, 1Gbe(???? 4 ???????) ???????? ??????(Raw) 132TB 55TB ????????? 2 (?? 22 ????????) 4 (?? 22 ????????) ??????? 3TB 7200rpm 600GB 15000rpm ??????????? 4 4 ????????????????????????ZFS Backup Appliance ???? Exadata ?????????????????????????????????????????????·???????????????????? OTN ???????????????????????????????????????? Oracle Exadata Backup Configuration Utility v1.0.1 Oracle ZFS Storage Appliance Plugin Downloads US???????????

    Read the article

  • Is Infiniband going to get squeezed by iWARP and external QPI?

    - by andy.grover
    The Inquirer certainly thinks so.However, I'm not so sure it makes sense to compare Infiniband to an as-yet-unannounced optical external QPI. QPI is currently a processor interconnect. CPUs, RAM, and devices connected by it are conceptually part of the same machine -- they run a single OS, for example. They are both "networks" or "fabrics" but they have very different design trade-offs.Another widely-used bus in the system is closer to Infiniband than QPI -- PCI Express. Isn't it more likely that PCIe could take on IB? There are companies already who have solutions that use external PCI Express for cluster interconnect, but these have not gained significant market share. Why would QPI, a technology whose sweet spot is even further from Infiniband's than PCIe, be able to challenge Infiniband? It's hard to speculate without much information, but right now it doesn't seem likely to me.The other prediction made in the article is that Intel's 10GbE iWARP card could squeeze IB on the low end, due to its greater compatibility and lower cost.It's definitely never a good idea to bet against Ethernet when it comes to mass-market, commodity networking. Ethernet will win. 10GbE will win. But, there are now two competing ways to implement the low-latency RDMA Verbs interface on top of Ethernet. iWARP is essentially RDMA over TCP/IP over Ethernet. The new alternative is IBoE (Infiniband over Ethernet, aka RoCEE, aka "Rocky"). This encapsulates the IB packet protocol directly in the Ethernet frame. It loses the layer 3 routability of iWARP, but better maintains software compatibility with existing apps that use IB, and is simpler to implement in both software and hardware. iWARP has a substantial head start, but I believe that IBoE silicon will eventually be cheaper, and more likely to be implemented in commodity Ethernet hardware.I think IBoE is going to take low-end market share from traditional IB, but I think this is a situation IB hardware vendors have no problem accepting. Commoditized IBoE NICs invite greater use of RDMA features, and when higher performance is needed, customers can upgrade to "real" IB, maintaining IB's justification for higher prices. (IB max interconnect speeds have historically been 2-4x higher than Ethernet, and I don't see that changing.)(ObDisclosure: My current employer now sells IB hardware. I previously also worked at Intel. My opinions are my own, duh.)

    Read the article

  • Howo to get Multipath IO with Dell MD3600i into active/active setup?

    - by Disco
    I'm desperately trying to improve performance of my SAN connection. Here's what i have: [root@xnode1 dell]# multipath -ll mpath1 (36d4ae520009bd7cc0000030e4fe8230b) dm-2 DELL,MD36xxi [size=5.5T][features=3 queue_if_no_path pg_init_retries 50][hwhandler=1 rdac][rw] \_ round-robin 0 [prio=200][active] \_ 18:0:0:0 sdb 8:16 [active][ready] \_ 19:0:0:0 sdd 8:48 [active][ghost] \_ 20:0:0:0 sdf 8:80 [active][ghost] \_ 21:0:0:0 sdh 8:112 [active][ready] And multipath.conf : defaults { udev_dir /dev polling_interval 5 prio_callout none rr_min_io 100 max_fds 8192 user_friendly_names yes path_grouping_policy multibus default_features "1 fail_if_no_path" } blacklist { device { vendor "*" product "Universal Xport" } } devices { device { vendor "DELL" product "MD36xxi" path_checker rdac path_selector "round-robin 0" hardware_handler "1 rdac" failback immediate features "2 pg_init_retries 50" no_path_retry 30 rr_min_io 100 prio_callout "/sbin/mpath_prio_rdac /dev/%n" } } And sessions. [root@xnode1 dell]# iscsiadm -m session tcp: [13] 10.0.51.220:3260,1 iqn.1984-05.com.dell:powervault.md3600i.6d4ae520009bd7cc000000004fd7507c tcp: [14] 10.0.50.221:3260,2 iqn.1984-05.com.dell:powervault.md3600i.6d4ae520009bd7cc000000004fd7507c tcp: [15] 10.0.51.221:3260,2 iqn.1984-05.com.dell:powervault.md3600i.6d4ae520009bd7cc000000004fd7507c tcp: [16] 10.0.50.220:3260,1 iqn.1984-05.com.dell:powervault.md3600i.6d4ae520009bd7cc000000004fd7507c I'm getting very poor read performance : dd if=/dev/mapper/mpath1 of=/dev/null bs=1M count=1000 The SAN is configured as follows: CTRL0,PORT0 : 10.0.50.220 CTRL0,PORT1 : 10.0.50.221 CTRL1,PORT0 : 10.0.51.220 CTRL1,PORT1 : 10.0.51.221 And on the host : IF0 : 10.0.50.1 IF1 : 10.0.51.1 (Dual 10GbE Ethernet Card Intel DA2) It's connected to a 10gbE switch dedicated for SAN traffic. My questions being; why the connection is set up as 'ghost' and not 'ready' like an active/active configuration ?

    Read the article

  • Limiting interface bandwidth with tc under Linux

    - by Matt
    I have a linux router which has a 10GBe interface on the outside and bonded Gigabit ethernet interfaces on the inside. We have currently budget for 2GBit/s. If we exceed that rate by more than 5% average for a month then we'll be charged for the whole 10Gbit/s capacity. Quite a step up in dollar terms. So, I want to limit this to 2GBit/s on 10GBe interface. TBF filter might be ideal, but this comment is of concern. On all platforms except for Alpha, it is able to shape up to 1mbit/s of normal traffic with ideal minimal burstiness, sending out data exactly at the configured rates. Should I be using TBF or some other filter to apply this rate to the interface and how would I do it. I don't understand the example given here: Traffic Control HOWTO In particular "Example 9. Creating a 256kbit/s TBF" tc qdisc add dev eth0 handle 1:0 root dsmark indices 1 default_index 0 tc qdisc add dev eth0 handle 2:0 parent 1:0 tbf burst 20480 limit 20480 mtu 1514 rate 32000bps How is the 256K bit/s rate calculated? In this example, 32000bps = 32k bytes per second. Since tc uses bps = bytes per second. I guess burst and limit come into play but how would you go about choosing sensible numbers to reach the desired rate? This is not a mistake. I tested this and it gave a rate close to 256K but not exactly that.

    Read the article

  • How to best tune my SAN/Initiators for best performance?

    - by Disco
    Recent owner of a Dell PowerVault MD3600i i'm experiencing some weird results. I have a dedicated 24x 10GbE Switch (PowerConnect 8024), setup to jumbo frames 9K. The MD3600 has 2 RAID controllers, each has 2x 10GbE ethernet nics. There's nothing else on the switch; one VLAN for SAN traffic. Here's my multipath.conf defaults { udev_dir /dev polling_interval 5 selector "round-robin 0" path_grouping_policy multibus getuid_callout "/sbin/scsi_id -g -u -s /block/%n" prio_callout none path_checker readsector0 rr_min_io 100 max_fds 8192 rr_weight priorities failback immediate no_path_retry fail user_friendly_names yes # prio rdac } blacklist { device { vendor "*" product "Universal Xport" } # devnode "^sd[a-z]" } devices { device { vendor "DELL" product "MD36xxi" path_grouping_policy group_by_prio prio rdac # polling_interval 5 path_checker rdac path_selector "round-robin 0" hardware_handler "1 rdac" failback immediate features "2 pg_init_retries 50" no_path_retry 30 rr_min_io 100 prio_callout "/sbin/mpath_prio_rdac /dev/%n" } } And iscsid.conf : node.startup = automatic node.session.timeo.replacement_timeout = 15 node.conn[0].timeo.login_timeout = 15 node.conn[0].timeo.logout_timeout = 15 node.conn[0].timeo.noop_out_interval = 5 node.conn[0].timeo.noop_out_timeout = 10 node.session.iscsi.InitialR2T = No node.session.iscsi.ImmediateData = Yes node.session.iscsi.FirstBurstLength = 262144 node.session.iscsi.MaxBurstLength = 16776192 node.conn[0].iscsi.MaxRecvDataSegmentLength = 262144 After my tests; i can barely come to 200 Mb/s read/write. Should I expect more than that ? Providing it has dual 10 GbE my thoughts where to come around the 400 Mb/s. Any ideas ? Guidelines ? Troubleshooting tips ?

    Read the article

  • QLogic QLE8152 won't link up with a fiber loop

    - by Mike Pennington
    I have a Dell R710 running Debian Linux 6.0 (Squeeze). I installed a QLogic QLE8152 CNA in the PCI-E riser today and I have been trying to get the CNA's ethernet layer to come up after I put a fiber loop on it; I also scoped it with a light meter, and the SFP+ transceiver is getting the correct amount of light. The whole time I have been working on this problem, the lights on the CNA are blinking green at a rate of about 1 flash every 3 seconds. When I plug the fiber into a Juniper EX4500 10GE in a rack next to it, the Juniper's link stays down as well. I have to confess that this is my first wrangling with a CNA, so perhaps I'm doing something fundamentally wrong, but here is what I have found so far... First the basics... [mpenning@Finger ~]$ uname -a Linux Finger 2.6.32-5-amd64 #1 SMP Mon Oct 3 03:59:20 UTC 2011 x86_64 GNU/Linux [mpenning@Finger ~]$ cat /etc/issue Debian GNU/Linux 6.0 \n \l [mpenning@Finger ~]$ lspci -v shows that the card is properly installed (physically) 04:00.0 Ethernet controller: QLogic Corp. 10GbE Converged Network Adapter (TCP/IP Networking) (rev 02) Subsystem: QLogic Corp. Device 017e Flags: bus master, fast devsel, latency 0, IRQ 34 I/O ports at e000 [size=256] Memory at df8f0000 (64-bit, non-prefetchable) [size=16K] Memory at df900000 (64-bit, non-prefetchable) [size=1M] Expansion ROM at df800000 [disabled] [size=256K] Capabilities: [44] Power Management version 3 Capabilities: [4c] Express Endpoint, MSI 00 Capabilities: [88] MSI: Enable- Count=1/32 Maskable- 64bit+ Capabilities: [98] Vital Product Data Capabilities: [a0] MSI-X: Enable+ Count=8 Masked- Capabilities: [100] Advanced Error Reporting Capabilities: [138] Device Serial Number 00-c0-dd-ff-fe-1c-53-b4 Capabilities: [144] Power Budgeting <?> Kernel driver in use: qlge The output from ethtool shows that eth4 is the QLE8152... [mpenning@Finger ~]$ sudo ethtool eth4 Settings for eth4: Supported ports: [ FIBRE ] Supported link modes: 10000baseT/Full Supports auto-negotiation: No Advertised link modes: 10000baseT/Full Advertised pause frame use: No Advertised auto-negotiation: No Speed: 10000Mb/s Duplex: Full Port: FIBRE PHYAD: 0 Transceiver: external Auto-negotiation: on Current message level: 0x000060f7 (24823) Link detected: no [mpenning@Finger ~]$ sudo ethtool -i eth4 driver: qlge version: v1.00.00-b3 firmware-version: v1.35.11 bus-info: 0000:04:00.0 [mpenning@Finger ~]$ Finally, I tried modprobe -r qlge; modprobe -r qla2xxx and then modprobe qla2xxx; modprobe qlge to kick the system again... I don't see a smoking gun in /var/log/messages... Nov 17 19:46:21 finger kernel: [ 6212.298275] qlge 0000:04:00.1: PCI INT B disabled Nov 17 19:46:22 finger kernel: [ 6213.779974] qlge 0000:04:00.0: PCI INT A disabled Nov 17 19:46:33 finger kernel: [ 6224.554074] qla2xxx 0000:04:00.3: PCI INT D disabled Nov 17 19:46:33 finger kernel: [ 6224.555322] qla2xxx 0000:04:00.2: PCI INT C disabled Nov 17 19:46:54 finger kernel: [ 6245.625854] QLogic Fibre Channel HBA Driver: 8.03.01-k6 Nov 17 19:46:54 finger kernel: [ 6245.625888] qla2xxx 0000:04:00.2: PCI INT C -> GSI 35 (level, low) -> IRQ 35 Nov 17 19:46:54 finger kernel: [ 6245.626837] qla2xxx 0000:04:00.2: MSI-X vector count: 7 Nov 17 19:46:54 finger kernel: [ 6245.626841] qla2xxx 0000:04:00.2: Found an ISP8001, irq 35, iobase 0xffffc90012664000 Nov 17 19:46:54 finger kernel: [ 6245.627113] qla2xxx 0000:04:00.2: Configuring PCI space... Nov 17 19:46:54 finger kernel: [ 6245.639429] qla2xxx 0000:04:00.2: Configure NVRAM parameters... Nov 17 19:46:54 finger kernel: [ 6245.642597] qla2xxx 0000:04:00.2: Verifying loaded RISC code... Nov 17 19:46:54 finger kernel: [ 6245.642708] qla2xxx 0000:04:00.2: FW: Loading from flash (a0000)... Nov 17 19:46:55 finger kernel: [ 6246.273340] qla2xxx 0000:04:00.2: Allocated (64 KB) for FCE... Nov 17 19:46:55 finger kernel: [ 6246.273401] qla2xxx 0000:04:00.2: Allocated (64 KB) for EFT... Nov 17 19:46:55 finger kernel: [ 6246.273486] qla2xxx 0000:04:00.2: Allocated (1350 KB) for firmware dump... Nov 17 19:46:55 finger kernel: [ 6246.273856] scsi9 : qla2xxx Nov 17 19:46:55 finger kernel: [ 6246.274631] qla2xxx 0000:04:00.2: Nov 17 19:46:55 finger kernel: [ 6246.274633] QLogic Fibre Channel HBA Driver: 8.03.01-k6 Nov 17 19:46:55 finger kernel: [ 6246.274634] QLogic QLE8152 - QLogic PCI-Express Dual Channel 10GbE CNA Nov 17 19:46:55 finger kernel: [ 6246.274636] ISP8001: PCIe (5.0GT/s x4) @ 0000:04:00.2 hdma+, host#=9, fw=5.01.04 (8d4) Nov 17 19:46:55 finger kernel: [ 6246.274666] qla2xxx 0000:04:00.3: PCI INT D -> GSI 37 (level, low) -> IRQ 37 Nov 17 19:46:55 finger kernel: [ 6246.274748] qla2xxx 0000:04:00.3: MSI-X vector count: 7 Nov 17 19:46:55 finger kernel: [ 6246.274751] qla2xxx 0000:04:00.3: Found an ISP8001, irq 37, iobase 0xffffc900125c4000 Nov 17 19:46:55 finger kernel: [ 6246.275098] qla2xxx 0000:04:00.3: Configuring PCI space... Nov 17 19:46:55 finger kernel: [ 6246.287329] qla2xxx 0000:04:00.3: Configure NVRAM parameters... Nov 17 19:46:55 finger kernel: [ 6246.290624] qla2xxx 0000:04:00.3: Verifying loaded RISC code... Nov 17 19:46:55 finger kernel: [ 6246.290736] qla2xxx 0000:04:00.3: FW: Loading from flash (a0000)... Nov 17 19:46:55 finger kernel: [ 6246.920204] qla2xxx 0000:04:00.3: Allocated (64 KB) for FCE... Nov 17 19:46:55 finger kernel: [ 6246.920264] qla2xxx 0000:04:00.3: Allocated (64 KB) for EFT... Nov 17 19:46:55 finger kernel: [ 6246.920345] qla2xxx 0000:04:00.3: Allocated (1350 KB) for firmware dump... Nov 17 19:46:55 finger kernel: [ 6246.920749] scsi10 : qla2xxx Nov 17 19:46:55 finger kernel: [ 6246.921715] qla2xxx 0000:04:00.3: Nov 17 19:46:55 finger kernel: [ 6246.921716] QLogic Fibre Channel HBA Driver: 8.03.01-k6 Nov 17 19:46:55 finger kernel: [ 6246.921717] QLogic QLE8152 - QLogic PCI-Express Dual Channel 10GbE CNA Nov 17 19:46:55 finger kernel: [ 6246.921719] ISP8001: PCIe (5.0GT/s x4) @ 0000:04:00.3 hdma+, host#=10, fw=5.01.04 (8d4) Nov 17 19:46:58 finger kernel: [ 6249.519911] qlge 0000:04:00.0: PCI INT A -> GSI 34 (level, low) -> IRQ 34 Nov 17 19:46:58 finger kernel: [ 6249.533970] qlge 0000:04:00.0: QLogic 10 Gigabit PCI-E Ethernet Driver Nov 17 19:46:58 finger kernel: [ 6249.533975] qlge 0000:04:00.0: Driver name: qlge, Version: v1.00.00-b3. Nov 17 19:46:58 finger kernel: [ 6249.534856] qlge 0000:04:00.0: ql_display_dev_info: Function #0, Port 0, NIC Roll 0, NIC Rev = 1, XG Roll = 0, XG Rev = 1. Nov 17 19:46:58 finger kernel: [ 6249.534860] qlge 0000:04:00.0: ql_display_dev_info: MAC address 00:c0:dd:1c:53:b4 Nov 17 19:46:58 finger kernel: [ 6249.534888] qlge 0000:04:00.1: PCI INT B -> GSI 44 (level, low) -> IRQ 44 Nov 17 19:46:58 finger kernel: [ 6249.549109] qlge 0000:04:00.1: ql_display_dev_info: Function #1, Port 1, NIC Roll 0, NIC Rev = 1, XG Roll = 0, XG Rev = 1. Nov 17 19:46:58 finger kernel: [ 6249.549112] qlge 0000:04:00.1: ql_display_dev_info: MAC address 00:c0:dd:1c:53:b6 Any assistance getting a link up on this is greatly appreciated...

    Read the article

  • Planning trunk capacity for multiple GbE switches

    - by wuckachucka
    Without measuring throughput (it's at the top of the list; this is just theoretical), I want to know the most standard method for trunking VLANs on multiple Gigabit (GbE) switches to a core Layer 3 GbE switch. Say you have three VLANs: VLAN10 (10.0.0.0/24) Servers: your typical Windows DC/file server, Exchange, and an Accounting/SQL server. VLAN20: (10.0.1.0/24) Sales: needs access to everything on VLAN10; doesn't need access to VLAN30 and vice-versa. VLAN20: (10.0.1.0/24) Support: needs access to everything on VLAN10; doesn't need access to VLAN20 and vice-versa. Here's how I think this should work in my head: Switch #1: Ports 2-20 are assigned to VLAN20; all the Sales workstations and printers are connected here. Optional 10GbE combo port #1 is trunked to L3 switch's 10 GbE combo port #1. Switch #2: Ports 2-20 are assigned to VLAN30; all the Support workstations and printers are connected here. Optional 10GbE combo port #1 is trunked to L3 switch's 10 GbE combo port #2. Core L3 switch: Ports 2-10 are assigned to VLAN10; all three servers are connected here. With a standard 10/100 x 24 switch, it'll usually come with one or two 1 GbE uplink ports; carrying over this logic to a 10/100/1000 x 24, the "optional" 10 GbE combo ports that most higher-end switches can get shouldn't really be an option. Keep in mind I haven't tested anything yet, I'm primarily moving in this direction for growth (don't want to buy 10/100 switches and have to replace those within a couple of years) and security (being able to control access between VLANs with L3 routing/packet filtering ACLs). Does this sound right? Do I really need the 10 GbE ports? It seems very non-standard and expensive, but it "feels" right when you think about 40 or 50 workstations trunking up to the L3 switch over 1 GbE standard ports. If say 20 workstations want to download a 10 GB image from the servers concurrently, wouldn't the trunk be the bottleneck? At least if the trunk was 10 GbE, you'd have 10x1GbE nodes being able to reach their theoretical max. What about switch stacking? Some of the D-Links I've been looking at have HDMI interfaces for stacking. As far as I know, stacking two switches creates one logical switch, but is this just for management I/O or does the switches use the (assuming it's HDMI 1.3) 10.2 Gbps for carrying data back and forth?

    Read the article

  • Building a virtualized SPARC environment

    - by Owen Allen
    If you're interested in making effective use of virtualization tools like Oracle VM Server, there's a whitepaper on oracle.com that you should check out. The whitepaper starts with a few specific technologies and hardware: Oracle VM Server for SPARC, T4 Servers, Ops Center, Solaris 11, Sun Network 10GbE Switches, and Sun ZFS Storage Appliances. It then explains how to use them to plan and set up a virtualized environment, in which guests are grouped in Server Pools with high availability and are managed through Ops Center. It explains how Ops Center simplifies the management of logical domains by using custom plans to create new logical domains and managing their life cycle through its user interface. So, if you're interested in setting up a cloud and you want to avoid surprises along the way, have a look.

    Read the article

  • Nexenta/OpenSolaris filer kernel panic/crash

    - by ewwhite
    I've an x4540 Sun storage server running NexentaStor Enterprise. It's serving NFS over 10GbE CX4 for several VMWare vSphere hosts. There are 30 virtual machines running. For the past few weeks, I've had random crashes spaced 10-14 days apart. This system used to open OpenSolaris and was stable in that arrangement. The crashes trigger the automated system recovery feature on the hardware, forcing a hard system reset. Here's the output from mdb debugger: panic[cpu5]/thread=ffffff003fefbc60: Deadlock: cycle in blocking chain ffffff003fefb570 genunix:turnstile_block+795 () ffffff003fefb5d0 unix:mutex_vector_enter+261 () ffffff003fefb630 zfs:dbuf_find+5d () ffffff003fefb6c0 zfs:dbuf_hold_impl+59 () ffffff003fefb700 zfs:dbuf_hold+2e () ffffff003fefb780 zfs:dmu_buf_hold+8e () ffffff003fefb820 zfs:zap_lockdir+6d () ffffff003fefb8b0 zfs:zap_update+5b () ffffff003fefb930 zfs:zap_increment+9b () ffffff003fefb9b0 zfs:zap_increment_int+68 () ffffff003fefba10 zfs:do_userquota_update+8a () ffffff003fefba70 zfs:dmu_objset_do_userquota_updates+de () ffffff003fefbaf0 zfs:dsl_pool_sync+112 () ffffff003fefbba0 zfs:spa_sync+37b () ffffff003fefbc40 zfs:txg_sync_thread+247 () ffffff003fefbc50 unix:thread_start+8 () Any ideas what this means?

    Read the article

  • please demystify the 10Gb ethernet interfaces, cables

    - by maruti
    this really is a Dell question but tempted to ask the experts @ serverfault. choosen a Dell powerconnect 8024 10GbE switch. per the spec sheet this has 10GbaseT ports. "24x 10GBASE-T (10Gb/1Gb/100Mb) with 4x Combo Ports of SFP+ (10Gb/1Gb) or 10GBASE-T" the HBA on my storage server has 10G CX4 copper ports Dell does not sell any cables and this adds to my confusion. from the picture Dell 8024 seems to have RJ-45 type ports on the front panel? my question: is it a RJ-45 + CX4 cable or CX4 + CX4 cable?

    Read the article

  • ZFS on top of iSCSI

    - by Solipsism
    I'm planning on building out a file server using ZFS and BSD, and I was hoping to make it more expandable by attaching drives stored in other machines in the same rack via iSCSI (e.g., one machine is running ZFS, and others have iSCSI targets available to be connected to by the ZFS box and added to zpools). Looking for other people who have tried this has pretty much lead me to resources about exposing iSCSI shares on top of ZFS, but nothing about the reverse. Primarily I have the following questions: Is iSCSI over gigabit ethernet fast enough for this purpose, or would I have to switch to 10GbE to get decent performance? What would happen when one of the machines running iSCSI targets disconnects from the network? Is there a better way to do this that I just am not clever enough to have realized? Thanks for any help.

    Read the article

  • How to Avoid Your Next 12-Month Science Project

    - by constant
    While most customers immediately understand how the magic of Oracle's Hybrid Columnar Compression, intelligent storage servers and flash memory make Exadata uniquely powerful against home-grown database systems, some people think that Exalogic is nothing more than a bunch of x86 servers, a storage appliance and an InfiniBand (IB) network, built into a single rack. After all, isn't this exactly what the High Performance Computing (HPC) world has been doing for decades? On the surface, this may be true. And some people tried exactly that: They tried to put together their own version of Exalogic, but then they discover there's a lot more to building a system than buying hardware and assembling it together. IT is not Ikea. Why is that so? Could it be there's more going on behind the scenes than merely putting together a bunch of servers, a storage array and an InfiniBand network into a rack? Let's explore some of the special sauce that makes Exalogic unique and un-copyable, so you can save yourself from your next 6- to 12-month science project that distracts you from doing real work that adds value to your company. Engineering Systems is Hard Work! The backbone of Exalogic is its InfiniBand network: 4 times better bandwidth than even 10 Gigabit Ethernet, and only about a tenth of its latency. What a potential for increased scalability and throughput across the middleware and database layers! But InfiniBand is a beast that needs to be tamed: It is true that Exalogic uses a standard, open-source Open Fabrics Enterprise Distribution (OFED) InfiniBand driver stack. Unfortunately, this software has been developed by the HPC community with fastest speed in mind (which is good) but, despite the name, not many other enterprise-class requirements are included (which is less good). Here are some of the improvements that Oracle's InfiniBand development team had to add to the OFED stack to make it enterprise-ready, simply because typical HPC users didn't have the need to implement them: More than 100 bug fixes in the pieces that were not related to the Message Passing Interface Protocol (MPI), which is the protocol that HPC users use most of the time, but which is less useful in the enterprise. Performance optimizations and tuning across the whole IB stack: From Switches, Host Channel Adapters (HCAs) and drivers to low-level protocols, middleware and applications. Yes, even the standard HPC IB stack could be improved in terms of performance. Ethernet over IB (EoIB): Exalogic uses InfiniBand internally to reach high performance, but it needs to play nicely with datacenters around it. That's why Oracle added Ethernet over InfiniBand technology to it that allows for creating many virtual 10GBE adapters inside Exalogic's nodes that are aggregated and connected to Exalogic's IB gateway switches. While this is an open standard, it's up to the vendor to implement it. In this case, Oracle integrated the EoIB stack with Oracle's own IB to 10GBE gateway switches, and made it fully virtualized from the beginning. This means that Exalogic customers can completely rewire their server infrastructure inside the rack without having to physically pull or plug a single cable - a must-have for every cloud deployment. Anybody who wants to match this level of integration would need to add an InfiniBand switch development team to their project. Or just buy Oracle's gateway switches, which are conveniently shipped with a whole server infrastructure attached! IPv6 support for InfiniBand's Sockets Direct Protocol (SDP), Reliable Datagram Sockets (RDS), TCP/IP over IB (IPoIB) and EoIB protocols. Because no IPv6 = not very enterprise-class. HA capability for SDP. High Availability is not a big requirement for HPC, but for enterprise-class application servers it is. Every node in Exalogic's InfiniBand network is connected twice for redundancy. If any cable or port or HCA fails, there's always a replacement link ready to take over. This requires extra magic at the protocol level to work. So in addition to Weblogic's failover capabilities, Oracle implemented IB automatic path migration at the SDP level to avoid unnecessary failover operations at the middleware level. Security, for example spoof-protection. Another feature that is less important for traditional users of InfiniBand, but very important for enterprise customers. InfiniBand Partitioning and Quality-of-Service (QoS): One of the first questions we get from customers about Exalogic is: “How can we implement multi-tenancy?” The answer is to partition your IB network, which effectively creates many networks that work independently and that are protected at the lowest networking layer possible. In addition to that, QoS allows administrators to prioritize traffic flow in multi-tenancy environments so they can keep their service levels where it matters most. Resilient IB Fabric Management: InfiniBand is a self-managing network, so a lot of the magic lies in coming up with the right topology and in teaching the subnet manager how to properly discover and manage the network. Oracle's Infiniband switches come with pre-integrated, highly available fabric management with seamless integration into Oracle Enterprise Manager Ops Center. In short: Oracle elevated the OFED InfiniBand stack into an enterprise-class networking infrastructure. Many years and multiple teams of manpower went into the above improvements - this is something you can only get from Oracle, because no other InfiniBand vendor can give you these features across the whole stack! Exabus: Because it's not About the Size of Your Network, it's How You Use it! So let's assume that you somehow were able to get your hands on an enterprise-class IB driver stack. Or maybe you don't care and are just happy with the standard OFED one? Anyway, the next step is to actually leverage that InfiniBand performance. Here are the choices: Use traditional TCP/IP on top of the InfiniBand stack, Develop your own integration between your middleware and the lower-level (but faster) InfiniBand protocols. While more bandwidth is always a good thing, it's actually the low latency that enables superior performance for your applications when running on any networking infrastructure: The lower the latency, the faster the response travels through the network and the more transactions you can close per second. The reason why InfiniBand is such a low latency technology is that it gets rid of most if not all of your traditional networking protocol stack: Data is literally beamed from one region of RAM in one server into another region of RAM in another server with no kernel/drivers/UDP/TCP or other networking stack overhead involved! Which makes option 1 a no-go: Adding TCP/IP on top of InfiniBand is like adding training wheels to your racing bike. It may be ok in the beginning and for development, but it's not quite the performance IB was meant to deliver. Which only leaves option 2: Integrating your middleware with fast, low-level InfiniBand protocols. And this is what Exalogic's "Exabus" technology is all about. Here are a few Exabus features that help applications leverage the performance of InfiniBand in Exalogic: RDMA and SDP integration at the JDBC driver level (SDP), for Oracle Weblogic (SDP), Oracle Coherence (RDMA), Oracle Tuxedo (RDMA) and the new Oracle Traffic Director (RDMA) on Exalogic. Using these protocols, middleware can communicate a lot faster with each other and the Oracle database than by using standard networking protocols, Seamless Integration of Ethernet over InfiniBand from Exalogic's Gateway switches into the OS, Oracle Weblogic optimizations for handling massive amounts of parallel transactions. Because if you have an 8-lane Autobahn, you also need to improve your ramps so you can feed it with many cars in parallel. Integration of Weblogic with Oracle Exadata for faster performance, optimized session management and failover. As you see, “Exabus” is Oracle's word for describing all the InfiniBand enhancements Oracle put into Exalogic: OFED stack enhancements, protocols for faster IB access, and InfiniBand support and optimizations at the virtualization and middleware level. All working together to deliver the full potential of InfiniBand performance. Who else has 100% control over their middleware so they can develop their own low-level protocol integration with InfiniBand? Even if you take an open source approach, you're looking at years of development work to create, test and support a whole new networking technology in your middleware! The Extras: Less Hassle, More Productivity, Faster Time to Market And then there are the other advantages of Engineered Systems that are true for Exalogic the same as they are for every other Engineered System: One simple purchasing process: No headaches due to endless RFPs and no “Will X work with Y?” uncertainties. Everything has been engineered together: All kinds of bugs and problems have been already fixed at the design level that would have only manifested themselves after you have built the system from scratch. Everything is built, tested and integrated at the factory level . Less integration pain for you, faster time to market. Every Exalogic machine world-wide is identical to Oracle's own machines in the lab: Instant replication of any problems you may encounter, faster time to resolution. Simplified patching, management and operations. One throat to choke: Imagine finger-pointing hell for systems that have been put together using several different vendors. Oracle's Engineered Systems have a single phone number that customers can call to get their problems solved. For more business-centric values, read The Business Value of Engineered Systems. Conclusion: Buy Exalogic, or get ready for a 6-12 Month Science Project And here's the reason why it's not easy to "build your own Exalogic": There's a lot of work required to make such a system fly. In fact, anybody who is starting to "just put together a bunch of servers and an InfiniBand network" is really looking at a 6-12 month science project. And the outcome is likely to not be very enterprise-class. And it won't have Exalogic's performance either. Because building an Engineered System is literally rocket science: It takes a lot of time, effort, resources and many iterations of design/test/analyze/fix to build such a system. That's why InfiniBand has been reserved for HPC scientists for such a long time. And only Oracle can bring the power of InfiniBand in an enterprise-class, ready-to use, pre-integrated version to customers, without the develop/integrate/support pain. For more details, check the new Exalogic overview white paper which was updated only recently. P.S.: Thanks to my colleagues Ola, Paul, Don and Andy for helping me put together this article! var flattr_uid = '26528'; var flattr_tle = 'How to Avoid Your Next 12-Month Science Project'; var flattr_dsc = 'While most customers immediately understand how the magic of Oracle's Hybrid Columnar Compression, intelligent storage servers and flash memory make Exadata uniquely powerful against home-grown database systems, some people think that Exalogic is nothing more than a bunch of x86 servers, a storage appliance and an InfiniBand (IB) network, built into a single rack.After all, isn't this exactly what the High Performance Computing (HPC) world has been doing for decades?On the surface, this may be true. And some people tried exactly that: They tried to put together their own version of Exalogic, but then they discover there's a lot more to building a system than buying hardware and assembling it together. IT is not Ikea.Why is that so? Could it be there's more going on behind the scenes than merely putting together a bunch of servers, a storage array and an InfiniBand network into a rack? Let's explore some of the special sauce that makes Exalogic unique and un-copyable, so you can save yourself from your next 6- to 12-month science project that distracts you from doing real work that adds value to your company.'; var flattr_tag = 'Engineered Systems,Engineered Systems,Infiniband,Integration,latency,Oracle,performance'; var flattr_cat = 'text'; var flattr_url = 'http://constantin.glez.de/blog/2012/04/how-avoid-your-next-12-month-science-project'; var flattr_lng = 'en_GB'

    Read the article

  • Ops Center 12c - Provisioning Solaris Using a Card-Based NIC

    - by scottdickson
    It's been a long time since last I added something here, but having some conversations this last week, I got inspired to update things. I've been spending a lot of time with Ops Center for managing and installing systems these days.  So, I suspect a number of my upcoming posts will be in that area. Today, I want to look at how to provision Solaris using Ops Center when your network is not connected to one of the built-in NICs.  We'll talk about how this can work for both Solaris 10 and Solaris 11, since they are pretty similar.  In both cases, WANboot is a key piece of the story. Here's what I want to do:  I have a Sun Fire T2000 server with a Quad-GbE nxge card installed.  The only network is connected to port 2 on that card rather than the built-in network interfaces.  I want to install Solaris on it across the network, either Solaris 10 or Solaris 11.  I have met with a lot of customers lately who have a similar architecture.  Usually, they have T4-4 servers with the network connected via 10GbE connections. Add to this mix the fact that I use Ops Center to manage the systems in my lab, so I really would like to add this to Ops Center.  If possible, I would like this to be completely hands free.  I can't quite do that yet. Close, but not quite. WANBoot or Old-Style NetBoot? When a system is installed from the network, it needs some help getting the process rolling.  It has to figure out what its network configuration (IP address, gateway, etc.) ought to be.  It needs to figure out what server is going to help it boot and install, and it needs the instructions for the installation.  There are two different ways to bootstrap an installation of Solaris on SPARC across the network.   The old way uses a broadcast of RARP or more recently DHCP to obtain the IP configuration and the rest of the information needed.  The second is to explicitly configure this information in the OBP and use WANBoot for installation WANBoot has a number of benefits over broadcast-based installation: it is not restricted to a single subnet; it does not require special DHCP configuration or DHCP helpers; it uses standard HTTP and HTTPS protocols which traverse firewalls much more easily than NFS-based package installation.  But, WANBoot is not available on really old hardware and WANBoot requires the use o Flash Archives in Solaris 10.  Still, for many people, this is a great approach. As it turns out, WANBoot is necessary if you plan to install using a NIC on a card rather than a built-in NIC. Identifying Which Network Interface to Use One of the trickiest aspects to this process, and the one that actually requires manual intervention to set up, is identifying how the OBP and Solaris refer to the NIC that we want to use to boot.  The OBP already has device aliases configured for the built-in NICs called net, net0, net1, net2, net3.  The device alias net typically points to net0 so that when you issue the command  "boot net -v install", it uses net0 for the boot.  Our task is to figure out the network instance for the NIC we want to use.  We will need to get to the OBP console of the system we want to install in order to figure out what the network should be called.  I will presume you know how to get to the ok prompt.  Once there, we have to see what networks the OBP sees and identify which one is associated with our NIC using the OBP command show-nets. SunOS Release 5.11 Version 11.0 64-bit Copyright (c) 1983, 2011, Oracle and/or its affiliates. All rights reserved. {4} ok banner Sun Fire T200, No Keyboard Copyright (c) 1998, 2010, Oracle and/or its affiliates. All rights reserved. OpenBoot 4.30.4.b, 32640 MB memory available, Serial #69057548. Ethernet address 0:14:4f:1d:bc:c, Host ID: 841dbc0c. {4} ok show-nets a) /pci@7c0/pci@0/pci@2/network@0,1 b) /pci@7c0/pci@0/pci@2/network@0 c) /pci@780/pci@0/pci@8/network@0,3 d) /pci@780/pci@0/pci@8/network@0,2 e) /pci@780/pci@0/pci@8/network@0,1 f) /pci@780/pci@0/pci@8/network@0 g) /pci@780/pci@0/pci@1/network@0,1 h) /pci@780/pci@0/pci@1/network@0 q) NO SELECTION Enter Selection, q to quit: d /pci@780/pci@0/pci@8/network@0,2 has been selected. Type ^Y ( Control-Y ) to insert it in the command line. e.g. ok nvalias mydev ^Y for creating devalias mydev for /pci@780/pci@0/pci@8/network@0,2 {4} ok devalias ... net3 /pci@7c0/pci@0/pci@2/network@0,1 net2 /pci@7c0/pci@0/pci@2/network@0 net1 /pci@780/pci@0/pci@1/network@0,1 net0 /pci@780/pci@0/pci@1/network@0 net /pci@780/pci@0/pci@1/network@0 ... name aliases By looking at the devalias and the show-nets output, we can see that our Quad-GbE card must be the device nodes starting with  /pci@780/pci@0/pci@8/network@0.  The cable for our network is plugged into the 3rd slot, so the device address for our network must be /pci@780/pci@0/pci@8/network@0,2. With that, we can create a device alias for our network interface.  Naming the device alias may take a little bit of trial and error, especially in Solaris 11 where the device alias seems to matter more with the new virtualized network stack. So far in my testing, since this is the "next" network interface to be used, I have found success in naming it net4, even though it's a NIC in the middle of a card that might, by rights, be called net6 (assuming the 0th interface on the card is the next interface identified by Solaris and this is the 3rd interface on the card).  So, we will call it net4.  We need to assign a device alias to it: {4} ok nvalias net4 /pci@780/pci@0/pci@8/network@0,2 {4} ok devalias net4 /pci@780/pci@0/pci@8/network@0,2 ... We also may need to have the MAC for this particular interface, so let's get it, too.  To do this, we go to the device and interrogate its properties. {4} ok cd /pci@780/pci@0/pci@8/network@0,2 {4} ok .properties assigned-addresses 82060210 00000000 03000000 00000000 01000000 82060218 00000000 00320000 00000000 00008000 82060220 00000000 00328000 00000000 00008000 82060230 00000000 00600000 00000000 00100000 local-mac-address 00 21 28 20 42 92 phy-type mif ... From this, we can see that the MAC for this interface is  00:21:28:20:42:92.  We will need this later. This is all we need to do at the OBP.  Now, we can configure Ops Center to use this interface. Network Boot in Solaris 10 Solaris 10 turns out to be a little simpler than Solaris 11 for this sort of a network boot.  Since WANBoot in Solaris 10 fetches a specified In order to install the system using Ops Center, it is necessary to create a OS Provisioning profile and its corresponding plan.  I am going to presume that you already know how to do this within Ops Center 12c and I will just cover the differences between a regular profile and a profile that can use an alternate interface. Create a OS Provisioning profile for Solaris 10 as usual.  However, when you specify the network resources for the primary network, click on the name of the NIC, probably GB_0, and rename it to GB_N/netN, where N is the instance number you used previously in creating the device alias.  This is where the trial and error may come into play.  You may need to try a few instance numbers before you, the OBP, and Solaris all agree on the instance number.  Mark this as the boot network. For Solaris 10, you ought to be able to then apply the OS Provisioning profile to the server and it should install using that interface.  And if you put your cards in the same slots and plug the networks into the same NICs, this profile is reusable across multiple servers. Why This Works If you watch the console as Solaris boots during the OSP process, Ops Center is going to look for the device alias netN.  Since WANBoot requires a device alias called just net, Ops Center uses the value of your netN device alias and assigns that device to the net alias.  That means that boot net will automatically use this device.  Very cool!  Here's a trace from the console as Ops Center provisions a server: Sun Sun Fire T200, No KeyboardCopyright (c) 1998, 2010, Oracle and/or its affiliates. All rights reserved.OpenBoot 4.30.4.b, 32640 MB memory available, Serial #69057548.Ethernet address 0:14:4f:1d:bc:c, Host ID: 841dbc0c.auto-boot? =            false{0} ok  {0} ok printenv network-boot-argumentsnetwork-boot-arguments =  host-ip=10.140.204.234,router-ip=10.140.204.1,subnet-mask=255.255.254.0,hostname=atl-sewr-52,client-id=0100144F1DBC0C,file=http://10.140.204.22:5555/cgi-bin/wanboot-cgi{0} ok {0} ok devalias net net                      /pci@780/pci@0/pci@1/network@0{0} ok devalias net4 net4                     /pci@780/pci@0/pci@8/network@0,2{0} ok devalias net /pci@780/pci@0/pci@8/network@0,2{0} ok setenv network-boot-arguments host-ip=10.140.204.234,router-ip=10.140.204.1,subnet-mask=255.255.254.0,hostname=atl-sewr-52,client-id=0100144F1DBC0C,file=http://10.140.204.22:8004/cgi-bin/wanboot-cginetwork-boot-arguments =  host-ip=10.140.204.234,router-ip=10.140.204.1,subnet-mask=255.255.254.0,hostname=atl-sewr-52,client-id=0100144F1DBC0C,file=http://10.140.204.22:8004/cgi-bin/wanboot-cgi{0} ok {0} ok boot net - installBoot device: /pci@780/pci@0/pci@8/network@0,2  File and args: - install/pci@780/pci@0/pci@8/network@0,2: 1000 Mbps link up<time unavailable> wanboot info: WAN boot messages->console<time unavailable> wanboot info: configuring /pci@780/pci@0/pci@8/network@0,2 See what happened?  Ops Center looked for the network device alias called net4 that we specified in the profile, took the value from it, and made it the net device alias for the boot.  Pretty cool! WANBoot and Solaris 11 Solaris 11 requires an additional step since the Automated Installer in Solaris 11 uses the MAC address of the network to figure out which manifest to use for system installation.  In order to make sure this is available, we have to take an extra step to associate the MAC of the NIC on the card with the host.  So, in addition to creating the device alias like we did above, we also have to declare to Ops Center that the host has this new MAC. Declaring the NIC Start out by discovering the hardware as usual.  Once you have discovered it, take a look under the Connectivity tab to see what networks it has discovered.  In the case of this system, it shows the 4 built-in networks, but not the networks on the additional cards.  These are not directly visible to the system controller.  In order to add the additional network interface to the hardware asset, it is necessary to Declare it.  We will declare that we have a server with this additional NIC, but we will also  specify the existing GB_0 network so that Ops Center can associate the right resources together.  The GB_0 acts as sort of a key to tie our new declaration to the old system already discovered.  Go to the Assets tab, select All Assets, and then in the Actions tab, select Add Asset.  Rather than going through a discovery this time, we will manually declare a new asset. When we declare it, we will give the hostname, IP address, system model that match those that have already been discovered.  Then, we will declare both GB_0 with its existing MAC and the new GB_4 with its MAC.  Remember that we collected the MAC for GB_4 when we created its device alias. After you declare the asset, you will see the new NIC in the connectivity tab for the asset.  You will notice that only the NICs you listed when you declared it are seen now.  If you want Ops Center to see all of the existing NICs as well as the additional one, declare them as well.  Add the other GB_1, GB_2, GB_3 links and their MACs just as you did GB_0 and GB_4.  Installing the OS  Once you have declared the asset, you can create an OS Provisioning profile for Solaris 11 in the same way that you did for Solaris 10.  The only difference from any other provisioning profile you might have created already is the network to use for installation.  Again, use GB_N/netN where N is the interface number you used for your device alias and in your declaration.  And away you go.  When the system boots from the network, the automated installer (AI) is able to see which system manifest to use, based on the new MAC that was associated, and the system gets installed. {0} ok {0} ok printenv network-boot-argumentsnetwork-boot-arguments =  host-ip=10.140.204.234,router-ip=10.140.204.1,subnet-mask=255.255.254.0,hostname=atl-sewr-52,client-id=01002128204292,file=http://10.140.204.22:5555/cgi-bin/wanboot-cgi{0} ok {0} ok devalias net net                      /pci@780/pci@0/pci@1/network@0{0} ok devalias net4 net4                     /pci@780/pci@0/pci@8/network@0,2{0} ok devalias net /pci@780/pci@0/pci@8/network@0,2{0} ok setenv network-boot-arguments host-ip=10.140.204.234,router-ip=10.140.204.1,subnet-mask=255.255.254.0,hostname=atl-sewr-52,client-id=01002128204292,file=http://10.140.204.22:5555/cgi-bin/wanboot-cginetwork-boot-arguments =  host-ip=10.140.204.234,router-ip=10.140.204.1,subnet-mask=255.255.254.0,hostname=atl-sewr-52,client-id=01002128204292,file=http://10.140.204.22:5555/cgi-bin/wanboot-cgi{0} ok {0} ok boot net - installBoot device: /pci@780/pci@0/pci@8/network@0,2  File and args: - install/pci@780/pci@0/pci@8/network@0,2: 1000 Mbps link up<time unavailable> wanboot info: WAN boot messages->console<time unavailable> wanboot info: configuring /pci@780/pci@0/pci@8/network@0,2...SunOS Release 5.11 Version 11.0 64-bitCopyright (c) 1983, 2011, Oracle and/or its affiliates. All rights reserved.Remounting root read/writeProbing for device nodes ...Preparing network image for useDownloading solaris.zlib--2012-02-17 15:10:17--  http://10.140.204.22:5555/var/js/AI/sparc//solaris.zlibConnecting to 10.140.204.22:5555... connected.HTTP request sent, awaiting response... 200 OKLength: 126752256 (121M) [text/plain]Saving to: `/tmp/solaris.zlib'100%[======================================>] 126,752,256 28.6M/s   in 4.4s    2012-02-17 15:10:21 (27.3 MB/s) - `/tmp/solaris.zlib' saved [126752256/126752256] Conclusion So, why go to all of this trouble?  More and more, I find that customers are wiring their data center to only use higher speed networks - 10GbE only to the hosts.  Some customers are moving aggressively toward consolidated networks combining storage and network on CNA NICs.  All of this means that network-based provisioning cannot rely exclusively on the built-in network interfaces.  So, it's important to be able to provision a system using other than the built-in networks.  Turns out, that this is pretty straight-forward for both Solaris 10 and Solaris 11 and fits into the Ops Center deployment process quite nicely. Hopefully, you will be able to use this as you build out your own private cloud solutions with Ops Center.

    Read the article

  • Fibre channel long distance woes

    - by Marki
    I need a fresh pair of eyes. We're using a 15km fibre optic line across which fibrechannel and 10GbE is multiplexed (passive optical CWDM). For FC we have long distance lasers suitable up to 40km (Skylane SFCxx0404F0D). The multiplexer is limited by the SFPs which can do max. 4Gb fibrechannel. The FC switch is a Brocade 5000 series. The respective wavelengths are 1550,1570,1590 and 1610nm for FC and 1530nm for 10GbE. The problem is the 4GbFC fabrics are almost never clean. Sometimes they are for a while even with a lot of traffic on them. Then they may suddenly start producing errors (RX CRC, RX encoding, RX disparity, ...) even with only marginal traffic on them. I am attaching some error and traffic graphs. Errors are currently in the order of 50-100 errors per 5 minutes when with 1Gb/s traffic. Optics Here is the power output of one port summarized (collected using sfpshow on different switches) SITE-A units=uW (microwatt) SITE-B ********************************************** FAB1 SW1 TX 1234.3 RX 49.1 SW3 1550nm (ko) RX 95.2 TX 1175.6 FAB2 SW2 TX 1422.0 RX 104.6 SW4 1610nm (ok) RX 54.3 TX 1468.4 What I find curious at this point is the asymmetry in the power levels. While SW2 transmits with 1422uW which SW4 receives with 104uW, SW2 only receives the SW4 signal with similar original power only with 54uW. Vice versa for SW1-3. Anyway the SFPs have RX sensitivity down to -18dBm (ca. 20uW) so in any case it should be fine... But nothing is. Some SFPs have been diagnosed as malfunctioning by the manufacturer (the 1550nm ones shown above with "ko"). The 1610nm ones apparently are ok, they have been tested using a traffic generator. The leased line has also been tested more than once. All is within tolerances. I'm awaiting the replacements but for some reason I don't believe it will make things better as the apparently good ones don't produce ZERO errors either. Earlier there was active equipment involved (some kind of 4GFC retimer) before putting the signal on the line. No idea why. That equipment was eliminated because of the problems so we now only have: the long distance laser in the switch, (new) 10m LC-SC monomode cable to the mux (for each fabric), the leased line, the same thing but reversed on the other side of the link. FC switches Here is a port config from the Brocade portcfgshow (it's like that on both sides, obviously) Area Number: 0 Speed Level: 4G Fill Word(On Active) 0(Idle-Idle) Fill Word(Current) 0(Idle-Idle) AL_PA Offset 13: OFF Trunk Port ON Long Distance LS VC Link Init OFF Desired Distance 32 Km Reserved Buffers 70 Locked L_Port OFF Locked G_Port OFF Disabled E_Port OFF Locked E_Port OFF ISL R_RDY Mode OFF RSCN Suppressed OFF Persistent Disable OFF LOS TOV enable OFF NPIV capability ON QOS E_Port OFF Port Auto Disable: OFF Rate Limit OFF EX Port OFF Mirror Port OFF Credit Recovery ON F_Port Buffers OFF Fault Delay: 0(R_A_TOV) NPIV PP Limit: 126 CSCTL mode: OFF Forcing the links to 2GbFC produces no errors, but we bought 4GbFC and we want 4GbFC. I don't know where to look anymore. Any ideas what to try next or how to proceed? If we can't make 4GbFC work reliably I wonder what the people working with 8 or 16 do... I don't assume that "a few errors here and there" are acceptable. Oh and BTW we are in contact with everyone of the manufacturers (FC switch, MUX, SFPs, ...) Except for the SFPs to be changed (some have been changed before) nobody has a clue. Brocade SAN Health says the fabric is ok. MUX, well, it's passive, it's only a prism, nature at it's best. Any shots in the dark? APPENDIX: Answers to your questions @Chopper3: This is the second generation of Brocades exhibiting the problem. Before we had 5000s, now we have 5100s. In the beginning when we still had the active MUX we rented a longdistance laser once to put it into the switch directly in order to make tests for a day, during that day of course it was clean. But as I said, sometimes it's clean just like that. And sometimes it's not. Alternative switches would mean to rebuild the entire SAN with those only to test. Alternative SFPs, well they're hard to come by just like that. @longneck: The line is rented. It's a dark fibre (9um monomode) so there's noone else on it. Sure there are splices. I can't go and look but I have to trust they have been done correctly. As I said the line has been checked and rechecked (using an optical time-domain reflectometer). Obviously you don't have all this equipment yourself because it's way too expensive. @mdpc: What would be the "wrong" type of cable according to you? Up to the switch everything is monomode, yes. The connectors are the correct ones too. Yeah I know there are the green ones where the fibre is cut off at a certain angle etc. But we have the correct ones for all that I know. Progress Report #1 We have had two fabrics (=2x2 switches) with Brocade 5100s with FabricOS 6.4.1 and two fabrics (another 2x4 switches) on FabricOS 7.0.2. On the longdistance ISLs (one in each fabric) it turned out that with FOS 6.4.1 setting it to long distance issues warnings about the VC Init setting and consequently the fill word. But those are only warnings. FOS 7.0.2 requires you to do modifications to VCI and the fillword for long distance links. Setting FOS 6.4.1 to the LS (long-distance static distance) setting with wrong VCI and fillword setting made the whole fabric inoperational (stuck in an SCN loop, use fabriclog -s to see, you don't see it anywhere else, no port error counters or anything increasing). Currently I'm giving the one fabric with the IMHO more correct settings a beating and it seems to do fine, whereas the other one without much traffic still has errors here and there. In short: We have eliminated the active part of the MUX (the FC retimer). We are putting the long distance SFPs into the end equipment themselves. Just to be sure we bought new monomode cables to connect the end equipment to the remaining passive part of the MUX. We are now trying out several long distance configs. It's almost black magic. Everything that happens is mostly empirical, noone seems to have a clue what are the exact reasons to do something. ("We have tried this, and it didn't work, then we tried that and it worked, so we stuck with that." But noone really seems to know why.) I'll keep you updated. Progress Report #2 We got the new lasers for one of the fabrics on warranty. It's ultra clean even on 4GbFC. They're transmitting with roughly 2mW (3dBm) whereas the others are only at 1.5mW (1.5dBm) although that should really be enough. The other fabric (where the lasers are apparently ok) still produces one or two CRCs infrequently. Using sfpshow the SFP producing the actual RX errors shows Status/Ctrl: 0x82 Alarm flags[0,1] = 0x5, 0x40 Warn Flags[0,1] = 0x5, 0x40 Now I'll have to find out what that means. Not sure if it was there before. Well I'll first clear my head with a week of vacation. 8-)

    Read the article

  • Best Practices for Building a Virtualized SPARC Computing Environment

    - by Scott Elvington
    Normal 0 false false false EN-US X-NONE X-NONE /* Style Definitions */ table.MsoNormalTable {mso-style-name:"Table Normal"; mso-tstyle-rowband-size:0; mso-tstyle-colband-size:0; mso-style-noshow:yes; mso-style-priority:99; mso-style-qformat:yes; mso-style-parent:""; mso-padding-alt:0in 5.4pt 0in 5.4pt; mso-para-margin:0in; mso-para-margin-bottom:.0001pt; mso-pagination:widow-orphan; font-size:11.0pt; font-family:"Calibri","sans-serif"; mso-ascii-font-family:Calibri; mso-ascii-theme-font:minor-latin; mso-fareast-font-family:"Times New Roman"; mso-fareast-theme-font:minor-fareast; mso-hansi-font-family:Calibri; mso-hansi-theme-font:minor-latin; mso-bidi-font-family:"Times New Roman"; mso-bidi-theme-font:minor-bidi;} Oracle just published Best Practices for Building a Virtualized SPARC Computing Environment, a white paper that provides guidance on the complete hardware and software stack for deploying and managing your physical and virtual SPARC infrastructure. The solution is based on Oracle SPARC T4 servers, Oracle Solaris 11 with Oracle VM for SPARC 2.2, Sun ZFS storage appliances, Sun 10GbE 72 port switches and Oracle Enterprise Manager Ops Center 12c. The paper emphasizes the value and importance of planning the resources (compute, network and storage) that will comprise the virtualized environment to achieve the desired capacity, performance and availability characteristics. The document also details numerous operational best practices that will help you deliver on those characteristics with unique capabilities provided by Enterprise Manager Ops Center including policy-based guest placement, pool resource balancing and automated guest recovery in the event of server failure. Plenty of references to supplementary documentation are included to help point you to additional resources. Whether you’re building the first stages of your private cloud or a general-purpose virtualized SPARC computing environment, these documented best practices will help ensure success. Please join Phil Bullinger and Steve Wilson from Oracle to learn more about breakthrough efficiency in private cloud infrastructure and how SPARC based virtualization can help you get started on your cloud journey. Stay Connected: Twitter |  Face book |  You Tube |  Linked in |  Newsletter

    Read the article

  • Basic multicast network performance problems

    - by davedavedave
    I've been using mpong from 29west's mtools package to get some basic idea of multicast latency across various Cisco switches: 1Gb 2960G, 10Gb 4900M and 10Gb Nexus N5548P. The 1Gb is just for comparison. I have the following results for ~400 runs of mpong on each switch (sending 65536 "ping"-like messages to a receiver which then sends back -- all over multicast). Numbers are latencies measured in microseconds. Switch Average StdDev Min Max 2960 (1Gb) 109.68463 0.092816 109.4328 109.9464 4900M (10Gb) 705.52359 1.607976 703.7693 722.1514 NX 5548(10Gb) 58.563774 0.328242 57.77603 59.32207 The result for 4900M is very surprising. I've tried unicast ping and I see the 4900 has ~10us higher latency than the N5548P (average 73us vs 64us). Iperf (with no attempt to tune it) shows both 10Gb switches give me 9.4Gbps line speed. The two machines are connected to the same switch and we're not doing any multicast routing. OS is RHEL 6. 10Gb NICs are HP 10GbE PCI-E G2 Dual-port NICs (I believe they are rebranded Mellanox cards). The 4900 switch is used in a project with tight access control so I'm waiting for approval before I can access it and check the config. The other two I have full access to configure. I've looked at the Cisco document[2] detailing differences between NX-OS and IOS w.r.t multicast so I've got some ideas to try out but this isn't an area where I have much expertise. Does anyone have any idea what I should be looking at once I get access to the switch? [1] http://docwiki.cisco.com/wiki/Cisco_NX-OS/IOS_Multicast_Comparison

    Read the article

  • VMWare vSphere 5: 4 pNICs for iSCSI vs. 2 pNICs

    - by gravyface
    New SAN for me, never used before: it's an IBM DS3512, dual controller with a quad 1GbE NIC per controller that a client bought and needs help setting up. Hosts (x2) have 8 pNICs and while I usually reserve 2 pNICs for iSCSI per host (and 2 for VM, 2 for management, 2 for vMotion, staggered across adapters), these extra ports on the SAN have me wondering if storage I/O would be significantly improved with 2 additional NICs per host, or if the limitations of the vmkernel/initiator would prevent the additional multipaths from ever being realized. I'm not seeing alot of 4 pNIC iSCSI implementations per host; 2 is the de facto standard from what I've read/seen online. I could and probably will do some I/O testing, but just wondering if there's a "wall" that someone else has discovered long ago (i.e. before 10GbE) that makes a 4 NIC iSCSI per host setup somewhat pointless. Just to clarify: I'm not looking for a how-to, but an explanation (link to paper, VMWare recommendation, benchmark, etc.) as to why 2-NIC configurations are the norm vs. 4-NIC iSCSI configurations. i.e. storage vendor limitations, VMKernel/initiator limitations, etc.

    Read the article

1 2  | Next Page >