Search Results

Search found 16644 results on 666 pages for 'traffic management'.

Page 231/666 | < Previous Page | 227 228 229 230 231 232 233 234 235 236 237 238  | Next Page >

  • How to handle server failure in an n-tier architecture?

    - by andy
    Imagine I have an n-tier architecture in an auto-scaled cloud environment with say: a load balancer in a failover pair reverse proxy tier web app tier db tier Each tier needs to connect to the instances in the tier below. What are the standard ways of connecting tiers to make them resilient to failure of nodes in each tier? i.e. how does each tier get the IP addresses of each node in the tier below? For example if all reverse proxies should route traffic to all web app nodes, how could they be set up so that they don't send traffic to dead web app nodes, and so that when new web app nodes are brought online they can send traffic to it? I could run an agent that would update all the configs to all the nodes, but it seems inefficient. I could put an LB pair between each tier, so the tier above only needs to connect to the load balancers, but how do I handle the problem of the LBs dying? This just seems to shunt the problem of tier A needing to know the IPs of all nodes in tier B, to all nodes in tier A needing to know the IPs of all LBs between tiers A and B. For some applications, they can implement retry logic if they contact a node in the tier below that doesn't respond, but is there any way that some middleware could direct traffic to only live nodes in the following tier? If I was hosting on AWS I could use an ELB between tiers, but I want to know how I could achieve the same functionality myself. I've read (briefly) about heartbeat and keepalived - are these relevant here? What are the virtual IPs they talk about and how are they managed? Are there still single points of failure using them?

    Read the article

  • Connect to Postgres remotely, open port 5432 for Postgres in iptables

    - by Victor
    I am trying to connect to Postgres remotely but I need to open port 5432 in iptables. My current iptables configuration is as follows: *filter # Allows all loopback (lo0) traffic and drop all traffic to 127/8 that doesn't use lo0 -A INPUT -i lo -j ACCEPT -A INPUT ! -i lo -d 127.0.0.0/8 -j REJECT # Accepts all established inbound connections -A INPUT -m state --state ESTABLISHED,RELATED -j ACCEPT # Allows all outbound traffic # You can modify this to only allow certain traffic -A OUTPUT -j ACCEPT # Allows HTTP and HTTPS connections from anywhere (the normal ports for websites) -A INPUT -p tcp --dport 80 -j ACCEPT -A INPUT -p tcp --dport 443 -j ACCEPT # Allows SSH connections # # THE -dport NUMBER IS THE SAME ONE YOU SET UP IN THE SSHD_CONFIG FILE # -A INPUT -p tcp -m state --state NEW --dport 30000 -j ACCEPT # Allow ping -A INPUT -p icmp -m icmp --icmp-type 8 -j ACCEPT # log iptables denied calls -A INPUT -m limit --limit 5/min -j LOG --log-prefix "iptables denied: " --log-level 7 # Reject all other inbound - default deny unless explicitly allowed policy -A INPUT -j REJECT -A FORWARD -j REJECT COMMIT What would I have to add in iptables to open the port? I'm trying to install phppgadmin on a different server to access the postgres database. Thank you.

    Read the article

  • Inter-VLAN Malicious Code Scanning

    - by Jackthedog
    I am trying to find an inbuilt solution on a Cisco Catayst 3750X Switch to scan all traffic routed from one VLAN to another for malicious code. The situation is that we currently have a development environment which is currently being redesigned to upgrade the network infrastructure to use the 3750X switches to manage server and workstation connectivity as well as inter-VLAN routing. We also have another system that is responsible for taking the builds created on the development environment and imaging various HDDs. Because these are two separate systems, we have a requirement in the workplace to anti-virus scan any data transferred between these systems. This is done by copying the data from the originating system to external USB HDD, scanning in a standalone workstation and then copying the data on to the receiving system. As you can imagine this is extremely tedious and impractical most of the time... (I don't make the rules). Anyway, with this redesign going on, we would like to join the imaging system to the network infrastructure of the development system, keeping separation by the use of VLANs and restricting traffic by using ACLs. As we still have the requirement to scan all traffic I would like to configure some sort of malicious code scanning when ever traffic is routed between these VLANs. I am aware I could install a separate in-line IPS/IDS device, however both systems will be using multiple ports on the switch (obviously), and we won't be able to put a device on each port. I would would prefer not to add additional hardware if the 3750x switch is capable of doing the job. Is anyone aware of any Cisco solution that I could use here, that ideally can be incorporated into the 3750x switch? Thanks in advance.

    Read the article

  • How can I get my routers to forward ports correctly?

    - by Giffyguy
    My network currently looks like this (simplified): Note that Router #2 is connected to the LAN interface of Router #1. This should be familiar to anyone who has seen a standard static-IP setup with an additional firewall for a residence or other small building. Router #1 is actually my cable gateway, but since it is a fully functional router/firewall, I am going to refer to it as a router. Now, I need to open various ports in both firewalls for incoming communication to my server - port 80 is a good example. So I've opened up port 80 in Router #2, and so far all incoming traffic at the public IP X.X.X.129 is being routed correctly. The problem is that I also need my server to respond to incoming traffic at the public IP X.X.X.130 on the WAN interface of Router #1. Naturally, I can't just tell Router #1 to forward port 80 to another public IP. Port forwarding is only supported when the traffic is being directed to the LAN subnet. I am willing to restructure my network topology if required, with the following conditions: Router #1 cannot have its WAN IP reassigned - X.X.X.130 is mandatory. Router #1 cannot be moved or disconnected from the cloud. The server cannot be given a second IP address. I would prefer the server to have a private IP address - e.g. 10.0.0.10 I'd like to keep Router #2, but it can have a private IP - e.g. 10.0.1.10 Following these rules, I need to get my server to receive incoming traffic on port 80 from both public IP addresses. Does anyone on SU know if this is possible? So far my only theories have been to set up a static route on either router, or to somehow combine my two subnets into a single subnet.

    Read the article

  • Request bursting from web application Load Tests

    - by MaseBase
    I'm migrating our web and database hosting to a new environment on all new machines. I've recently performed a Load Test using WAPT to generate load from multiple distributed clients. The server has plenty of room to handle the traffic load, but I'm seeing an odd pattern of incoming traffic during the load tests. Here is the gist of our setup: Firewall server running MS Forefront TMG 2010 on Win 2k8 server Request routing done by IIS Application Request Routing on firewall machine Web server is a Hyper-V VM on the Database server (which is the host OS) These machines are hefty with dual-CPU's with six cores (12 total procs) Web server running IIS 7.5 Web applications built in ASP.NET 2.0, with 1 ISAPI filter (Url Rewrite) in front What I'm seeing during the load tests is that the requests all come through in bursts. Even though I have 7 different distributed clients sending traffic loads, the requests come through about 300-500 requests at a time. The performance monitor shows nearly all of the counters moving through this pattern, where a burst of requests comes in the req/sec jumps to 70, the queued requests jumps to 500, the current requests jumps up, the CPU jumps up, everything. Then once it's handled that group of requests, it has a lull for nearly 10 seconds where nearly nothing is happening. 0-5 req/sec, 0 queued requests, minimal CPU usage. Then after 10 seconds of inactivity, another burst comes through, spiking all of the counters once again. What I can't figure out is why the requests are coming through in bursts when I know that the load being generated is not sent that way, especially considering the various load-generating clients sending traffic all in different intervals with random think time's between each request. Is there something in the layers between Hyper-V or perhaps in the hardware which might cause this coalesce of requests together? Here is what i'm looking at, the highlighted metric is Requests/sec, but the others critical counter go with it: Requests Queued (which I'd obviously like to keep as close to 0 as possible). Any ideas on this?

    Read the article

  • Hyper-V Virtual Machine Networking issues related to Max Ethernet Frame Size

    - by Goatmale
    I fixed an issue today earlier today but i'm interested in learning WHY it worked. We set up a new Hyper-V virtual machine only to discover that HTTP traffic wasn't working. HTTPS, pings, everything else was working fine. After months of prodding around I took a shot in the dark. On the Hyper-V host server, the physical NIC card had an advanced setting of "Max Ethernet Frame Size" set to 1500. After setting this setting to 1514 the issue was fixed. Alternatively, setting this to 1512 did not solve the issue; 1514 is the magic number. My best guess it that when this setting was set to 1500 it was allowing incoming pings because the data payload was a lot smaller of say, HTTP traffic. As far as HTTPS traffic, I read about something called "Path MTU discovery" which i'm going to assume why is HTTPs traffic was getting through fine, albeit slower. Looking at this post, people agree that 1518 is the max total frame size. Why didn't I need to change this to 1518 instead of 1514 bytes? Why is the default frame size 1500 if that's the max size of the Ethernet payload and not the max size.

    Read the article

  • How to populate RRD database with CPU and MEM usage data?

    - by Tomaszs
    I have a Lighttpd server (on Centos) and would like to display 4 graphs: lighttpd traffic, lighttpd requests per second, CPU usage and MEM usage. I've set place for rrd database for lighttpd config like this: rrdtool.binary = "/usr/bin/rrdtool" rrdtool.db-name = "/var/www/lighttpd.rrd" And put into my WWW cgi-bin sh file that gets data from lighttpd RRD file and creates graphs of traffic and requests per second like this: #!/bin/sh RRDTOOL=/usr/bin/rrdtool OUTDIR=//var/www/graphs INFILE=/var/www/lighttpd.rrd OUTPRE=lighttpd-traffic WIDTH=400 HEIGHT=100 DISP="-v bytes --title TrafficWebserver \ DEF:binraw=$INFILE:InOctets:AVERAGE \ DEF:binmaxraw=$INFILE:InOctets:MAX \ DEF:binminraw=$INFILE:InOctets:MIN \ DEF:bout=$INFILE:OutOctets:AVERAGE \ DEF:boutmax=$INFILE:OutOctets:MAX \ DEF:boutmin=$INFILE:OutOctets:MIN \ CDEF:bin=binraw,-1,* \ CDEF:binmax=binmaxraw,-1,* \ CDEF:binmin=binminraw,-1,* \ CDEF:binminmax=binmaxraw,binminraw,- \ CDEF:boutminmax=boutmax,boutmin,- \ AREA:binmin#ffffff: \ STACK:binmax#f00000: \ LINE1:binmin#a0a0a0: \ LINE1:binmax#a0a0a0: \ LINE2:bin#efb71d:incoming \ GPRINT:bin:MIN:%.2lf \ GPRINT:bin:AVERAGE:%.2lf \ GPRINT:bin:MAX:%.2lf \ AREA:boutmin#ffffff: \ STACK:boutminmax#00f000: \ LINE1:boutmin#a0a0a0: \ LINE1:boutmax#a0a0a0: \ LINE2:bout#a0a735:outgoing \ GPRINT:bout:MIN:%.2lf \ GPRINT:bout:AVERAGE:%.2lf \ GPRINT:bout:MAX:%.2lf \ " $RRDTOOL graph $OUTDIR/$OUTPRE-hour.png -a PNG --start -14400 $DISP -w $WIDTH -h $HEIGHT $RRDTOOL graph $OUTDIR/$OUTPRE-day.png -a PNG --start -86400 $DISP -w $WIDTH -h $HEIGHT $RRDTOOL graph $OUTDIR/$OUTPRE-month.png -a PNG --start -2592000 $DISP -w $WIDTH -h $HEIGHT OUTPRE=lighttpd-requests DISP="-v req --title RequestsperSecond -u 1 \ DEF:req=$INFILE:Requests:AVERAGE \ DEF:reqmax=$INFILE:Requests:MAX \ DEF:reqmin=$INFILE:Requests:MIN \ CDEF:reqminmax=reqmax,reqmin,- \ AREA:reqmin#ffffff: \ STACK:reqminmax#00f000: \ LINE1:reqmin#a0a0a0: \ LINE1:reqmax#a0a0a0: \ LINE2:req#00a735:requests" $RRDTOOL graph $OUTDIR/$OUTPRE-hour.png -a PNG --start -14400 $DISP -w $WIDTH -h $HEIGHT $RRDTOOL graph $OUTDIR/$OUTPRE-day.png -a PNG --start -86400 $DISP -w $WIDTH -h $HEIGHT $RRDTOOL graph $OUTDIR/$OUTPRE-month.png -a PNG --start -2592000 $DISP -w $WIDTH -h $HEIGHT Basically it's not my script, i get it from somewhere from the internet. Now i would like to do the same for CPU usage and MEM usage. I don't like to use any additional packages! As you can see lighttpd populates lighttpd.rrd file with traffic data and requests per second. Now i would like to the system to populate second rrd file with CPU and MEM usage, so i can add to sh file code to generate graphs for this data. How can I populate RRD file with CPU and MEM usage data? Please, NO THIRD-PARTY tools !

    Read the article

  • iptables-restore: line 1 failed

    - by Doug
    Hello, I am new to servers, and I was following this guide and it failed on the first command instructed. Could anyone give me a hand? http://wiki.debian.org/iptables ~ZORO~:/etc# iptables-restore < /etc/iptables.test.rules iptables-restore: line 1 failed Edit: iptables.test.rules ~ZORO~:/etc# cat /etc/iptables.test.rules *filter # Allows all loopback (lo0) traffic and drop all traffic to 127/8 that doesn't use lo0 -A INPUT -i lo -j ACCEPT -A INPUT -i ! lo -d 127.0.0.0/8 -j REJECT # Accepts all established inbound connections -A INPUT -m state --state ESTABLISHED,RELATED -j ACCEPT # Allows all outbound traffic # You could modify this to only allow certain traffic -A OUTPUT -j ACCEPT # Allows HTTP and HTTPS connections from anywhere (the normal ports for websites) -A INPUT -p tcp --dport 80 -j ACCEPT -A INPUT -p tcp --dport 443 -j ACCEPT # Allows SSH connections for script kiddies # THE -dport NUMBER IS THE SAME ONE YOU SET UP IN THE SSHD_CONFIG FILE -A INPUT -p tcp -m state --state NEW --dport 30000 -j ACCEPT # Now you should read up on iptables rules and consider whether ssh access # for everyone is really desired. Most likely you will only allow access from certain IPs. # Allow ping -A INPUT -p icmp -m icmp --icmp-type 8 -j ACCEPT # log iptables denied calls (access via 'dmesg' command) -A INPUT -m limit --limit 5/min -j LOG --log-prefix "iptables denied: " --log-level 7 # Reject all other inbound - default deny unless explicitly allowed policy: -A INPUT -j REJECT -A FORWARD -j REJECT COMMIT

    Read the article

  • WEP/WPA/WPA2 and wifi sniffing

    - by jcea
    Hi, I know that WEP traffic can be "sniffed" by any user of the WIFI. I know that WPA/WPA2 traffic is encrypted using a different link key for each user, so they can't sniff traffic... unless they capture the initial handshake. If you are using a PSK (preshared key) schema, then you recover the link key trivially from this initial handshake. If you don't know the PSK, you can capture the handshake and try to crack the PSK by bruteforce offline. Is my understanding correct so far?. I know that WPA2 has AES mode and can use "secure" tokens like X.509 certificates and such, and it is said to be secure against sniffing because capturing the handshake doesn't help you. So, is WPA2+AES secure (so far) against sniffing, and how it actually works?. That is, how is the (random) link key negociated?. When using X.509 certificates or a (private and personal) passphrase. Do WPA/WPA2 have other sniffer-secure modes beside WPA2+AES? How is broadcast traffic managed to be received by all the WIFI users, if each has a different link key?. Thanks in advance! :).

    Read the article

  • nginx: dump HTTP requests for debugging

    - by Alexander Gladysh
    Ubuntu 10.04.2 nginx 0.7.65 I see some weird HTTP requests coming to my nginx server. To better understand what is going on, I want to dump whole HTTP request data for such queries. (I.e. dump all request headers and body somewhere I can read them.) Can I do this with nginx? Alternatively, is there some HTTP server that allows me to do this out of the box, to which I can proxy these requests by the means of nginx? Update: Note that this box has a bunch of normal traffic, and I would like to avoid capturing all of it on low level (say, with tcpdump) and filtering it out later. I think it would be much easier to filter good traffic first in a rewrite rule (fortunately I can write one quite easily in this case), and then deal with bogus traffic only. And I do not want to channel bogus traffic to another box just to be able to capture it there with tcpdump. Update 2: To give a bit more details, bogus request have parameter named (say) foo in their GET query (the value of the parameter can differ). Good requests are guaranteed not to have this parameter ever. If I can filter by this in tcpdump or ngrep somehow — no problem, I'll use these.

    Read the article

  • Privacy, VPN and routers

    - by user123189
    Ever since this ACTA push-up the things are starting to heat up around torrents and privacy. I am using Tribler now, but this is not secure enough for me. Not enough privacy. I've been using in the past a swedish VPN PPTP connection. What I observed is that, when the VPN connection was down, Internet traffic wasn't cut off, rather the downloads were continuing, this time with my real IP, wearing off my protection. 1st : How to enforce a VPN connection that will cut all traffic when down? That is, the moment the connection is down, all internet traffic should cease as if I'd pull the network plug out. 2nd: Is PPTP good enough or should I ask for SSTP or IKEV2 ? 3rd: Should I disable IPv6 ? Is VPN no longer private if I keep IPv6 active? I 'heard some stuff' about dual vpn routers to be able to improve privacy; but nothing more about how to configure one for such a task. 4th: Is there any kind of "black box" hardware equipment that can be used in hiding IP, encrypting traffic and so on ?

    Read the article

  • Encrypting peer-to-peer application with iptables and stunnel

    - by Jonathan Oliver
    I'm running legacy applications in which I do not have access to the source code. These components talk to each other using plaintext on a particular port. I would like to be able to secure the communications between the two or more nodes using something like stunnel to facilitate peer-to-peer communication rather than using a more traditional (and centralized) VPN package like OpenVPN, etc. Ideally, the traffic flow would go like this: app@hostA:1234 tries to open a TCP connection to app@hostB:1234. iptables captures and redirects the traffic on port 1234 to stunnel running on hostA at port 5678. stunnel@hostA negotiates and establishes a connection with stunnel@hostB:4567. stunnel@hostB forwards any decrypted traffic to app@hostB:1234. In essence, I'm trying to set this up to where any outbound traffic (generated on the local machine) to port N forwards through stunnel to port N+1, and the receiving side receives on port N+1, decrypts, and forwards to the local application at port N. I'm not particularly concerned about losing the hostA origin IP address/machine identity when stunnel@hostB forwards to app@hostB because the communications payload contains identifying information. The other trick in this is that normally with stunnel you have a client/server architecture. But this application is much more P2P because nodes can come and go dynamically and hard-coding some kind of "connection = hostN:port" in the stunnel configuration won't work.

    Read the article

  • Setting up vsftpd, hangs on list command

    - by Victor
    I installed vsftpd and configured it. When I try to connect to the ftp server using Transmit, it manages to connect but hangs on Listing "/" Then, I get a message stating: Could not retrieve file listing for “/”. Control connection timed out. Does it have anything to do with my iptables? My rules are as listed: *filter # Allows all loopback (lo0) traffic and drop all traffic to 127/8 that doesn't use lo0 -A INPUT -i lo -j ACCEPT -A INPUT ! -i lo -d 127.0.0.0/8 -j REJECT # Accepts all established inbound connections -A INPUT -m state --state ESTABLISHED,RELATED -j ACCEPT # Allows all outbound traffic # You can modify this to only allow certain traffic -A OUTPUT -j ACCEPT # Allows HTTP and HTTPS connections from anywhere (the normal ports for websites) -A INPUT -p tcp -m tcp --dport 80 -j ACCEPT -A INPUT -p tcp -m tcp --dport 443 -j ACCEPT -A INPUT -p tcp -m tcp --dport 21 -j ACCEPT # Allows SSH connections # # THE -dport NUMBER IS THE SAME ONE YOU SET UP IN THE SSHD_CONFIG FILE # -A INPUT -p tcp -m state --state NEW --dport 30000 -j ACCEPT # Allow ping -A INPUT -p icmp -m icmp --icmp-type 8 -j ACCEPT # log iptables denied calls -A INPUT -m limit --limit 5/min -j LOG --log-prefix "iptables denied: " --log-level 7 # Reject all other inbound - default deny unless explicitly allowed policy -A INPUT -j REJECT -A FORWARD -j REJECT COMMIT

    Read the article

  • iSCSI performance questions

    - by RyanLambert
    Hi everyone, apologies for the long-winded post in advance... Attempting to troubleshoot some iSCSI sluggishness on a brand new vSphere deployment (still in test). Layout is as such: 3 VSphere hosts, each with 2x 10GB NICs plugged into a pair of Nexus 5020s with a 10gig back-to-back between them. NICs are port-channeled in an active/active redundant fashion (using vPC-mac pinning for those of you familiar with N1KV) Both NICs carry service console, vmotion, iSCSI, and guest traffic. iSCSI is on a single subnet/single VLAN that is not routed through our IP network (strictly layer2) Had this been a 1gig deployment, we probably would have split the iSCSI traffic off onto separate NICs, but the price/port gets rather ridiculous when you start throwing 4+ NICs to a server in a 10gigabit infrastructure, and I'm not really convinced it's necessary. Open to dialogue/tech facts re: this, though. At this point even a single VM guest will boot slowly to iSCSI storage (EMC CX4 on the same Nexus 5020 10gig switches), and restores of VMs from iSCSI take about twice as long as we'd expect them to. Our server folks mentioned that if we split the iSCSI off onto its own NIC, performance seems significantly better. From a network perspective, I've run through the variables I can think of (port configuration errors, MTU problems, congestion etc.) and I'm coming up dry. There really is no other traffic on these hosts other than the very specific test being performed at the time. Important thing to note is that guest traffic works just fine... it seems storage is the only thing affected by whatever gremlin exists. Concluding that we're not 'overutilizing' the network infrastructure since we're doing hardly anything, I'm just looking for some helpful tips/ideas we can use to resolve this... preferably without hurling extra 10gig NICs that are going to sit around 10% utilization while we've got 70+% left on our others.

    Read the article

  • Offloading backups to secondary network

    - by user1467163
    I'm trying to solve a problem- Currently, we are constantly backing up and have no budget for additional servers. Our production network is still a 10/100 and handles voip, SQL plus our backup traffic, and I'd like to offload the backup traffic onto a secondary network- all of our servers have secondary NIC's that are not in use, and all support gigabit (Our switching hardware does not- a topic for another day). I'd like to move my backups off the production network, but I am having a hard time getting the computers to communicate. I am using a Netgear GS724T switch for the backup network- Chosen for cost and because I have used them extensively on networks saturated with ghosting traffic, so I know it's up to the task. I have defined a VLAN, with ports that are not members of any other VLAN. All traffic is untagged on the VLAN. I have set the servers with 192.168.1.10 and 192.168.1.11 addresses, 255.255.255.0 netmask and I have tried a blank GW, using the local IP of the server 192.168.1.whatever address, and I have tried using the switch's production-side IP as the GW. The machines cannot find each other. DNS addresses are blank because I am going purely by IP for now... Any ideas how to get these machines to talk? they are Windows machines, running Server 2008R2 and 2003R2. Thanks!

    Read the article

  • iptables: How to combine DNAT and SNAT to use a secondary IP address?

    - by Que_273
    There are lots of questions on here about iptables DNAT/SNAT setups but I haven't found one that solves my current problem. I have services bound to the IP address of eth0 (e.g. 192.168.0.20) and I also have a IP address on eth0:0 (192.168.0.40) which is shared with another server. Only one server is active, so this alias interface comes and goes depending on which server is active. In order to get traffic accepted by the service a DNAT rule is used to change the destination IP. iptables -t nat -A PREROUTING -d 192.168.0.40 -p udp --dport 7100 -j DNAT --to-destination 192.168.0.20 I also wish all outbound traffic from this service to appear to come from the shared IP, so that return responses will work in the event of a active-standby failover. iptables -t nat -A POSTROUTING -p udp --sport 7100 -j SNAT --to-source 192.168.0.40 My problem is that the SNAT rule is not always run. Inbound traffic causes a connection tracking entry like this. [root]# conntrack -L -p udp udp 17 170 src=192.168.0.185 dst=192.168.0.40 sport=7100 dport=7100 src=192.168.0.20 dst=192.168.0.185 sport=7100 dport=7100 [ASSURED] mark=0 secmark=0 use=2 which means the POSTROUTING chain is not run and outbound traffic leaves with the real IP address as the source. I am thinking I can set up a NOTRACK rule in the raw table to prevent conntracking for this port number, but is there a better or more efficient way to make this work? Edit - Alternative question: Is there a way (in CentOS/Linux) to have an interface that can be bound to but not used, such that it can be attached to the network or detached when a shared IP address is swapped between servers?

    Read the article

  • Allow SFTP in iptables

    - by Kevin Orriss
    I have just purchased a VPS from linode and am going through the setup guide. I have everything running (apache2, php, mysql etc) but I am being denied access via SFTP when using fileZilla to upload a file. Now this is my second time installing the server as I missed a section out the first time. I was able to connect to my server through SFTP on filezilla the first time and the thing I missed out was adding a new user and editing the iptables in the firewall. So it would seem that the guide I have been following has blocked SFTP but allowed SSH. Here is the iptables file: *filter # Allow all loopback (lo0) traffic and drop all traffic to 127/8 that doesn't use lo0 -A INPUT -i lo -j ACCEPT -A INPUT ! -i lo -d 127.0.0.0/8 -j REJECT # Accept all established inbound connections -A INPUT -m state --state ESTABLISHED,RELATED -j ACCEPT # Allow all outbound traffic - you can modify this to only allow certain traffic -A OUTPUT -j ACCEPT # Allow HTTP and HTTPS connections from anywhere (the normal ports for websites and SSL). -A INPUT -p tcp --dport 80 -j ACCEPT -A INPUT -p tcp --dport 443 -j ACCEPT # Allow SSH connections # # The -dport number should be the same port number you set in sshd_config # -A INPUT -p tcp -m state --state NEW --dport 22 -j ACCEPT # Allow ping -A INPUT -p icmp -m icmp --icmp-type 8 -j ACCEPT # Log iptables denied calls -A INPUT -m limit --limit 5/min -j LOG --log-prefix "iptables denied: " --log-level 7 # Reject all other inbound - default deny unless explicitly allowed policy -A INPUT -j REJECT -A FORWARD -j REJECT COMMIT All I would like is a line I need to put in there which allows SFTP over port 22. Thank you for reading this.

    Read the article

  • How much free memory should I have on my webserver?

    - by neanderslob
    I have a webserver that's currently hosting two Wordpress sites and some java-based collaboration software. The server has 2G of memory and is currently using about 1.8G of the available memory. Right now what's on here is pretty much a pilot project that's getting negligible traffic so I think it's pretty clear that I'll be needing more memory. I was wondering, if I was to release it, how I might anticipate my memory needs based on the traffic it gets. I've poked around on Google and what I've found has been a bit tenuous. Is there a good heuristic that one should use when calculating memory demands as a function of the base (no traffic) load on the server? For reference, the output of free -m can be seen below: total used free shared buffers cached Mem: 2048 1832 215 0 0 0 -/+ buffers/cache: 1832 215 Swap: 0 0 0 To me this looks like actual memory used and isn't an illusion due to caching or anything else. I figure the demands of my collaboration software will have to be experimentally tested so here's free -m without that software running: total used free shared buffers cached Mem: 2048 1109 938 0 0 0 -/+ buffers/cache: 1109 938 Swap: 0 0 0 My plan B to figure this out is to add a bunch of swap space to the server, give it some traffic and adjust according the the amount that swap gets used. I was just wondering if anyone had a good rule of thumb to estimate how much memory I should plan on in advance...or if what I'm thinking is nuts. Many thanks in advance (I'm really quite new to this).

    Read the article

  • Windows Server 2003 IPSec Tunnel Connected, But Not Working (Possibly NAT/RRAS Related)

    - by Kevinoid
    Configuration I have setup a "raw" IPSec tunnel between a Windows Server 2003 (SBS) machine and a Netgear FVG318 according to the instructions in Microsoft KB816514. The configuration is as follows (using the same conventions as the article): NetA | SBS2003 | FVG318 | NetB 10.0.0.0/24 | 216.x.x.x | 69.y.y.y | 10.0.254.0/24 Both the Main Mode and Quick Mode Security Associations are successfully completed and appear in the IP Security Monitor. I am also able to ping the SBS2003 server on its private address from any computer on NetB. The Problem Any traffic sent from a computer on NetA to NetB, or from SBS2003 to NetB (excluding ICMP Ping responses), is sent out on the public network interface outside the IPSec tunnel (no encryption or header authentication, as if the tunnel were not there). Pings sent from a computer on NetB to a computer on NetA successfully reach computers on NetA, but the responses are silently discarded by SBS2003 (they do not go out in the clear and do not generate any encrypted traffic). Possible Solutions Incorrect Configuration I could have mistyped something, somewhere, or KB816514 could be incorrect in some way. I have tried very hard to eliminate the first option. Have re-created the configuration several times, tried tweaking and adjusting all the settings I could without success (most prevent the SA from being established). NAT/RRAS I have seen multiple posts elsewhere suggesting that this could be due to interaction between NAT and the IPSec filters. Possibly the NetA private addresses get rewritten to 216.x.x.x before being compared with the Quick Mode IPSec filters and don't get tunneled because of the mismatch. In fact, The Cable Guy article from June 2005 "TCP/IP Packet Processing Paths" suggests that this is the case, (see step 2 and 4 of the Transit Traffic path). If this is the case, is there a way to exclude NetA-NetB traffic from NAT? Any thoughts, ideas, suggestions, and/or comments are appreciated.

    Read the article

  • Configuring iptables rules for HAProxy and others

    - by MLister
    I have the following relevant settings for HAProxy: defaults log global mode http option httplog option dontlognull retries 3 option redispatch maxconn 500 contimeout 5s clitimeout 15s srvtimeout 15s frontend public bind *:80 option http-server-close option http-pretend-keepalive option forwardfor # ACLs ... I have three backends (including a Nginx server) configured in HAProxy, all listening on different ports of 127.0.0.1. And my iptables config is this: *filter # Allows all loopback (lo0) traffic and drop all traffic to 127/8 that doesn't use lo0 -A INPUT -i lo -j ACCEPT -A INPUT -i lo -d 127.0.0.0/8 -j REJECT # Accepts all established inbound connections -A INPUT -m state --state ESTABLISHED,RELATED -j ACCEPT # Allows all outbound traffic # You can modify this to only allow certain traffic -A OUTPUT -j ACCEPT # Allows HTTP and HTTPS connections from anywhere (the normal ports for websites) -A INPUT -p tcp --dport 80 -j ACCEPT -A INPUT -p tcp --dport 443 -j ACCEPT # Allows SSH connections # # THE -dport NUMBER IS THE SAME ONE YOU SET UP IN THE SSHD_CONFIG FILE # -A INPUT -p tcp -m state --state NEW --dport 22 -j ACCEPT # Allow ping -A INPUT -p icmp -m icmp --icmp-type 8 -j ACCEPT # log iptables denied calls -A INPUT -m limit --limit 5/min -j LOG --log-prefix "iptables denied: " --log-level 7 # Reject all other inbound - default deny unless explicitly allowed policy -A INPUT -j REJECT -A FORWARD -j REJECT COMMIT My questions are: Would the above iptables config work with the settings/options in my HAProxy config? I am also runnning a postgres and a redis server on the same machine, what settings do I need to adjust for these two to enable them work with iptables?

    Read the article

  • How do I tunnel an HTTPS proxy through a virtual machine (VMWare)

    - by Kyle
    I have a personal setup at home using VMWare Workstation. I also have a set of Virtual Private Machines that run Squid, and therefore provide me HTTPS proxy tunnels. Using Proxifier, I can tunnel all traffic for given applications through these tunnels. However, I also have a few virtual machines for dev/staging/experimentation/etc. I generally just use NAT to provide Internet access to the machines, and if I need to use these proxies, I can just setup Proxifier (or a Linux equivalent) to pipe the traffic through them. No problem. But... I got to thinking: Wouldn't it be great if I could assign these proxy tunnels to a virtual machine, so that when I start up the VM, it has instant-on access through the tunnel and not my local connection? (EDIT: Of course, it would USE my local connection, but it would tunnel traffic through the proxy.) To be more clear: I want a solution that binds the proxy to a VM, so that when I start the VM, I don't have to use a proxy client to connect to the tunnel - I am already piping all traffic from that VM through that proxy. I did a bit of searching, and the closest thing I could find was this: How to route public static IP to a virtual machine on a vmware ESXi host? Which wasn't all that applicable. The proxies are protected by user/pass but do not filter by IP. Again, they are HTTPS proxies setup through Squid. Any ideas on how to make this happen? Thanks a ton.

    Read the article

  • LDAPS being redirected to 389

    - by Ikkoras
    We're trying to perform an LDAPS bind to a server which blocks 389 with a firewall so all traffic must travel over 636. In our test lab we're connecting to a test ldap (located on the same server) which does not have this firewall so both ports are exposed. Running ldp.exe on the test server we generate the trace below which seems to suggest that it is successfully binding over 636. However if we monitor the traffic with wireshark all the traffic is being sent to 389 with no attempt to even contact 636. Other tools will bind only with SSL on 636 or without SSL on 389 whjich seems to suggest it is behaving correctly but Wireshark shows 389. Only the test server we are using RawCap to capture the local loopback traffic. Any ideas? 0x0 = ldap_unbind(ld); ld = ldap_sslinit("WIN-GF49504Q77T.test.com", 636, 1); Error 0 = ldap_set_option(hLdap, LDAP_OPT_PROTOCOL_VERSION, 3); Error 0 = ldap_connect(hLdap, NULL); Error 0 = ldap_get_option(hLdap,LDAP_OPT_SSL,(void*)&lv); Host supports SSL, SSL cipher strength = 128 bits Established connection to WIN-GF49504Q77T.test.com. Retrieving base DSA information... Getting 1 entries: Dn: (RootDSE)

    Read the article

  • Quick guide to Oracle IRM 11g: Configuring SSL

    - by Simon Thorpe
    Quick guide to Oracle IRM 11g index So far in this guide we have an IRM Server up and running, however I skipped over SSL configuration in the previous article because I wanted to focus in more detail now. You can, if you wish, not bother with setting up SSL, but considering this is a security technology it is worthwhile doing. Contents Setting up a one way, self signed SSL certificate in WebLogic Setting up an official SSL certificate in Apache 2.x Configuring Apache to proxy traffic to the IRM server There are two common scenarios in which an Oracle IRM server is configured. For a development or evaluation system, people usually communicate directly to the WebLogic Server running the IRM service. However in a production environment and for some proof of concept evaluations that require a setup reflecting a production system, the traffic to the IRM server travels via a web server proxy, commonly Apache. In this guide we are building an Oracle Enterprise Linux based IRM service and this article will go over the configuration of SSL in WebLogic and also in Apache. Like in the past articles, we are going to use two host names in the configuration below,irm.company.com will refer to the public Apache server irm.company.internal will refer to the internal WebLogic IRM server Setting up a one way, self signed SSL certificate in WebLogic First lets look at creating just a simple self signed SSL certificate to be used in WebLogic. This is a quick and easy way to get SSL working in your environment, however the downside is that no browsers are going to trust this certificate you create and you'll need to manually install the certificate onto any machine's communicating with the server. This is fine for development or when you have only a few users evaluating the system, but for any significant use it's usually better to have a fully trusted certificate in use and I explain that in the next section. But for now lets go through creating, installing and testing a self signed certificate. We use a library in Java to create the certificates, open a console and running the following commands. Note you should choose your own secure passwords whenever you see password below. [oracle@irm /] source /oracle/middleware/wlserver_10.3/server/bin/setWLSEnv.sh [oracle@irm /] cd /oracle/middleware/user_projects/domains/irm_domain/config/fmwconfig/ [oracle@irm /] java utils.CertGen -selfsigned -certfile MyOwnSelfCA.cer -keyfile MyOwnSelfKey.key -keyfilepass password -cn "irm.oracle.demo" [oracle@irm /] java utils.ImportPrivateKey -keystore MyOwnIdentityStore.jks -storepass password -keypass password -alias trustself -certfile MyOwnSelfCA.cer.pem -keyfile MyOwnSelfKey.key.pem -keyfilepass password [oracle@irm /] keytool -import -trustcacerts -alias trustself -keystore TrustMyOwnSelf.jks -file MyOwnSelfCA.cer.der -keyalg RSA We now have two Java Key Stores, MyOwnIdentityStore.jks and TrustMyOwnSelf.jks. These contain keys and certificates which we will use in WebLogic Server. Now we need to tell the IRM server to use these stores when setting up SSL connections for incoming requests. Make sure the Admin server is running and login into the WebLogic Console at http://irm.company.intranet:7001/console and do the following; In the menu on the left, select the + next to Environment to expose the submenu, then click on Servers. You will see two servers in the list, AdminServer(admin) and IRM_server1. If the IRM server is running, shut it down either by hitting CONTROL + C in the console window it was started from, or you can switch to the CONTROL tab, select IRM_server1 and then select the Shutdown menu and then Force Shutdown Now. In the Configuration tab select IRM_server1 and switch to the Keystores tab. By default WebLogic Server uses it's own demo identity and trust. We are now going to switch to the self signed one's we've just created. So select the Change button and switch to Custom Identity and Custom Trust and hit save. Now we have to complete the resulting fields, the setting's i've used in my evaluation server are below. IdentityCustom Identity Keystore: /oracle/middleware/user_projects/domains/irm_domain/config/fmwconfig/MyOwnIdentityStore.jks Custom Identity Keystore Type: JKS Custom Identity Keystore Passphrase: password Confirm Custom Identity Keystore Passphrase: password TrustCustom Trust Keystore: /oracle/middleware/user_projects/domains/irm_domain/config/fmwconfig/TrustMyOwnSelf.jks Custom Trust Keystore Type: JKS Custom Trust Keystore Passphrase: password Confirm Custom Trust Keystore Passphrase: password Now click on the SSL tab for the IRM_server1 and enter in the alias and passphrase, in my demo here the details are; IdentityPrivate Key Alias: trustself Private Key Passphrase: password Confirm Private Key Passphrase: password And hit save. Now lets test a connection to the IRM server over HTTPS using SSL. Go back to a console window and start the IRM server, a quick reminder on how to do this is... [oracle@irm /] cd /oracle/middleware/user_projects/domains/irm_domain/bin [oracle@irm /] ./startManagedWeblogic IRM_server1 Once running, open a browser and head to the SSL port of the server. By default the IRM server will be listening on the URL https://irm.company.intranet:16101/irm_rights. Note in the example image on the right the port is 7002 because it's a system that has the IRM services installed on the Admin server, this isn't typical (or advisable). Your system is going to have a separate managed server which will be listening on port 16101. Once you open this address you will notice that your browser is going to complain that the server certificate is untrusted. The images on the right show how Firefox displays this error. You are going to be prompted every time you create a new SSL session with the server, both from the browser and more annoyingly from the IRM Desktop. If you plan on always using a self signed certificate, it is worth adding it to the Windows certificate store so that when you are accessing sealed content you do not keep being informed this certificate is not trusted. Follow these instructions (which are for Internet Explorer 8, they may vary for your version of IE.) Start Internet Explorer and open the URL to your IRM server over SSL, e.g. https://irm.company.intranet:16101/irm_rights. IE will complain that about the certificate, click on Continue to this website (not recommended). From the IE Tools menu select Internet Options and from the resulting dialog select Security and then click on Trusted Sites and then the Sites button. Add to the list of trusted sites a URL which mates the server you are accessing, e.g. https://irm.company.intranet/ and select OK. Now refresh the page you were accessing and next to the URL you should see a red cross and the words Certificate Error. Click on this button and select View Certificates. You will now see a dialog with the details of the self signed certificate and the Install Certificate... button should be enabled. Click on this to start the wizard. Click next and you'll be asked where you should install the certificate. Change the option to Place all certificates in the following store. Select browse and choose the Trusted Root Certification Authorities location and hit OK. You'll then be prompted to install the certificate and answer yes. You also need to import the root signed certificate into the same location, so once again select the red Certificate Error option and this time when viewing the certificate, switch to the Certification Path tab and you should see a CertGenCAB certificate. Select this and then click on View Certificate and go through the same process as above to import the certificate into the store. Finally close all instances of the IE browser and re-access the IRM server URL again, this time you should not receive any errors. Setting up an official SSL certificate in Apache 2.x At this point we now have an IRM server that you can communicate with over SSL. However this certificate isn't trusted by any browser because it's path of trust doesn't end in a recognized certificate authority (CA). Also you are communicating directly to the WebLogic Server over a non standard SSL port, 16101. In a production environment it is common to have another device handle the initial public internet traffic and then proxy this to the WebLogic server. The diagram below shows a very simplified view of this type of deployment. What i'm going to walk through next is configuring Apache to proxy traffic to a WebLogic server and also to use a real SSL certificate from an official CA. First step is to configure Apache to handle incoming requests over SSL. In this guide I am configuring the IRM service in Oracle Enterprise Linux 5 update 3 and Apache 2.2.3 which came with OpenSSL and mod_ssl components. Before I purchase an SSL certificate, I need to generate a certificate request from the server. Oracle.com uses Verisign and for my own personal needs I use cheaper certificates from GoDaddy. The following instructions are specific to Apache, but there are many references out there for other web servers. For Apache I have OpenSSL and the commands are; [oracle@irm /] cd /usr/bin [oracle@irm bin] openssl genrsa -des3 -out irm-apache-server.key 2048 Generating RSA private key, 2048 bit long modulus ............................+++ .........+++ e is 65537 (0x10001) Enter pass phrase for irm-apache-server.key: Verifying - Enter pass phrase for irm-apache-server.key: [oracle@irm bin] openssl req -new -key irm-apache-server.key -out irm-apache-server.csr Enter pass phrase for irm-apache-server.key: You are about to be asked to enter information that will be incorporated into your certificate request. What you are about to enter is what is called a Distinguished Name or a DN. There are quite a few fields but you can leave some blank For some fields there will be a default value, If you enter '.', the field will be left blank. ----- Country Name (2 letter code) [GB]:US State or Province Name (full name) [Berkshire]:CA Locality Name (eg, city) [Newbury]:San Francisco Organization Name (eg, company) [My Company Ltd]:Oracle Organizational Unit Name (eg, section) []:Security Common Name (eg, your name or your server's hostname) []:irm.company.com Email Address []:[email protected] Please enter the following 'extra' attributes to be sent with your certificate request A challenge password []:testing An optional company name []: You must make sure to remember the pass phrase you used in the initial key generation, you will need this when later configuring Apache. In the /usr/bin directory there are now two new files. The irm-apache-server.csr contains our certificate request and is what you cut and paste, or upload, to your certificate authority when you purchase and validate your SSL certificate. In response you will typically get two files. Your server certificate and another certificate file that will likely contain a set of certificates from your CA which validate your certificate's trust. Next we need to configure Apache to use these files. Typically there is an ssl.conf file which is where all the SSL configuration is done. On my Oracle Enterprise Linux server this file is located in /etc/httpd/conf.d/ssl.conf and i've added the following lines. <VirtualHost irm.company.com> # Setup SSL for irm.company.com ServerName irm.company.com SSLEngine On SSLCertificateFile /oracle/secure/irm.company.com.crt SSLCertificateKeyFile /oracle/secure/irm.company.com.key SSLCertificateChainFile /oracle/secure/gd_bundle.crt </VirtualHost> Restarting Apache (apachectl restart) and I can now attempt to connect to the Apache server in a web browser, https://irm.company.com/. If all is configured correctly I should now see an Apache test page delivered to me over HTTPS. Configuring Apache to proxy traffic to the IRM server Final piece in setting up SSL is to have Apache proxy requests for the IRM server but do so securely. So the requests to Apache will be over HTTPS using a legitimate certificate, but we can also configure Apache to proxy these requests internally across to the IRM server using SSL with the self signed certificate we generated at the start of this article. To do this proxying we use the WebLogic Web Server plugin for Apache which you can download here from Oracle. Download the zip file and extract onto the server. The file extraction reveals a set of zip files, each one specific to a supported web server. In my instance I am using Apache 2.2 32bit on an Oracle Enterprise Linux, 64 bit server. If you are not sure what version your Apache server is, run the command /usr/sbin/httpd -V and you'll see version and it its 32 or 64 bit. Mine is a 32bit server so I need to extract the file WLSPlugin1.1-Apache2.2-linux32-x86.zip. The from the resulting lib folder copy the file mod_wl.so into /usr/lib/httpd/modules/. First we want to test that the plug in will work for regular HTTP traffic. Edit the httpd.conf for Apache and add the following section at the bottom. LoadModule weblogic_module modules/mod_wl.so <IfModule mod_weblogic.c>    WebLogicHost irm.company.internal    WebLogicPort 16100    WLLogFile /tmp/wl-proxy.log </IfModule> <Location /irm_rights>    SetHandler weblogic-handler </Location> <Location /irm_desktop>    SetHandler weblogic-handler </Location> <Location /irm_sealing>    SetHandler weblogic-handler </Location> <Location /irm_services>    SetHandler weblogic-handler </Location> Now restart Apache again (apachectl restart) and now open a browser to http://irm.company.com/irm_rights. Apache will proxy the HTTP traffic from the port 80 of your Apache server to the IRM service listening on port 16100 of the WebLogic Managed server. Note above I have included all four of the Locations you might wish to proxy. http://irm.company.internalirm_rights is the URL to the management website, /irm_desktop is the URL used for the IRM Desktop to communicate. irm_sealing is for web services based document sealing and irm_services is for IRM server web services. The last two are typically only used when you have the IRM server integrated with another application and it is unlikely you'd be accessing these resources from the public facing Apache server. However, just in case, i've mentioned them above. Now let's enable SSL communication from Apache to WebLogic. In the ZIP file we extracted were some more modules we need to copy into the Apache folder. Looking back in the lib that we extracted, there are some more files. Copy the following into the /usr/lib/httpd/modules/ folder. libwlssl.so libnnz11.so libclntsh.so.11.1 Now the documentation states that should only need to do this, but I found that I also needed to create an environment variable called LD_LIBRARY_PATH and point this to the folder /usr/lib/httpd/modules/. If I didn't do this, starting Apache with the WebLogic module configured to SSL would throw the error. [crit] (20014)Internal error: WL SSL Init failed for server: (null) on 0 So I had to edit the file /etc/profile and add the following lines at the bottom. You may already have the LD_LIBRARY_PATH variable defined, therefore simply add this path to it. LD_LIBRARY_PATH=/usr/lib/httpd/modules/ export LD_LIBRARY_PATH Now the WebLogic plug in uses an Oracle Wallet to store the required certificates.You'll need to copy the self signed certificate from the IRM server over to the Apache server. Copy over the MyOwnSelfCA.cer.der into the same folder where you are storing your public certificates, in my example this is /oracle/secure. It's worth mentioning these files should ONLY be readable by root (the user Apache runs as). Now lets create an Oracle Wallet and import the self signed certificate from the IRM server. The file orapki was included in the bin folder of the Apache 1.1 plugin zip you extracted. orapki wallet create -wallet /oracle/secure/my-wallet -auto_login_only orapki wallet add -wallet /oracle/secure/my-wallet -trusted_cert -cert MyOwnSelfCA.cer.der -auto_login_only Finally change the httpd.conf to reflect that we want the WebLogic Apache plug-in to use HTTPS/SSL and not just plain HTTP. <IfModule mod_weblogic.c>    WebLogicHost irm.company.internal    WebLogicPort 16101    SecureProxy ON    WLSSLWallet /oracle/secure/my-wallet    WLLogFile /tmp/wl-proxy.log </IfModule> Then restart Apache once more and you can go back to the browser to test the communication. Opening the URL https://irm.company.com/irm_rights will proxy your request to the WebLogic server at https://irm.company.internal:16101/irm_rights. At this point you have a fully functional Oracle IRM service, the next step is to create a sealed document and test the entire system.

    Read the article

  • Connection Pooling is Busted

    - by MightyZot
    A few weeks ago we started getting complaints about performance in an application that has performed very well for many years.  The application is a n-tier application that uses ADODB with the SQLOLEDB provider to talk to a SQL Server database.  Our object model is written in such a way that each public method validates security before performing requested actions, so there is a significant number of queries executed to get information about file cabinets, retrieve images, create workflows, etc.  (PaperWise is a document management and workflow system.)  A common factor for these customers is that they have remote offices connected via MPLS networks. Naturally, the first thing we looked at was the query performance in SQL Profiler.  All of the queries were executing within expected timeframes, most of them were so fast that the duration in SQL Profiler was zero.  After getting nowhere with SQL Profiler, the situation was escalated to me.  I decided to take a peek with Process Monitor.  Procmon revealed some “gaps” in the TCP/IP traffic.  There were notable delays between send and receive pairs.  The send and receive pairs themselves were quite snappy, but quite often there was a notable delay between a receive and the next send.  You might expect some delay because, presumably, the application is doing some thinking in-between the pairs.  But, comparing the procmon data at the remote locations with the procmon data for workstations on the local network showed that the remote workstations were significantly delayed.  Procmon also showed a high number of disconnects. Wireshark traces showed that connections to the database were taking between 75ms and 150ms.  Not only that, but connections to a file share containing images were taking 2 seconds!  So, I asked about a trust.  Sure enough there was a trust between two domains and the file share was on the second domain.  Joining a remote workstation to the domain hosting the share containing images alleviated the time delay in accessing the file share.  Removing the trust had no affect on the connections to the database. Microsoft Network Monitor includes filters that parse TDS packets.  TDS is the protocol that SQL Server uses to communicate.  There is a certificate exchange and some SSL that occurs during authentication.  All of this was evident in the network traffic.  After staring at the network traffic for a while, and examining packets, I decided to call it a night.  On the way home that night, something about the traffic kept nagging at me.  Then it dawned on me…at the beginning of the dance of packets between the client and the server all was well.  Connection pooling was working and I could see multiple queries getting executed on the same connection and ethereal port.  After a particular query, connecting to two different servers, I noticed that ADODB and SQLOLEDB started making repeated connections to the database on different ethereal ports.  SQL Server would execute a single query and respond on a port, then open a new port and execute the next query.  Connection pooling appeared to be broken. The next morning I wrote a test to confirm my hypothesis.  Turns out that the sequence causing the connection nastiness goes something like this: Make a connection to the database. Open a result set that returns enough records to require multiple roundtrips to the server. For each result, query for some other data in the database (this will open a new implicit connection.) Close the inner result set and repeat for every item in the original result set. Close the original connection. Provided that the first result set returns enough data to require multiple roundtrips to the server, ADODB and SQLOLEDB will start making new connections to the database for each query executed in the loop.  Originally, I thought this might be due to Microsoft’s denial of service (ddos) attack protection.  After turning those features off to no avail, I eventually thought to switch my queries to client-side cursors instead of server-side cursors.  Server-side cursors are the default, by the way.  Voila!  After switching to client-side cursors, the disconnects were gone and the above sequence yielded two connections as expected. While the real problem is the amount of time it takes to make connections over these MPLS networks (100ms on average), switching to client-side cursors made the problem go away.  Believe it or not, this is actually documented by Microsoft, and rather difficult to find.  (At least it was while we were trying to troubleshoot the problem!)  So, if you’re noticing performance issues on slower networks, or networks with slower switching, take a look at the traffic in a tool like Microsoft Network Monitor.  If you notice a high number of disconnects, and you’re using fire-hose or server-side cursors, then try switching to client-side cursors and you may see the problem go away. Most likely, Microsoft believes this to be appropriate behavior, because ADODB can’t guarantee that all of the data has been retrieved when you execute the inner queries.  I’m not convinced, though, because the problem remains even after replacing all of the implicit connections with explicit connections and closing those connections in-between each of the inner queries.  In that case, there doesn’t seem to be a reason why ADODB can’t use a single connection from the connection pool to make the additional queries, bringing the total number of connections to two.  Instead ADO appears to make an assumption about the state of the connection. I’ve reported the behavior to Microsoft and am awaiting to hear from the appropriate team, so that I can demonstrate the problem.  Maybe they can explain to us why this is appropriate behavior.  :)

    Read the article

  • OpenVPN - Windows 8 to Windows 2008 Server, not connecting

    - by niico
    I have followed this tutorial about setting up an OpenVPN Server on Windows Server - and a client on Windows (in this case Windows 8). The server appears to be running fine - but it is not connecting with this error: Mon Jul 22 19:09:04 2013 Warning: cannot open --log file: C:\Program Files\OpenVPN\log\my-laptop.log: Access is denied. (errno=5) Mon Jul 22 19:09:04 2013 OpenVPN 2.3.2 x86_64-w64-mingw32 [SSL (OpenSSL)] [LZO] [PKCS11] [eurephia] [IPv6] built on Jun 3 2013 Mon Jul 22 19:09:04 2013 MANAGEMENT: TCP Socket listening on [AF_INET]127.0.0.1:25340 Mon Jul 22 19:09:04 2013 Need hold release from management interface, waiting... Mon Jul 22 19:09:05 2013 MANAGEMENT: Client connected from [AF_INET]127.0.0.1:25340 Mon Jul 22 19:09:05 2013 MANAGEMENT: CMD 'state on' Mon Jul 22 19:09:05 2013 MANAGEMENT: CMD 'log all on' Mon Jul 22 19:09:05 2013 MANAGEMENT: CMD 'hold off' Mon Jul 22 19:09:05 2013 MANAGEMENT: CMD 'hold release' Mon Jul 22 19:09:05 2013 Socket Buffers: R=[65536->65536] S=[65536->65536] Mon Jul 22 19:09:05 2013 UDPv4 link local: [undef] Mon Jul 22 19:09:05 2013 UDPv4 link remote: [AF_INET]66.666.66.666:9999 Mon Jul 22 19:09:05 2013 MANAGEMENT: >STATE:1374494945,WAIT,,, Mon Jul 22 19:10:05 2013 TLS Error: TLS key negotiation failed to occur within 60 seconds (check your network connectivity) Mon Jul 22 19:10:05 2013 TLS Error: TLS handshake failed Mon Jul 22 19:10:05 2013 SIGUSR1[soft,tls-error] received, process restarting Mon Jul 22 19:10:05 2013 MANAGEMENT: >STATE:1374495005,RECONNECTING,tls-error,, Mon Jul 22 19:10:05 2013 Restart pause, 2 second(s) Note I have changed the IP and port no (it uses a non-standard port for security reasons). That port is open on the hardware firewall. The server logs are showing a connection attempt from my client: TLS: Initial packet from [AF_INET]118.68.xx.xx:65011, sid=081af4ed xxxxxxxx Mon Jul 22 14:19:15 2013 118.68.xx.xx:65011 TLS Error: TLS key negotiation failed to occur within 60 seconds (check your network connectivity) How can I problem solve this & find the problem? Thx Update - Client config file: ############################################## # Sample client-side OpenVPN 2.0 config file # # for connecting to multi-client server. # # # # This configuration can be used by multiple # # clients, however each client should have # # its own cert and key files. # # # # On Windows, you might want to rename this # # file so it has a .ovpn extension # ############################################## # Specify that we are a client and that we # will be pulling certain config file directives # from the server. client # Use the same setting as you are using on # the server. # On most systems, the VPN will not function # unless you partially or fully disable # the firewall for the TUN/TAP interface. ;dev tap dev tun # Windows needs the TAP-Win32 adapter name # from the Network Connections panel # if you have more than one. On XP SP2, # you may need to disable the firewall # for the TAP adapter. ;dev-node MyTap # Are we connecting to a TCP or # UDP server? Use the same setting as # on the server. ;proto tcp proto udp # The hostname/IP and port of the server. # You can have multiple remote entries # to load balance between the servers. remote 00.00.00.00 1194 ;remote 00.00.00.00 9999 ;remote my-server-2 1194 # Choose a random host from the remote # list for load-balancing. Otherwise # try hosts in the order specified. ;remote-random # Keep trying indefinitely to resolve the # host name of the OpenVPN server. Very useful # on machines which are not permanently connected # to the internet such as laptops. resolv-retry infinite # Most clients don't need to bind to # a specific local port number. nobind # Downgrade privileges after initialization (non-Windows only) ;user nobody ;group nobody # Try to preserve some state across restarts. persist-key persist-tun # If you are connecting through an # HTTP proxy to reach the actual OpenVPN # server, put the proxy server/IP and # port number here. See the man page # if your proxy server requires # authentication. ;http-proxy-retry # retry on connection failures ;http-proxy [proxy server] [proxy port #] # Wireless networks often produce a lot # of duplicate packets. Set this flag # to silence duplicate packet warnings. ;mute-replay-warnings # SSL/TLS parms. # See the server config file for more # description. It's best to use # a separate .crt/.key file pair # for each client. A single ca # file can be used for all clients. ca "C:\\Program Files\\OpenVPN\\config\\ca.crt" cert "C:\\Program Files\\OpenVPN\\config\\my-laptop.crt" key "C:\\Program Files\\OpenVPN\\config\\my-laptop.key" # Verify server certificate by checking # that the certicate has the nsCertType # field set to "server". This is an # important precaution to protect against # a potential attack discussed here: # http://openvpn.net/howto.html#mitm # # To use this feature, you will need to generate # your server certificates with the nsCertType # field set to "server". The build-key-server # script in the easy-rsa folder will do this. ns-cert-type server # If a tls-auth key is used on the server # then every client must also have the key. ;tls-auth ta.key 1 # Select a cryptographic cipher. # If the cipher option is used on the server # then you must also specify it here. ;cipher x # Enable compression on the VPN link. # Don't enable this unless it is also # enabled in the server config file. comp-lzo # Set log file verbosity. verb 3 # Silence repeating messages ;mute 20 Server config file: ################################################# # Sample OpenVPN 2.0 config file for # # multi-client server. # # # # This file is for the server side # # of a many-clients <-> one-server # # OpenVPN configuration. # # # # OpenVPN also supports # # single-machine <-> single-machine # # configurations (See the Examples page # # on the web site for more info). # # # # This config should work on Windows # # or Linux/BSD systems. Remember on # # Windows to quote pathnames and use # # double backslashes, e.g.: # # "C:\\Program Files\\OpenVPN\\config\\foo.key" # # # # Comments are preceded with '#' or ';' # ################################################# # Which local IP address should OpenVPN # listen on? (optional) ;local 00.00.00.00 # Which TCP/UDP port should OpenVPN listen on? # If you want to run multiple OpenVPN instances # on the same machine, use a different port # number for each one. You will need to # open up this port on your firewall. std 1194 port 1194 # TCP or UDP server? ;proto tcp proto udp # "dev tun" will create a routed IP tunnel, # "dev tap" will create an ethernet tunnel. # Use "dev tap0" if you are ethernet bridging # and have precreated a tap0 virtual interface # and bridged it with your ethernet interface. # If you want to control access policies # over the VPN, you must create firewall # rules for the the TUN/TAP interface. # On non-Windows systems, you can give # an explicit unit number, such as tun0. # On Windows, use "dev-node" for this. # On most systems, the VPN will not function # unless you partially or fully disable # the firewall for the TUN/TAP interface. ;dev tap dev tun # Windows needs the TAP-Win32 adapter name # from the Network Connections panel if you # have more than one. On XP SP2 or higher, # you may need to selectively disable the # Windows firewall for the TAP adapter. # Non-Windows systems usually don't need this. ;dev-node MyTap # SSL/TLS root certificate (ca), certificate # (cert), and private key (key). Each client # and the server must have their own cert and # key file. The server and all clients will # use the same ca file. # # See the "easy-rsa" directory for a series # of scripts for generating RSA certificates # and private keys. Remember to use # a unique Common Name for the server # and each of the client certificates. # # Any X509 key management system can be used. # OpenVPN can also use a PKCS #12 formatted key file # (see "pkcs12" directive in man page). ca "C:\\Program Files\\OpenVPN\\config\\ca.crt" cert "C:\\Program Files\\OpenVPN\\config\\server.crt" key "C:\\Program Files\\OpenVPN\\config\\server.key" # Diffie hellman parameters. # Generate your own with: # openssl dhparam -out dh1024.pem 1024 # Substitute 2048 for 1024 if you are using # 2048 bit keys. dh "C:\\Program Files\\OpenVPN\\config\\dh2048.pem" # Configure server mode and supply a VPN subnet # for OpenVPN to draw client addresses from. # The server will take 10.8.0.1 for itself, # the rest will be made available to clients. # Each client will be able to reach the server # on 10.8.0.1. Comment this line out if you are # ethernet bridging. See the man page for more info. server 10.8.0.0 255.255.255.0 # Maintain a record of client <-> virtual IP address # associations in this file. If OpenVPN goes down or # is restarted, reconnecting clients can be assigned # the same virtual IP address from the pool that was # previously assigned. ifconfig-pool-persist ipp.txt # Configure server mode for ethernet bridging. # You must first use your OS's bridging capability # to bridge the TAP interface with the ethernet # NIC interface. Then you must manually set the # IP/netmask on the bridge interface, here we # assume 10.8.0.4/255.255.255.0. Finally we # must set aside an IP range in this subnet # (start=10.8.0.50 end=10.8.0.100) to allocate # to connecting clients. Leave this line commented # out unless you are ethernet bridging. ;server-bridge 10.8.0.4 255.255.255.0 10.8.0.50 10.8.0.100 # Configure server mode for ethernet bridging # using a DHCP-proxy, where clients talk # to the OpenVPN server-side DHCP server # to receive their IP address allocation # and DNS server addresses. You must first use # your OS's bridging capability to bridge the TAP # interface with the ethernet NIC interface. # Note: this mode only works on clients (such as # Windows), where the client-side TAP adapter is # bound to a DHCP client. ;server-bridge # Push routes to the client to allow it # to reach other private subnets behind # the server. Remember that these # private subnets will also need # to know to route the OpenVPN client # address pool (10.8.0.0/255.255.255.0) # back to the OpenVPN server. ;push "route 192.168.10.0 255.255.255.0" ;push "route 192.168.20.0 255.255.255.0" # To assign specific IP addresses to specific # clients or if a connecting client has a private # subnet behind it that should also have VPN access, # use the subdirectory "ccd" for client-specific # configuration files (see man page for more info). # EXAMPLE: Suppose the client # having the certificate common name "Thelonious" # also has a small subnet behind his connecting # machine, such as 192.168.40.128/255.255.255.248. # First, uncomment out these lines: ;client-config-dir ccd ;route 192.168.40.128 255.255.255.248 # Then create a file ccd/Thelonious with this line: # iroute 192.168.40.128 255.255.255.248 # This will allow Thelonious' private subnet to # access the VPN. This example will only work # if you are routing, not bridging, i.e. you are # using "dev tun" and "server" directives. # EXAMPLE: Suppose you want to give # Thelonious a fixed VPN IP address of 10.9.0.1. # First uncomment out these lines: ;client-config-dir ccd ;route 10.9.0.0 255.255.255.252 # Then add this line to ccd/Thelonious: # ifconfig-push 10.9.0.1 10.9.0.2 # Suppose that you want to enable different # firewall access policies for different groups # of clients. There are two methods: # (1) Run multiple OpenVPN daemons, one for each # group, and firewall the TUN/TAP interface # for each group/daemon appropriately. # (2) (Advanced) Create a script to dynamically # modify the firewall in response to access # from different clients. See man # page for more info on learn-address script. ;learn-address ./script # If enabled, this directive will configure # all clients to redirect their default # network gateway through the VPN, causing # all IP traffic such as web browsing and # and DNS lookups to go through the VPN # (The OpenVPN server machine may need to NAT # or bridge the TUN/TAP interface to the internet # in order for this to work properly). ;push "redirect-gateway def1 bypass-dhcp" # Certain Windows-specific network settings # can be pushed to clients, such as DNS # or WINS server addresses. CAVEAT: # http://openvpn.net/faq.html#dhcpcaveats # The addresses below refer to the public # DNS servers provided by opendns.com. ;push "dhcp-option DNS 208.67.222.222" ;push "dhcp-option DNS 208.67.220.220" # Uncomment this directive to allow differenta # clients to be able to "see" each other. # By default, clients will only see the server. # To force clients to only see the server, you # will also need to appropriately firewall the # server's TUN/TAP interface. ;client-to-client # Uncomment this directive if multiple clients # might connect with the same certificate/key # files or common names. This is recommended # only for testing purposes. For production use, # each client should have its own certificate/key # pair. # # IF YOU HAVE NOT GENERATED INDIVIDUAL # CERTIFICATE/KEY PAIRS FOR EACH CLIENT, # EACH HAVING ITS OWN UNIQUE "COMMON NAME", # UNCOMMENT THIS LINE OUT. ;duplicate-cn # The keepalive directive causes ping-like # messages to be sent back and forth over # the link so that each side knows when # the other side has gone down. # Ping every 10 seconds, assume that remote # peer is down if no ping received during # a 120 second time period. keepalive 10 120 # For extra security beyond that provided # by SSL/TLS, create an "HMAC firewall" # to help block DoS attacks and UDP port flooding. # # Generate with: # openvpn --genkey --secret ta.key # # The server and each client must have # a copy of this key. # The second parameter should be '0' # on the server and '1' on the clients. ;tls-auth ta.key 0 # This file is secret # Select a cryptographic cipher. # This config item must be copied to # the client config file as well. ;cipher BF-CBC # Blowfish (default) ;cipher AES-128-CBC # AES ;cipher DES-EDE3-CBC # Triple-DES # Enable compression on the VPN link. # If you enable it here, you must also # enable it in the client config file. comp-lzo # The maximum number of concurrently connected # clients we want to allow. ;max-clients 100 # It's a good idea to reduce the OpenVPN # daemon's privileges after initialization. # # You can uncomment this out on # non-Windows systems. ;user nobody ;group nobody # The persist options will try to avoid # accessing certain resources on restart # that may no longer be accessible because # of the privilege downgrade. persist-key persist-tun # Output a short status file showing # current connections, truncated # and rewritten every minute. status openvpn-status.log # By default, log messages will go to the syslog (or # on Windows, if running as a service, they will go to # the "\Program Files\OpenVPN\log" directory). # Use log or log-append to override this default. # "log" will truncate the log file on OpenVPN startup, # while "log-append" will append to it. Use one # or the other (but not both). ;log openvpn.log ;log-append openvpn.log # Set the appropriate level of log # file verbosity. # # 0 is silent, except for fatal errors # 4 is reasonable for general usage # 5 and 6 can help to debug connection problems # 9 is extremely verbose verb 3 # Silence repeating messages. At most 20 # sequential messages of the same message # category will be output to the log. ;mute 20 I have changed IP's for security

    Read the article

< Previous Page | 227 228 229 230 231 232 233 234 235 236 237 238  | Next Page >