Results 1 to 24 of 24
  1. #1
    Join Date
    Jan 2008
    Location
    Adelaide, Australia
    Posts
    428

    512MB Memory VPS... Out of memory already?

    Okay, Im relatively new to all of this, so please excuse me!

    In short, I have a 512mb DV server with Mediatemple, which I am running 24 (ish) domains off (most of them static websites) and a teamspeak server. I would say MAX theres 10 users online at a time)

    Now, I know its running out of memory because i get frequent QoS Alerts in plesk (kmemsize is apparently the memory size):
    Oct 01, 2009 11:52:57 AM Black zone kmemsize

    I have attached my results (when I did top).

    My questions are:


    1. Should I be expecting to be out of memory running what I am?

    2. Is there a way to see the problem domains (memory wise)?

    3. Are there any ways I can reduce the memory? (I have followed this already: http://kb.mediatemple.net/questions/...ur+(dv)+server.)

    4. Where is the memory usage coming from (I am finding it very hard to understand TOP)
    Attached Thumbnails Attached Thumbnails top.gif  

  2. #2
    Join Date
    Dec 2003
    Location
    Boston, MA
    Posts
    603
    memory looks as if your apache processes are high and could use some tweaking as well as sql. I dont know what your pushing on either but from 10 domains? This looks abnormal. Are they massive forums or smaller websites?
    Axcelx Technologies - James
    Boston Colocation | Boston VPS
    Massachusetts Server Colocation and Dedicated Servers

  3. #3
    Join Date
    Mar 2009
    Posts
    3,807
    Are you running any PHP/SQL intensive sites?

  4. #4
    Do you have SuPHP running in your PHP config? That can be a bit of a memory hog. Without more info it's hard to tell.
    Try checking your server logs, it should give you out of memory errors, it could be caused by a PHP spike.

  5. #5
    Join Date
    Apr 2009
    Location
    Mauritius
    Posts
    407
    isn't kmemsize the number of Inodes. If inodes are low then this could happen.
    HSV Informatics Ltd - Since 2011
    Shared Hosting | VPS | Dedicated Servers

  6. #6
    Join Date
    Mar 2009
    Posts
    395
    Quote Originally Posted by Paul Millar View Post
    Now, I know its running out of memory because i get frequent QoS Alerts in plesk (kmemsize is apparently the memory size).
    kmemsize is one of the (many) memory parameters in a virtuozzo vps. Please post the output of this command:

    cat /proc/user_beancounters

  7. #7
    Join Date
    Feb 2005
    Location
    Australia
    Posts
    5,842
    Quote Originally Posted by mainpipe View Post
    kmemsize is one of the (many) memory parameters in a virtuozzo vps. Please post the output of this command:

    cat /proc/user_beancounters
    Yep. There are several of these niggling extra limitations on Virtuozzo UBC (not SLM) and OpenVZ. But it's a secondary parameter - provided you're not approaching the important memory limits some providers will increase your kmemsize limit on request.
    Chris

    "Some problems are so complex that you have to be highly intelligent and well informed just to be undecided about them." - Laurence J. Peter

  8. #8
    I find it odd you have no swap space
    Garry Dolley @ ARP Networks | gdolley _at_ arpnetworks . com | #arpnetworks on Freenode | @arpnetworks, @bsdvps
    State of the art services, strong community and friendly support
    FreeBSD, OpenBSD, Linux VPS • ARP Metal™ Dedicated Servers • Colocation • IP Transit
    Native IPv6 • VNC Console • Tunnel VNC over SSH • Serial Console over SSH • DNS Manager

  9. #9
    Join Date
    Mar 2009
    Posts
    395
    Quote Originally Posted by garrycc View Post
    I find it odd you have no swap space
    Virtuozzo (Parallels) based vps's don't have swap.

  10. #10
    Join Date
    Jan 2008
    Location
    Adelaide, Australia
    Posts
    428
    Regarding what is actually being ran. There are two smallish forums (http://rs-db.co.cc/forums/) and (http://guildreckoning.eu/forums/)

    And something which does seem to be quite php intensive: - but is it really that intensive? Theres like max 2 users on it at once.
    http://guildreckoning.eu/

    Beancounter thing attached.
    Attached Thumbnails Attached Thumbnails cat.gif  

  11. #11
    Join Date
    Feb 2005
    Location
    Australia
    Posts
    5,842
    A search on "beancounters" in this forum will find you lots of useful discussion - see this thread in particular for the handy interpretation script.

    IMO, the 12MB kernel memory barrier (kmemsize) on a 512MB VPS is way too low - I have 18MB on a 128MB VPS. Ask your provider if they're willing to increase this for you.

    Edit: You'll probably need at least double your current kmemsize limits to have a hope of using all your "guaranteed" memory. Welcome to the wonderful world of UBC.
    Last edited by foobic; 10-01-2009 at 10:33 PM.
    Chris

    "Some problems are so complex that you have to be highly intelligent and well informed just to be undecided about them." - Laurence J. Peter

  12. #12
    Join Date
    Jan 2008
    Location
    Adelaide, Australia
    Posts
    428
    Quote Originally Posted by foobic View Post
    A search on "beancounters" in this forum will find you lots of useful discussion - see this thread in particular for the handy interpretation script.

    IMO, the 12MB kernel memory barrier (kmemsize) on a 512MB VPS is way too low - I have 18MB on a 128MB VPS. Ask your provider if they're willing to increase this for you.

    Edit: You'll probably need at least double your current kmemsize limits to have a hope of using all your "guaranteed" memory. Welcome to the wonderful world of UBC.
    But it says I'm already using my guaranteed memory (512Mb)

  13. #13
    Join Date
    Jan 2008
    Location
    Adelaide, Australia
    Posts
    428
    Regarding Logs.

    Code:
    [Fri Oct 02 00:47:00 2009] [error] python_handler: no interpreter callback found.
    [Fri Oct 02 00:47:18 2009] [error] (12)Cannot allocate memory: fork: Unable to fork new process
    [Fri Oct 02 00:47:28 2009] [error] (12)Cannot allocate memory: fork: Unable to fork new process
    [Fri Oct 02 00:47:38 2009] [error] (12)Cannot allocate memory: fork: Unable to fork new process
    [Fri Oct 02 00:47:41 2009] [error] (12)Cannot allocate memory: apr_socket_accept: (client socket)
    [Fri Oct 02 00:47:41 2009] [error] make_obcallback: could not import mod_python.apache.\n
    Traceback (most recent call last):
      File "/usr/lib/python2.4/site-packages/mod_python/apache.py", line 21, in ?
        import traceback
      File "/usr/lib/python2.4/traceback.py", line 3, in ?
        import linecache
    ImportError: No module named linecache
    [Fri Oct 02 00:47:41 2009] [error] python_handler: no interpreter callback found.
    [Fri Oct 02 00:47:48 2009] [notice] child pid 21964 exit signal Segmentation fault (11)
    [Fri Oct 02 00:48:12 2009] [error] (12)Cannot allocate memory: fork: Unable to fork new process
    [Fri Oct 02 00:48:14 2009] [error] (12)Cannot allocate memory: apr_socket_accept: (client socket)
    [Fri Oct 02 00:48:22 2009] [error] (12)Cannot allocate memory: fork: Unable to fork new process
    [Fri Oct 02 00:48:32 2009] [error] (12)Cannot allocate memory: fork: Unable to fork new process
    [Fri Oct 02 00:48:42 2009] [notice] child pid 23780 exit signal Segmentation fault (11)
    [Fri Oct 02 00:50:21 2009] [error] (12)Cannot allocate memory: fork: Unable to fork new process
    [Fri Oct 02 00:50:31 2009] [error] (12)Cannot allocate memory: fork: Unable to fork new process
    [Fri Oct 02 00:50:32 2009] [error] (12)Cannot allocate memory: apr_socket_accept: (client socket)
    [Fri Oct 02 00:50:32 2009] [error] make_obcallback: could not import mod_python.apache.\n
    Traceback (most recent call last):
      File "/usr/lib/python2.4/site-packages/mod_python/apache.py", line 21, in ?
        import traceback
      File "/usr/lib/python2.4/traceback.py", line 3, in ?
        import linecache
    ImportError: No module named linecache

  14. #14
    Join Date
    Feb 2005
    Location
    Australia
    Posts
    5,842
    Quote Originally Posted by Paul Millar View Post
    But it says I'm already using my guaranteed memory (512Mb)
    Your 512MB guarantee is correct (vmguarpages barrier). Your current / max usage is 192MB (oomguarpages held / maxheld). Your allocated memory - also what you see in top - is 512MB (privvmpages held) but that's what burst is for. Read the other threads and try out that script.
    Chris

    "Some problems are so complex that you have to be highly intelligent and well informed just to be undecided about them." - Laurence J. Peter

  15. #15
    Join Date
    Jan 2008
    Location
    Adelaide, Australia
    Posts
    428
    I'm confused!
    Surely that will tell me nothing? - As if im using my guaranteed memory then what I need to do is find out where the memory is going & try in fix it? (Eg. is it one of the applications on a website / a process that shouldnt be running / should I really be using up all the 512mb on what im running atm).

    Apologies for my "noobiness" when it comes to this.

  16. #16
    Join Date
    Feb 2005
    Location
    Australia
    Posts
    5,842
    On a UBC system the only indicators that actually matter are in /proc/user_beancounters. In your case this shows zillions of errors on kmemsize, meaning that your processes are being killed because you're hitting your kernel memory limit of 12MB. Nothing else really matters, and I seriously doubt your chances of significantly reducing your usage here, so all I can suggest is you get your kmemsize limits increased or try another supplier.
    Chris

    "Some problems are so complex that you have to be highly intelligent and well informed just to be undecided about them." - Laurence J. Peter

  17. #17
    Join Date
    May 2007
    Location
    Dublin, California
    Posts
    238
    Quote Originally Posted by Paul Millar View Post
    I'm confused!
    Surely that will tell me nothing? - As if im using my guaranteed memory then what I need to do is find out where the memory is going & try in fix it? (Eg. is it one of the applications on a website / a process that shouldnt be running / should I really be using up all the 512mb on what im running atm).

    Apologies for my "noobiness" when it comes to this.
    Have you attempted to contact MediaTemple for help with this issue?

  18. #18
    Join Date
    Jan 2008
    Location
    Adelaide, Australia
    Posts
    428
    Quote Originally Posted by kotrt99 View Post
    Have you attempted to contact MediaTemple for help with this issue?
    Not yet, I thought I'd better check out whether or not it was something either I was doing wrong or whether it is't.

  19. #19
    Join Date
    Jan 2008
    Location
    Adelaide, Australia
    Posts
    428
    Quote Originally Posted by foobic View Post
    On a UBC system the only indicators that actually matter are in /proc/user_beancounters. In your case this shows zillions of errors on kmemsize, meaning that your processes are being killed because you're hitting your kernel memory limit of 12MB. Nothing else really matters, and I seriously doubt your chances of significantly reducing your usage here, so all I can suggest is you get your kmemsize limits increased or try another supplier.
    Thank you

  20. #20
    Join Date
    Mar 2009
    Posts
    395
    Quote Originally Posted by foobic View Post
    On a UBC system the only indicators that actually matter are in /proc/user_beancounters. In your case this shows zillions of errors on kmemsize, meaning that your processes are being killed because you're hitting your kernel memory limit of 12MB. Nothing else really matters, and I seriously doubt your chances of significantly reducing your usage here, so all I can suggest is you get your kmemsize limits increased or try another supplier.
    True.

    I have an account with ServInt and with their $49 Essential plan they have the kmemsize limit set to 132MB. That is a significant difference.

    You need to monitor the beancounters, specifically the failcnts.

    You are also showing failures under the tcp send buffer.
    Last edited by mainpipe; 10-02-2009 at 02:08 PM.

  21. #21
    I also have a 512Mb VPS and my privmpages are high with all accounts suspended!

    Here's a screenshot:

    any ideas?
    Attached Thumbnails Attached Thumbnails beancounters.jpg  

  22. #22
    Join Date
    Mar 2005
    Location
    Labrador, Canada
    Posts
    951
    Different issue entirely (you may want to start a new thread for it). You're hitting your burst memory limit, so you would benefit from optimizing your VPS for memory usage, or upgrading.

  23. #23
    So it is normal to have those failcnts even if no sites are running on my vps?

  24. #24
    In UBC, virtual size of memory is accounted and subtracted from the memory allocated for VPS. Just do a stack size tuning for python and apache or whatever else you are running and reduce memory usage considerably(even upto 50%).

    As per the screen shots, python's VSZ is almost 10MB and why you are leaving ~10MB of stack unnecessarily? It will even work with 256KB.

Similar Threads

  1. 512MB memory not enough
    By Sciencemies in forum VPS Hosting
    Replies: 24
    Last Post: 08-20-2009, 10:05 PM
  2. 14x Sticks of 512MB DDR 266 ECC Memory
    By SwitchCentral in forum Other Offers & Requests
    Replies: 2
    Last Post: 08-13-2009, 11:19 PM
  3. Suggestion for cheap dedicated server with 512MB memory?
    By lolmannz in forum Dedicated Server
    Replies: 9
    Last Post: 12-30-2007, 08:37 PM
  4. Notebook Memory: 2x 512MB DDR PC2700
    By Anky in forum Other Web Hosting Related Offers
    Replies: 7
    Last Post: 09-13-2005, 10:18 PM
  5. $0 SETUP -- P4 1.7GHZ, 100GB transfer, 512MB of memory $99
    By infinity7 in forum Dedicated Hosting Offers
    Replies: 7
    Last Post: 01-12-2003, 03:20 AM

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •