Search Results

Search found 18191 results on 728 pages for 'single board'.

Page 657/728 | < Previous Page | 653 654 655 656 657 658 659 660 661 662 663 664  | Next Page >

  • How to rate-limit concurrent sessions with nginx or haproxy?

    - by bantic
    I'm currently using nginx to reverse-proxy requests from web clients that are doing long-polling to an upstream. Since we're doing long polling (as opposed to websockets), when a client connects it will make multiple http connections to the server in serial, re-establishing a connection every time the server sends it some data (or timing out and re-establishing if the server has nothing to say for 10 seconds). What I'd like to do is limit the number of concurrent web clients. Since the clients are constantly making new HTTP requests instead of keeping a single request open, it's a little tricky to count the total number of web clients (because it's not the same as total number of concurrently connected http clients). The method I've come up with is to track http requests by the originating IP address, and store the IP address somewhere with a TTL of 20 seconds. If a request comes in whose IP isn't recognized, then we check the total number of unexpired stored IP addresses; if that's less than the maximum then we allow this request through. And if a request comes in with an IP address that we can find in the look-up table that hasn't yet expired, then it is allowed through as well. All requests that are allowed through have their IPs added to the table (if not there before) and the TTL refreshed to 20 seconds again. I had actually whipped something together that worked correctly this way using nginx along with the Redis 2.0 Nginx Module (and the nginx lua module to simplify the conditional branching), using redis to store my IP addresses with a TTL (the SETEX command), and checking the table size with the DBSIZE command. This worked but the performance was horrible. nginx and redis ended up using lots of cpu and the machine could only handle a very small number of concurrent requests. The new stick-table and tracking counters that were added to Haproxy in version 1.5 (via a commission from serverfault) seem like they might be ideal to implement exactly this sort of rate limiting, because the stick-table can track IP addresses and automatically expire entries. However, I don't see an easy way to get a total count of the unexpired entries in the stick table, which would be necessary to know the number of connected web clients. I'm curious if anyone has any suggestions, for nginx or haproxy or even for something else not mentioned here that I haven't thought of yet.

    Read the article

  • How to tune system settings for mongoDB on Linux?

    - by jsh
    Trying to squeeze a lot out of one question here -- please bear with me. Although the MongoDB man pages make several useful recommendations about system settings like ulimit (http://docs.mongodb.org/manual/reference/ulimit/), and other production factors (http://docs.mongodb.org/manual/administration/production-notes/) they seem mysteriously silent on things like virtual memory and swap settings. The closest we get to a hint is that "...the operating system’s virtual memory subsystem manages MongoDB’s memory..." (http://docs.mongodb.org/manual/faq/fundamentals/#does-mongodb-require-a-lot-of-ram). Running the same job - high writes and high reads on about 10,000,000 records in a single collection -- on my 4-processor, 4GB RAM macbook and an 8-core ubuntu box with 64GB RAM I saw dramatically WORSE read performance on the linux box with factory settings, and could hear the disk constantly spinning, indicating high I/O and presumably swapping. Yes, other things were happening on the box, but there was plenty of free RAM, disk space, etc.; furthermore, I did not see evidence that Mongo was expanding to take advantage of all that free RAM as it is touted to do. Linux box default settings were as follows: vm.swappiness =60 vm.dirty_background_ratio = 10 vm.dirty_ratio = 20 vm.dirty_expire_centisecs =3000 vm.dirty_writeback_centisecs=500 I hazarded some guesses looking at docs and blogs for other types of databases (Oracle, MYSQL, etc.), experimented, and adjusted as below. vm.swappiness=10 vm.dirty_background_ratio=5 vm.dirty_ratio=5 vm.dirty_writeback_centisecs=250 vm.dirty_expire_centisecs=500 I saw some immediate apparent improvements in read time. However, when I ran my test jobs again, read performance continued to be painfully sluggish during heavy writes. Then, I REBUILT the collection from an available data source - and suddenly I can read at 1ms or less per record WHILE doing the write job! So the question is really two-fold: 1) What are appropriate VM settings for MongoDB on Linux? 2) (bonus) Does Mongo do some checking or optimization with the OS while data is being built? In other words, if I have built a large data set with suboptimal VM or I/O settings, does Mongo make assumptions during the memory-mapping process that will fail to take advantage of optimizations down the road? Obviously I don't fully grok memory mapping under the hood (I was hoping I wouldn't have to). Any help appreciated...thanks! -j

    Read the article

  • Creating a network link between 2 very close buildings

    - by Daniel Johnson
    I have a charity who have two adjacent medium sized modern detached houses (in the UK): the buildings stand next to each other and are less than 5 metres apart. They have DSL connected to a single computer in one of the buildings. They want to add a network with wireless, and want it to work across both buildings. Being a charity they need to keep costs down. The network would be used for sharing Word documents, e-mail, browsing and skyping. My initial thoughts were to connect the buildings with fibre. So: Option 1 Use fibre between the buildings. Sufficient cable and two TP-LINK MC100CM Fast Ethernet Media Converters. Cost ~£80.00. But there is the extra cost and hassle of running the cable down and up the external walls, lifting and relaying paving, and burying underground. Never having fitted fibre I'm also a little worried about going up the wall and then bending the cable at 90 degrees to go through the wall and into the building. Option 2 Use two TP-Link TL-WA7510N High Powered Outdoor 5Ghz 15dBi Wireless antennas to connect the buildings. There is a clear line of sight at first floor level. Cost ~£100. And much easier to fit than fibre! Is using the TL-WA7510Ns overkill? Is there something more suitable? I had hoped to use some Netgear stuff, e.g. two DGN2200, one in each house and also use them to provide the wireless link between the buildings. However, in bridge mode wireless client association is not available and repeater mode with client association only supports WEP security which isn't strong enough. Is there something similar that would be up to the job? Option 3 Connect the buildings with UTP cable. My concerns here are risk of electric shock due to a difference of potential between the buildings (or are they so close this shouldn't be an issue) and protection from lightning strikes. Is fitting lighting arrestors expensive? And what can be done to ameliorate against the risk of shock? This all falls outside my area of expertise so I would really appreciate some advice.

    Read the article

  • understanding my site's DNS records

    - by DaveM
    firstly apologies for using the word 'pointage' this is the word my french domain registrar uses so I may have used to wrong term. OK I would like to better understand what is going on on my 'pointage' record on my domain registrars site. for my (currently empty) web site it reports the following details... Type : Host : Destination A : www.mydomain.org : 62.210.176.146 A : mail.mydomain.org : 84.246.225.176 Mx : .mydomain.org : mail.mydomain.org I think I understand the MX record, that simply relays anything onto the mail.mydomain.org location. However why are the destination for the www and mail domains different. Even more confusing (for me) is the fact that if I ping either of www.mydomain.org or mail.mydomain.org the ping returns a different IP address. This IP address is consistent with that of my server (ie 92.39.247.92). So what exactly is going on ? I'm sure I could find the information on the web,I've read a few thing on the debianhelp site regarding DNS records, and it seems to suggest that the record should be a reverse lookup, but certains isn't the reverse of my servers IP ? but I don't what I should be looking for, so links to docs and search terms for google will be happily accepteed (even though they go against the grain of SO answers to question). thanks in advance. David. ps. I should add that everything seems to work just fine, and I've just descovered this part of the management page of my registrar. Edit: Addition of DNS records and ping results. The DNS record for the site. From what I've read there should only realy be a single 'A' record, so has something gone wrong ? should I change it (remove the extras and then just point www.facilitee.org - .facilitee.org and mail.facilitee.org - .facilitee.org here is the DNS record A www.facilitee.org ? 92.39.247.92 A .facilitee.org ? 92.39.247.92 A mail.facilitee.org ? 92.39.247.92 A webmail.facilitee.org ? 92.39.247.92 MX .facilitee.org ? mail.facilitee.org ping results... ~$ ping www.facilitee.org PING www.facilitee.org (92.39.247.92) 56(84) bytes of data. 64 bytes from vps4576-cloud.dns26.com (92.39.247.92): ~$ ping mail.facilitee.org PING mail.facilitee.org (92.39.247.92) 56(84) bytes of data. 64 bytes from vps4576-cloud.dns26.com (92.39.247.92): So the DNS and the ping correspond, but the 'pointage' doesn't. ~ how can I get a report of the pointage records other than from my registrar ?

    Read the article

  • Squid on windows loadbalancing only to one server

    - by Martin L.
    After thousands of googles and trying days i cant get the load balancer/failover in squid on windows to work. Iam using squid 2.7. My webservers are 2 single NIC lighttpd and one dual nic lighttpd. server1 in this example is running squid on port 80 and lighttpd on port 8080 (just to test) Requirements: All 3 webservers running lighttpd should be balanced two option for load balancing: Best would be if server1 is busy server2 takes over, if server2 is busy server3 takes over, etc.. Round robin style evenly distributed load. Eg server1 takes first call, server2 second etc.. All requests should be treated the same way (no url rewriting or so on) Sent host headers have to be redirected to every server as http host header, speaking of "server1", "server1.company.internal" and "10.211.1.1". My approach: acl all src all acl manager proto cache_object http_port 80 accel defaultsite=server1.company.internal vhost #reverse proxy entries cache_peer 10.211.2.1 parent 8080 0 no-query originserver round-robin login=PASS name=server1_nic1 cache_peer 10.211.1.2 parent 80 0 no-query originserver round-robin login=PASS name=server2_nic1 cache_peer 10.211.2.3 parent 8080 0 no-query originserver round-robin login=PASS name=server3_nic1 cache_peer 10.211.2.4 parent 8080 0 no-query originserver round-robin login=PASS name=server3_nic2 #decl of names of squid host acl registered_name_hostdomain dstdomain server1.company.internal acl registered_name_host dstdomain server1 #ip of squid host acl registered_name_ip dstdomain 10.211.2.1 # access: redirects the correct squid hostname http_access allow registered_name_hostdomain http_access allow registered_name_host http_access allow registered_name_ip http_access deny all cache_peer_access server1_nic1 allow registered_name_hostdomain cache_peer_access server1_nic1 allow registered_name_host cache_peer_access server1_nic1 allow registered_name_ip cache_peer_access server2_nic1 allow registered_name_hostdomain cache_peer_access server2_nic1 allow registered_name_host cache_peer_access server2_nic1 allow registered_name_ip cache_peer_access server3_nic1 allow registered_name_hostdomain cache_peer_access server3_nic1 allow registered_name_host cache_peer_access server3_nic1 allow registered_name_ip cache_peer_access server3_nic2 allow registered_name_hostdomain cache_peer_access server3_nic2 allow registered_name_host cache_peer_access server3_nic2 allow registered_name_ip cache_peer_access server1_nic1 deny all cache_peer_access server2_nic1 deny all cache_peer_access server3_nic1 deny all cache_peer_access server3_nic2 deny all never_direct allow all Problems: Load balancer does not load balance other than to first server. Only if the first server is killed in any way the second will take over. I have seen the others working at some point, but definitely not as the intended load balancing described above. If the cache_peer_access is not defined sometimes the wrong hostname is sent to the backend webserver and this always depends on the defaultsite= parameter. Probably because the host header on the request to squid is not set and its replaced by defaultsite. Leaving out defaultsite didnt solve the problem. The only workaround i found for this is the current approach with cache_peer_access. Questions: Does the cache_peer_access influence the round-robin? Is there a better workaround to pass the host header to the backed webservers? Which parameters do increase the speed of load balancing or does anyone have a better approach? -Martin

    Read the article

  • One user sometimes gets an unknown certificate error opening Outlook

    - by Chris
    Let me clarify a little. This isn't an unknown certificate error it's an unknown certificate error in so much as I can't figure out where the certificate comes from. This happens on a Win 7 Enterprise machine connecting to Exchange 2010 with Outlook 2010. The error he gets is that the root is not trusted because it's a self-signed cert. Take a look at this screenshot because even if I had generated this myself I wouldn't have put "SomeOrganizationalUnit" or "SomeCity" or "SomeState", etc. (Red block covers our domain name.) I'm a little concerned this is a symptom of a security breach. Exchange 2010 has three certificates installed but none of them are this certificate. They all have different expiration dates (one is expired) and different meta-data. edit: There are two scenarios that I see the certificate warning and one of them I can reliably repeat. When the user leaves his computer on over night Outlook pops the Security Warning window. I don't know what time this happens. Using Outlook Anywhere if I connect to Exchange externally via a cellular USB modem the Security Warning window will appear every time I close and reopen Outlook. Whether I say Yes or No does not make a difference on whether or not I can connect to Exchange and send/receive email. In other words, I can always connect to Exchange. I've checked my two Exchange servers and my Cisco router for a certificate that matches this one and I can't find it. edit 2: Here is a screenshot of the Security Alert window. (I've been calling it Security Warning... My mistake.) edit 3: I stopped seeing this error several weeks ago but I can't tie it to any single event (because I just sort of realized that warning had stopped showing up) but I think I found the source of the certificate. Last week I found out that the certificate on our website DomainA.com was invalid. I knew that our web admin had installed a valid certificate so when I look into the problem I found out I was being presented with the invalid certificate that this posting is in regards to. The Exchange server's domain is mail.DomainA.com so I can only guess that Outlook was passing this invalid certificate through as it did some kind of check on DomainA.com. This issue is still a mystery because the certificate warning stopped appearing several weeks ago whereas the invalid certificate issue on the website was only fixed last week. It ended up being a problem with the website control panel. The valid certificate was installed but not being served for some reason and instead the self-signed cert was being served.

    Read the article

  • System failure - need diagnostic recommendation

    - by Ladislav Mrnka
    I have big problem with my computer. Configuration is: Intel i7 + 6x2GB OCZ DDR3 Motheboard: Asus P6T Deluxe V2, HDD controller configured to AHCI Main drive: OCZ Vertex 2 (SSD) - contains all installed programs and system Second drive: Samsung SpinPoint - contains User profiles, ProgramData, virtual machines and databases Third drive: Samsung SpinPoint - data drive + backups OS: Windows 7 Ultimate x64 I have never had any problem with this computer until now. During weekend my computer completely crashed without any reason. Each time I tried to boot to Windows I got BSOD with message BAD_SYSTEM_CONFIG_INFO and automatic restart (I didn't install any new SW or HW). But after restart main OCZ drive was disconnected (not detected by BIOS). When I turned off and on computer, the drive was again connected. It also happend every single time I tried to repair installation somehow. It ended with some error and after restart drive was disconnected. The only thing which worked was format + fresh install. After installing almost everything I wanted to install Visual Studio 2010 Ultimate (complete installation without SQL Server Express). During installation of VS itself I always get BSOD - it is too fast so I'm not able to read description. After restart it searches for all disk drives for really long time and sometimes it changes boot drive so the system is not able to start - Bootmgr not found. After reconfiguring BIOS the system starts. There is no event describing the failure in Event viewer. Installing VS 2010 is absolutely necessary for me. I need help with diagnostic. I need to find where is the problem - I expect that the problem is in OCZ drive or in HDD controller on motherboard but I don't know how to find it. All components still have valid warranty. Can you recommend me some approach or tools to find the problem? Edit: I'm still looking for source of the problem. New information is that Windows are not able to perform check disk (Chkdsk) on the SSD system drive. After restarting it always starts checking drive and in part where files are checked it fails with BSOD - BAD_SYSTEM_CONFIG_INFO. After next restart and skipping check disk tests the system runs.

    Read the article

  • Windows Server 2008 R2 loses ability to connect to network share

    - by JamesB
    I could sure use some help with this one: I've got two Windows Server 2008 R2 x64 Terminal Servers, as well as several 2003 servers (DNS / Wins / AD / DC). On the two 2008 boxes, every now and then they will get in this mode where you can't map a drive to a random server. I say random server because it's not always the same server that you can't map to. Here is a summary of what I can and can't do: net view \\servername Sometimes this works, sometimes it does not. net view \\FQDN This always works. net view \\IPAddress This always works. ping servername Sometimes this works, sometimes it does not. ping FQDN This always works. ping IPAddress This always works. I've been looking all over for a solution to this. It sure seems like Microsoft would have a hotfix by now. The kicker to this is that it sometimes works great, especially after a reboot. It may run for 2 weeks just fine, but all of a sudden it will fail to resolve the remote server name. It will then be this way for a few days, then it might start working again. Also, while it's in the mode of not working, the other servers have no problem getting there. It's just these 2008 R2 Terminal Servers. Setting a static entry in the Hosts file and LMHosts does not make it work. All servers have static IPs and they are registered in DNS and Wins just fine. Here is a long thread on MS Technet of the exact same problem, but they don't have a good solution. Here is their workaround (It was from June of 2010): Good news - a hotfix is in the works and a workaround has been identified: Root cause is that since this is SMB1 all user sessions are on a single TCP connection to the remote server. The first user to initiate a connection to the remote SMB server has their logon-ID added to the structure defining the connection. If that user logs off all subsequent uses of that TCP session fail as the logon-id is no longer valid. As a workaround for now to keep the issue from happening you will want to have the user not logoff the Terminal Server only disconnect their sessions. Any word from anyone out there about a solution? Any help would sure be appreciated. Thanks, James

    Read the article

  • Conflicting ip routes with local table on attaching a virtual network interface

    - by user1071840
    I have an EC2 instance with these ip rules: $ sudo ip rule show 0: from all lookup local 32766: from all lookup main 32767: from all lookup default I can attach an elastic network interface to it with a private IP. Say the IP of my machine is 10.1.3.12 and the IP of the interface is 10.1.1.190. As soon as I attach the interface to my machine a new entry is added to the routing policy and local routing table: sudo ip rule show 0: from all lookup local 32765: from 10.1.1.190 lookup 10003 32766: from all lookup main 32767: from all lookup default $ sudo ip route show table local broadcast 10.1.1.0 dev eth3 proto kernel scope link src 10.1.1.190 local 10.1.1.190 dev eth3 proto kernel scope host src 10.1.1.190 broadcast 10.1.1.255 dev eth3 proto kernel scope link src 10.1.1.190 broadcast 10.1.3.0 dev eth0 proto kernel scope link src 10.1.3.12 local 10.1.3.12 dev eth0 proto kernel scope host src 10.1.3.12 broadcast 10.1.3.255 dev eth0 proto kernel scope link src 10.1.3.12 broadcast 127.0.0.0 dev lo proto kernel scope link src 127.0.0.1 local 127.0.0.0/8 dev lo proto kernel scope host src 127.0.0.1 local 127.0.0.1 dev lo proto kernel scope host src 127.0.0.1 broadcast 127.255.255.255 dev lo proto kernel scope link src 127.0.0.1 I can send traffic to this ENI directly from a host that can have the same IP as the host the ENI is attached to. This is where the problem starts. I ran tcpdump on the port in question and saw multiple SYNs going to the ENI with src '10.1.3.12' and destination '10.1.1.190' but didn't see even a single ACK. In my understanding if ACKs were being sent from the ENI they'd have destination as 10.1.3.12 i.e. the same as the local machine's IP and such packets will now be routed as local packets matching local routing policy: local 10.1.3.12 dev eth0 proto kernel scope host src 10.1.3.12 I'd like to send all the packets originating from 10.1.1.190 (my ENI) to go back on the same interface i.e. eth3 in this case. Contents of the nee table 10003 are: $ sudo ip route show table 10003 default via 10.1.1.1 dev eth3 I think I can do the following: I don't know if its possible but probably decrease the priority of local table so the packets match the table 10003. Use iptables to mangle these packets and update the local table route to include the mark information But I'm not sure if these are the right approaches.

    Read the article

  • Model M Keyboard inputs incorrect characters after logging in to Fedora

    - by mickburkejnr
    I recently bought a 24 year old IBM Model M keyboard. From what I gather, it'd been left on a shelf for the last 5 years, so you can imagine the amount of dust dirt and crap that was on it. Before cleaning it, I plugged it in to my laptop (running Fedora 17) using a PS/2 to USB adapter. What I found was, while it still works, the keys I press don't correspond to what is displayed on the screen. So for example, when I type S on the keyboard, I get ß display on the screen instead. At the time, I put this down to the adapter not working properly. Since then, I stripped the keys off the keyboard and cleaned the whole thing. It looks like it's just come out of a box! I then plugged it in to my computer (also running Fedora 17) via a standard PS/2 plug. The computer loaded up to the login screen, and I typed in my password. Pressed enter, and I logged straight in to my machine. At this point, I opened up a text editor and started typing some stuff. To my horror, the keystrokes I was entering weren't coming up as intended. What came up instead were characters that would map to the pressed key but only under a different keyboard language setting. I opened up a program to see what keyboard language had been selected, and the correct one for the keyboard was selected (which is UK in my case). I opened up a window that would show what characters mapped to what keys, and I pressed every single key on the keyboard, and every corresponding block representing each key lit up. I went back to the text editor to try again, but I was still getting these random characters. Whats more is that the backspace key would not work, although in the other utility it would flash when pressed. What I know is that at the login screen the keyboard must have entered the correct characters, otherwise I wouldn't have been able to log in. Further more, keys that don't respond while using a text editor as sending signals to the computer, as illustrated in that keyboard utility. The question is why random characters are displayed when they really shouldn't be? Would this be a hardware fault or a software issue?

    Read the article

  • How to enjoy DVD on Apple iPad

    - by user44251
    I believe many people spent a sleepless night yesterday waiting for the new Apple Tablet to come, just a few days ago or perhaps longer I noticed fierce debate about it, its name, size, capacity, processor, main features, price etc. And now, they can take a long breath with the new Apple Tablet named iPad officially released on 28, January, 2010 (Beijing Time). But I know a new battle just begins. iPad, sounds somewhat like iPod and it really shares some similarities in terms of shape like smart, light and portable. It has a 9.7-inch, LED-backlit, IPS display with a remarkable precise Multi-Touch screen. And yet, at just 1.5 lbs and 0.5 inches thin, it's easy to carry and use everywhere. It can greatly facilitates your experience with the web, emails, photos and videos. Right now, it can run almost 140.000 of the apps on the Apple store. It can even run the apps you have downloaded for your iPhone or iPod touch. But so far, I haven't seen any possibility that it can work with DVD, probability there is no built-in DVD-ROM or DVD player which can play DVD directly. As Apple iPad states, the video formats supported are MPEG-4 (MP4, M4V), H.264, MOV etc and audio formats accepted are AAC, Proteceted AAC, MP3, AIFF and WAV etc, those are formats that are commonly used with iMac. This could really a hard nut to crack if you want to watch your favourite DVD on this magic Apple iPad. But don't worry, there is still way out, you just need a few steps for ripping and importing DVD movies to Apple iPad with a simple application DVD to iPad converter What's on DVD to iPad Converter for Mac DVD to iPad converter for Mac is a powerful and professional application designed for the newly released Apple iPad which can rip, convert your DVD contents to Apple iPad compatible MPEG-4 (MP4, M4V), H.264, MOV etc, and other popular file formats like AVI, WMV, MPG, MKV, VOB, 3GP, FLV etc can also be converted so that you can put on your portable devices like iPod, iPhone, iRiver, BlackBerry etc. Besides, it can also extract audio from DVD videos and save as MP3, AIFF, AAC, WAV etc. Mac DVD to iPad converter has also been enhanced that can run both on PowerPC and Intel (Snow Leopard included). It can offer versatile editing features which allows you to make your own DVD videos. For example, you can cut your DVD to whatever length you like by Trim, crop off unwanted parts from DVD clips by Crop, add special effect like Gray, Emboss and Old film to make your videos more artistic. Besides, its built-in merging feature and batch mode allows you to join several DVD clips into a single one and do batch conversion. And more features can be expected if you afford a few minutes to try.

    Read the article

  • What is the probable failure - no BSOD, no event log, monitors sleeping, force reboot required

    - by Tyler
    Every 3 to 15 days, my PC freezes. This typically happens when the computer is idle, I'm coming home from work, back from vacation, etc. It's never happened while using my computer. The monitors are in power save mode The Caps Lock light on the (wireless) keyboard doesn't work Ctrl-alt-del has no effect, mouse (wireless) has no effect The hardware reset button and single press of power putton have no effect Computer does not appear on the network No BSOD, no memory dump Event logs have no errors or indications of problems near the time of crash. Only messages after reboot indicating that there was a reboot without a clean shutdown. Windows is set to never put the computer to sleep (just the display) Here are the vital stats of the build: OS Windows 8 Pro 64-bit CPU Intel i5-2400 Mobo Intel BOXDP67DE Micro ATX GPU MSI N460GTX Cyclone768D5/OC RAM CORSAIR XMS3 8GB (2 x 4GB) CMX8GX3M2A1333C9 PSU SeaSonic X Series X650 Gold System Drive Samsung 840 Pro 256 GB SSD Data Drive 2 x Western Digital WD20EARS 2TB in hardware RAID 1 Optical Lite-On DVD burner IHAS424-98 And here is the story of how the problem developed and what I've done to diagnose: January 2011, system built with Windows 7 64-bit, runs great. March 2011, Intel replaced the mobo because of the bad sata controllers. October 2012, upgrade to Windows 8 (problems start shortly after). January 2013, system freezes and causes network to fail for the whole house. Unplug the network cable and other devices and PCs can use the internet. Plug it back in, internet goes away for everyone. Reboot and everything is fine. March 2013, install Intel Gigabit CT PCI-E NIC, disable mobo nic in bios. Network strangeness goes away. Freezes are less frequent. Memtest shows no problems (20 passes). Early June 2013, replace Antec PSU with SeaSonic PSU. Mid June 2013, replace OCZ Vertex 2 SSD with Samsung SSD. Late June 2013, get frustrated and hope the community has some good ideas (I'm running out of budget to replace parts). My next plan of attack is setting "Turn off display" to Never and using a screen saver to see how that reacts on the next freeze. It makes me sad to waste power for up to 15 days though. Has anyone out there seen a problem like this? Any ideas on what kind of malfunction would act this way? Ideas of other diagnostic steps to take?

    Read the article

  • A gigabit network interface is CPU-limited to 25MB/s. How can I maximize the throughput?

    - by netvope
    I have a Acer Aspire R1600-U910H with a nForce gigabit network adapter. The maximum TCP throughput of it is about 25MB/s, and apparently it is limited by the single core Intel Atom 230; when the maximum throughput is reached, the CPU usage is about 50%-60%, which corresponds to full utilization considering this is a Hyper-threading enabled CPU. The same problem occurs on both Windows XP and on Ubuntu 8.04. On Windows, I have installed the latest nForce chipset driver, disabled power saving features, and enabled checksum offload. On Linux, the default driver has checksum offload enabled. There is no Linux driver available on Nvidia's website. ethtool -k eth0 shows that checksum offload is enabled: Offload parameters for eth0: rx-checksumming: on tx-checksumming: on scatter-gather: on tcp segmentation offload: on udp fragmentation offload: off generic segmentation offload: off The following is the output of powertop when the network is idle: Wakeups-from-idle per second : 61.9 interval: 10.0s no ACPI power usage estimate available Top causes for wakeups: 90.9% (101.3) <interrupt> : eth0 4.5% ( 5.0) iftop : schedule_timeout (process_timeout) 1.8% ( 2.0) <kernel core> : clocksource_register (clocksource_watchdog) 0.9% ( 1.0) dhcdbd : schedule_timeout (process_timeout) 0.5% ( 0.6) <kernel core> : neigh_table_init_no_netlink (neigh_periodic_timer) And when the maximum throughput of about 25MB/s is reached: Wakeups-from-idle per second : 11175.5 interval: 10.0s no ACPI power usage estimate available Top causes for wakeups: 99.9% (22097.4) <interrupt> : eth0 0.0% ( 5.0) iftop : schedule_timeout (process_timeout) 0.0% ( 2.0) <kernel core> : clocksource_register (clocksource_watchdog) 0.0% ( 1.0) dhcdbd : schedule_timeout (process_timeout) 0.0% ( 0.6) <kernel core> : neigh_table_init_no_netlink (neigh_periodic_timer) Notice the 20000 interrupts per second. Could this be the cause for the high CPU usage and low throughput? If so, how can I improve the situation? As a reference, the other computers in the network can usually transfer at 50+MB/s without problems. A computer with a Core 2 CPU generates only 5000 interrupts per second when it's transferring at 110MB/s. The number of interrupts is about 20 times less than the Atom system (if interrupts scale linearly with throughput.) And a minor question: How can I find out what is the driver in use for eth0?

    Read the article

  • Trying to get DNS services running on Windows Server 2008 R2, what am I getting wrong ?

    - by LaserBeak
    Ok, So I am basically trying to get a home server pc up that will provide Domain name services, act as Mail server and web server. I have one static IP, well it's not officially static but hasn't changed in two years so I'll call it static. I have done the following: Configured router NAT/virtual port forward UDP/TCP port 53 to the internal IP of my server 192.168.1.16, in adapter settings specified the manual settings: 192.168.1.16 IP, gateway 192.168.1.1, Subnet: 255.255.255.0 and loopback DNS: 127.0.0.1 Using my public my public IP Checked using http://www.canyouseeme.org/ that port 53 is open and is not being blocked by my ISP. It can see services on this port. Registered Domain name (mydomain.com.au) Updated whois database through the domain registrars site and registered NameServer names: ns0.mydomain.com.au and ns2.mydomain.com.au, both have been associated with my single public IP. (Waited 24 hours) Update the nameserver for mydomain.com.au: primary ns0.mydomain.com.au secondary: ns2.mydomain.com.au (waited 24+ hours) Installed Server 2008 R2, install web server role and DNS role. Webserver works when I enter my public IP into browser of any PC/mobile, get IIS7 welcome page. In DNS server: Created new forward lookup zone: ; ; Database file mydoman.com.au.dns for mydomain.com.au zone. ; Zone version: 10 ; @ IN SOA mydomain.com.au. mydomain.testdomain.com. ( 10 ; serial number 900 ; refresh 600 ; retry 86400 ; expire 3600 ) ; default TTL ; ; Zone NS records ; @ NS ns0.mydomain.com.au. @ NS ns1.mydomain.com.au. ; ; Zone records ; @ A 192.168.1.16 www A 192.168.1.16 The Domain name services will however not work, the whois database updated with ns0.mydomain.com.au etc. but when I type in my site name www.mydomain.com.au from an external machine it will not open site and I can't even ping it (Can't find host) When I check the ns0.mydomain.com.au NS record using a tool Like: http://www.squish.net/dnscheck/ I get: Security: Server ns0.mydomain.com.au (XXX.XXX.XXX.XX <- my public IP) is recursive Domain exists but there is no such record Any ideas, thanks...

    Read the article

  • MongoDB and datasets that don't fit in RAM no matter how hard you shove

    - by sysadmin1138
    This is very system dependent, but chances are near certain we'll scale past some arbitrary cliff and get into Real Trouble. I'm curious what kind of rules-of-thumb exist for a good RAM to Disk-space ratio. We're planning our next round of systems, and need to make some choices regarding RAM, SSDs, and how much of each the new nodes will get. But now for some performance details! During normal workflow of a single project-run, MongoDB is hit with a very high percentage of writes (70-80%). Once the second stage of the processing pipeline hits, it's extremely high read as it needs to deduplicate records identified in the first half of processing. This is the workflow for which "keep your working set in RAM" is made for, and we're designing around that assumption. The entire dataset is continually hit with random queries from end-user derived sources; though the frequency is irregular, the size is usually pretty small (groups of 10 documents). Since this is user-facing, the replies need to be under the "bored-now" threshold of 3 seconds. This access pattern is much less likely to be in cache, so will be very likely to incur disk hits. A secondary processing workflow is high read of previous processing runs that may be days, weeks, or even months old, and is run infrequently but still needs to be zippy. Up to 100% of the documents in the previous processing run will be accessed. No amount of cache-warming can help with this, I suspect. Finished document sizes vary widely, but the median size is about 8K. The high-read portion of the normal project processing strongly suggests the use of Replicas to help distribute the Read traffic. I have read elsewhere that a 1:10 RAM-GB to HD-GB is a good rule-of-thumb for slow disks, As we are seriously considering using much faster SSDs, I'd like to know if there is a similar rule of thumb for fast disks. I know we're using Mongo in a way where cache-everything really isn't going to fly, which is why I'm looking at ways to engineer a system that can survive such usage. The entire dataset will likely be most of a TB within half a year and keep growing.

    Read the article

  • Resetting root password on Fedora Core 3 - serial cable access only

    - by Sensible Eddie
    A little background: We have an old rackmount server running a customised version of Fedora, manufactured by a company called Navaho. The server is a TeamCAT, running some proprietary rubbish called Freedom2. We have to keep it going - the alternative is extraordinarily expensive, and the business is not likely to be running much longer to justify changing things. Through one means or another, it has fallen upon me to try and resolve our lack of root access. The previous admin has fallen under the proverbial bus, and nobody has any clue. We have no access to the root account for this server. ssh is running on the server, and there is one account admin that we can login with, however it has no permission to do anything (ironic...) The only other way into the server is with a null-modem serial cable. This works... up to a point. I can see the BIOS, I can see the post BIOS screen, and then I see "Starting grub", followed by another screen with about four lines of Linux information, but then it stops at that point. The server continues booting, and all services come online after around two minutes, but the serial terminal displays no more information. I understand it is possible to put Linux into "single user mode" to reset a root password, but I have no idea how to do this beyond trying to interrupt it at the grub stage listed above. When I have tried it just froze. It was almost like grub had appeared (since the server did not continue booting) but I couldn't see it on the serial terminal. Which made me think maybe the grub screen has some different serial settings? I don't know... it's the first time I've ever used serial for access! A friend of mine suggested trying to use a Fedora boot CD. We could boot from USB, so something along this approach is possible but again we still can only see what's going on with the serial terminal, so it might not be achievable. Does anyone have any suggestions for things I can try? I appreciate this is a bit of a long shot, but any assistance would be invaluable. *UPDATE 1 - 28/8/12 * - we will be making some attempts on this today and will post further details later!

    Read the article

  • How did what appears to be a virus get on my computer? (explanation of situation enclosed)

    - by Massimo
    My system is Windows XP SP3, updated with the latest patches. The PC is connected to a Cisco 877 ADSL router, which does NAT from the internal network to its single static public IP address. There are no forwarded ports, and the router's management console can only be accessed from the inside. I was doing two things: working on a remote office machine via VPN and browsing some web pages on the Cisco web site. The remote network is absolutely safe (it's a lab network, four virtual servers, no publicly accessible services and no users at all; also, none of what I'm going to describe ever happened there). The Cisco web site... well, I suppose is quite safe, too. Suddenly, something happened. Strange popups appears anywhere; programs claiming they're "antimalware", "antispyware" et so on begins autoinstalling; fake Windows Update and Security Center icons pop up in the system tray. svchost.exe began crashing repeatedly. Then, finally, after some minutes of this... BSOD. And, upon rebooting, BSOD again. Even in safe mode. Ok, that was obviously some virus/trojan/whatever. I had to install a new copy of Windows on another partition to clean things up. I found strange executables, services and DLLs almost anywhere. Amongst the other things, user32.dll and ndis.sys had been replaced. A fake software called "Antimalware Doctor" had been installed. There were services with completely random names or even GUIDs (!), and also ones called "IpSect" and "Darkness". There were executable files without an .exe extension. There were even two boot-class drivers, which I'm quite sure are the ones that finally caused the system to crash. A true massacre. Ok, now the questions: What the hell was that?!? It was something more than a simple virus! How did it manage to attack my computer, as I am behind a firewall and was not doing anything even only potentially harmful on the web at the time?

    Read the article

  • Cloud storage provider lost my data. How to back up next time?

    - by tomcam
    What do you do when cloud storage fails you? First, some background. A popular cloud storage provider (rhymes with Booger Link) damaged a bunch of my data. Getting it back was an uphill battle with all the usual accusations that it was my fault, etc. Finally I got the data back. Yes, I can back this up with evidence. Idiotically, I stayed with them, so I totally get that the rest of this is on me. The problem had been with a shared folder that works with all 12 computers my business and family use with the service. We'll call that folder the Tragic Briefcase. It is a sort of global folder that's publicly visible to all computers on the service. It's our main repository. Today I decided to deal with some residual effects of the Crash of '11. Part of the damage they did was that in just one of my computers (my primary, of course) all the documents in the Tragic Briefcase were duplicated in the Windows My Documents folder. I finally started deleting them. But guess what. Though they appeared to be duplicated in the file system, removing them from My Documents on the primary PC caused them to disappear from the Tragic Briefcase too. They efficiently disappeared from all the other computers' Tragic Briefcases as well. So now, 21 gigs of files are gone, and of course I don't know which ones. I want to avoid this in the future. Apart from using a different storage provider, the bigger picture is this: how do I back up my cloud data? A complete backup every week or so from web to local storage would cause me to exceed my ISP's bandwidth. Do I need to back up each of my 12 PCs locally? I do use Backupify for my primary Google Docs, but I have been storing taxes, confidential documents, Photoshop source, video source files, and so on using the web service. So it's a lot of data, but I need to keep it safe. Backup locally would also mean 2 backup drives or some kind of RAID per PC, right, because you can't trust a single point of failure? Assuming I move to DropBox or something of its ilk, what is the best way to make sure that if the next cloud storage provider messes up I can restore?

    Read the article

  • How should we serve files in a small bioinformatics cluster?

    - by cespinoza
    We have a small cluster of six ubuntu servers. We run bioinformatics analyses on these clusters. Each analysis takes about 24 hours to complete, each core i7 server can handle 2 at a time, takes as input about 5GB data and outputs about 10-25GB of data. We run dozens of these a week. The software is a hodgepodge of custom perl scripts and 3rd party sequence alignment software written in C/C++. Currently, files are served from two of the compute nodes (yes, we're using compute nodes as file servers)-- each node has 5 1TB sata drives mounted separately (no raid) and is pooled via glusterfs 2.0.1. They each have as 3 bonded intel ethernet pci gigabit ethernet cards, attached to a d-link DGS-1224T switch ($300 24 port consumer-level). We are not currently using jumbo frames (not sure why, actually). The two file-serving compute nodes are then mirrored via glusterfs. Each of the four other nodes mounts the files via glusterfs. The files are all large (4gb+), and are stored as bare files (no database/etc) if that matters. As you can imagine, this is a bit of a mess that grew organically without forethought and we want to improve it now that we're running out of space. Our analyses are I/O intensive and it is a bottle neck-- we're only getting 140mB/sec between the two fileservers, maybe 50mb/sec from the clients (which only have single NICs). We have a flexible budget which I can probably get up $5k or so. How should we spend our budget? We need at least 10TB of storage fast enough to serve all nodes. How fast/big does the cpu/memory of such a file server have to be? Should we use NFS, ATA over Ethernet, iSCSI, Glusterfs, or something else? Should we buy two or more servers and create some sort of storage cluster, or is 1 server enough for such a small number of nodes? Should we invest in faster NICs (say, PCI-express cards with multiple connectors)? The switch? Should we use raid, if so, hardware or software? and which raid (5, 6, 10, etc)? Any ideas appreciated. We're biologists, not IT gurus.

    Read the article

  • bluetooth connection using pybluez

    - by srj0408
    I am working on bluetooth not exactly on bluetooth stack-development but to use bluetooth in one of my project. I had done all that before using some of the py-bluez commands like hciconfig, hcitool scan , then simple-agents and using serial module inside python. But that was quite random. We were able to connect only one specific device based on its bluetooth address and there was no facility of reconnection once the devices are disconnected. Now i want to try out this stuff in a sequential manner like this (i am doing that all on a RPI and for at present on ubuntu 12.04.) i) Store some names in a file along with some other information with respect to that device. ii) Run a script to find out the device in locality with those names and if any one if found, report that. For this step, i had taken a reference from BTBook , made available from MIT. Below is the script for the same, but that script only search for the single name. from bluetooth import * target_name = "XT1033" target_address = None nearby_devices = discover_devices() for address in nearby_devices: if target_name == lookup_name( address ): target_address = address break if target_address is not None: print "found target bluetooth device with address ", target_address connect_socket(target_address); else: print "could not find target bluetooth device nearby" iii) Connect the device using client sock. But i dont have any device on which i can write a simple python script. My client can be any device that will be publishing data. Now i came through a script in the same book, that actually connect to a client requesting permission to connect to server. from bluetooth import * port = 1 server_sock=BluetoothSocket( RFCOMM ) server_sock.bind(("",port)) server_sock.listen(1) client_sock, client_info = server_sock.accept() print "Accepted connection from ", client_info data = client_sock.recv(1024) print "received [%s]" % data client_sock.close() server_sock.close() here client_sock, client_info = server_sock.accept() provide the client address and port requested to be connected. Can i pass address obtained from the earlier script to this, so that it connect server to the client? iv) Then if client get disconnected, re-connect(a simple polling can be used.) All this stuff can be done using bash and py-bluez functions but i want to do that in a sequential manner.I am not a master in python but i can do some small stuff. Can any one guide me for the same or can direct me to more usefull resource through which i can continue my coding part after finding the "X", "Y" named devices.

    Read the article

  • How to auto advance a PowerPoint slide after an exit animation is over?

    - by joooc
    PowerPoint entrance animation set up with "Start: With Previous" starts right when a new slide is advanced. However, if you set up an exit animation in the same way, it doesn't start with a slide ending sequence. Instead, the "Start: On Click" trigger needs to be used and after your exit animation is over you still need one extra click just to advance to the next slide. Workarounds to this are obvious: create a duplicate slide, make your ending animations from the original slide being your starting animations on the duplicate slide and let them be followed with whatever you want or create a transition slide with those ending animations only and set up "Change Advance slide - Automatically after - [the time it takes your animations to finish]". These workarounds will make it work for your audience, visually. However, it has an impact on slide numbers you might need to adjust accordingly and/or duplicate content changes. If you are the only one creating and using your presentation, this might be just fine. But if you are creating a presentation in collaborative mode with three other people and don't even know who will be the presenter at the end, you can mess things up. Let's be specific: most of my slides have 0.2s fly in entrance animation applied to blocks of content coming from right, bottom or left. Advancing to the next slide I want them to fly out in another 0.2s exit animation being followed by new slide 0.2s fly in entrance animation of the new blocks. The swapping of the blocks should be triggered while advancing to the next slide, as usually. As mentioned, I'm not able to achieve this without one extra click between the slides. I wrote a VBA script that should start together with an exit animation and will auto advance a slide after 0.3s when the exit animation is over. That way I should get rid of those extra clicks which are needed right now. Sub nextslide() iTime = 0.3 Start = Timer While Timer < Start + iTime DoEvents Wend With SlideShowWindows(1).View .GotoSlide (ActivePresentation.SlideShowWindow.View.Slide.SlideIndex + 1) End With End Sub It works well when binded on a box, button or another object. But I can't make it run on a single click (anywhere on the slide) so that it could start together with the exit animation onclick trigger. Creating a big transparent rectangular shape over the whole slide and binding the macro on it doesn't help either. By clicking it you only get the macro running, exit animation is not triggered. Anyway, I don't want to bind the macro to any other workaround object but the slide itself. Anyone knows how to trigger a PowerPoint VBA script on slide onclick event? Anyone knows a secret setting that will make the exit animation work as expected i.e. animating right before exiting a slide while transitioning to the next one? Anyone knows how to beat this dragon? Thank you!

    Read the article

  • Excel 2010: dynamic update of drop down list based upon datasource validation worksheet changes

    - by hornetbzz
    I have one worksheet for setting up the data sources of multiple data validation lists. in other words, I'm using this worksheet to provide drop down lists to multiple other worksheets. I need to dynamically update all worksheets upon any of a single or several changes on the data source worksheet. I may understand this should come with event macro over the entire workbook. My question is how to achieve this keeping the "OFFSET" formula across the whole workbook ? Thx To support my question, I put the piece of code that I'm trying to get it working : Provided the following informations : I'm using such a formula for a pseudo dynamic update of the drop down lists, for example : =OFFSET(MyDataSourceSheet!$O$2;0;0;COUNTA(MyDataSourceSheet!O:O)-1) I looked into the pearson book event chapter but I'm too noob for this. I understand this macro and implemented it successfully as a test with the drop down list on the same worksheet as the data source. My point is that I don't know how to deploy this over a complete workbook. Macro related to the datasource worksheet : Option Explicit Private Sub Worksheet_Change(ByVal Target As Range) ' Macro to update all worksheets with drop down list referenced upon ' this data source worksheet, base on ref names Dim cell As Range Dim isect As Range Dim vOldValue As Variant, vNewValue As Variant Dim dvLists(1 To 6) As String 'data validation area Dim OneValidationListName As Variant dvLists(1) = "mylist1" dvLists(2) = "mylist2" dvLists(3) = "mylist3" dvLists(4) = "mylist4" dvLists(5) = "mylist5" dvLists(6) = "mylist6" On Error GoTo errorHandler For Each OneValidationListName In dvLists 'Set isect = Application.Intersect(Target, ThisWorkbook.Names("STEP").RefersToRange) Set isect = Application.Intersect(Target, ThisWorkbook.Names(OneValidationListName).RefersToRange) ' If a change occured in the source data sheet If Not isect Is Nothing Then ' Prevent infinite loops Application.EnableEvents = False ' Get previous value of this cell With Target vNewValue = .Value Application.Undo vOldValue = .Value .Value = vNewValue End With ' LOCAL dropdown lists : For every cell with validation For Each cell In Me.UsedRange.SpecialCells(xlCellTypeAllValidation) With cell ' If it has list validation AND the validation formula matches AND the value is the old value If .Validation.Type = 3 And .Validation.Formula1 = "=" & OneValidationListName And .Value = vOldValue Then ' Debug ' MsgBox "Address: " & Target.Address ' Change the cell value cell.Value = vNewValue End If End With Next cell ' Call to other worksheets update macros Call Sheets(5).UpdateDropDownList(vOldValue, vNewValue) ' GoTo NowGetOut Application.EnableEvents = True End If Next OneValidationListName NowGetOut: Application.EnableEvents = True Exit Sub errorHandler: MsgBox "Err " & Err.Number & " : " & Err.Description Resume NowGetOut End Sub Macro UpdateDropDownList related to the destination worksheet : Sub UpdateDropDownList(Optional vOldValue As Variant, Optional vNewValue As Variant) ' Debug MsgBox "Received info for update : " & vNewValue ' For every cell with validation For Each cell In Me.UsedRange.SpecialCells(xlCellTypeAllValidation) With cell ' If it has list validation AND the validation formula matches AND the value is the old value ' If .Validation.Type = 3 And .Value = vOldValue Then If .Validation.Type = 3 And .Value = vOldValue Then ' Change the cell value cell.Value = vNewValue End If End With Next cell End Sub

    Read the article

  • How to move partition with Users to a new SATA controller?

    - by Al Kepp
    Our current situation: X79 chipset. Windows 7 installed on SATA SSD running on Marvell controller. C:\Users contains just Public and computer-name folders, all the other users are created in E:\Users. That partition is on two HDD's running on an Intel RAID controller. (The motherboard has got two SATA/RAID onboard controllers.) The goal is: Without reinstallation, I want to move SSD to a standard SATA controller. Then I want to get rid of RAID1, and move one HDD to the same internal Intel SATA controller and the current E:\Users move to that new place. I want it to stay as E:\users, but I need to reinstall the HDD to let it work in SATA mode without RAID. So I face several problems. I am sure all are solvable with free software utilities, but I don't know how exactly to do it. I can see the particular problems: I have got all users at E:\Users. When I turn off that E:\ disk, I won't be able to login to Windows. I need at least one admin account to be placed back at C:\users. The current C: runs on standard 120 GB SSD, but it is connected to a Marvell SATA/RAID controller. I am affraid the Windows won't let me put it to Intel SATA controller due to hardware/licence check, and I won't be able to use standard W7 recovery disk either, because there probably isn't marvel SATA/RAID driver on it. I haven't tried anything yet, because I am affraid I can end up with computer not working at all. (I want to move it to a standard ICH10 Intel SATA controller to let us have no problems in future with it. I think it is not very safe when we use any nonstandard hardware to boot the computer.) I need to somehow backup current E:\ disk and restore it to a new E:. I hope this will be the easy part (as long as the admin account will reside on C:). The E: RAID array is very large, but it is almost empty (less than 100 GB of data.) So I can make the partition smaller so it can easily fit to a single SATA HDD.

    Read the article

  • Tips for maximizing Nginx requests/sec?

    - by linkedlinked
    I'm building an analytics package, and project requirements state that I need to support 1 billion hits per day. Yep, "billion". In other words, no less than 12,000 hits per second sustained, and preferably some room to burst. I know I'll need multiple servers for this, but I'm trying to get maximum performance out of each node before "throwing more hardware at it". Right now, I have the hits-tracking portion completed, and well optimized. I pretty much just save the requests straight into Redis (for later processing with Hadoop). The application is Python/Django with a gunicorn for the gateway. My 2GB Ubuntu 10.04 Rackspace server (not a production machine) can serve about 1200 static files per second (benchmarked using Apache AB against a single static asset). To compare, if I swap out the static file link with my tracking link, I still get about 600 requests per second -- I think this means my tracker is well optimized, because it's only a factor of 2 slower than serving static assets. However, when I benchmark with millions of hits, I notice a few things -- No disk usage -- this is expected, because I've turned off all Nginx logs, and my custom code doesn't do anything but save the request details into Redis. Non-constant memory usage -- Presumably due to Redis' memory managing, my memory usage will gradually climb up and then drop back down, but it's never once been my bottleneck. System load hovers around 2-4, the system is still responsive during even my heaviest benchmarks, and I can still manually view http://mysite.com/tracking/pixel with little visible delay while my (other) server performs 600 requests per second. If I run a short test, say 50,000 hits (takes about 2m), I get a steady, reliable 600 requests per second. If I run a longer test (tried up to 3.5m so far), my r/s degrades to about 250. My questions -- a. Does it look like I'm maxing out this server yet? Is 1,200/s static files nginx performance comparable to what others have experienced? b. Are there common nginx tunings for such high-volume applications? I have worker threads set to 64, and gunicorn worker threads set to 8, but tweaking these values doesn't seem to help or harm me much. c. Are there any linux-level settings that could be limiting my incoming connections? d. What could cause my performance to degrade to 250r/s on long-running tests? Again, the memory is not maxing out during these tests, and HDD use is nil. Thanks in advance, all :)

    Read the article

  • How to prevent dual booted OSes from damaging each other?

    - by user1252434
    For better compatibility and performance in games I'm thinking about installing Windows additionally to Linux. I have security concerns about this, though. Note: "Windows" in the remaining text includes not only the OS but also any software running on it. Regardless of whether it comes included or is additionally installed, whether it is started intentionally or unintentionally (virus, malware). Is there an easy way to achieve the following requirements: Windows MUST NOT be able to kill my linux partition or my data disk neither single files (virus infection) nor overwriting the whole disk Windows MUST NOT be able to read data disk (- extra protection against spyware) Linux may or may not have access to the windows partition both Linux and Windows should have full access to the graphics card this rules out desktop VM solutions for gaming I want the manufacturer's windows graphics card driver Regarding Windows to be unable to destroy my linux install: this is not just the usual paranoia, that has happened to me in the past. So I don't accept "no ext4 driver" as an argument. Once bitten, twice shy. And even if destruction targeted at specific (linux) files is nearly impossible, there should be no way to shred the whole partition. I may accept the risk of malware breaking out of a barrier (e.g. VM) around the whole windows box, though. Currently I have a system disk (SSD) and a data disk (HDD), both SATA. I expect I have to add another disk. If i don't: even better. My CPU is a Intel Core i5, with VT-x and VT-d available, though untested. Ideas I've had so far: deactivate or hide other HDs until reboot at low level possible? can the boot loader (grub) do this for me? tiny VM layer: load windows in a VM that provides access to almost all hardware, except the HDs any ready made software solution for this? Preferably free. as I said: the main problem seems to be to provide full access to the graphics card hardware switch to cut power to disks commercial products expensive and lots of warnings against cheap home built solutions preferably all three hard disks with one switch (one push) mobile racks - won't wear of daily swapping be a problem?

    Read the article

< Previous Page | 653 654 655 656 657 658 659 660 661 662 663 664  | Next Page >