1. Please take a little time for this simple survey! Thank you for participating!
    Dismiss Notice
  2. Dear Pleskians, please read this carefully! New attachments and other rules Thank you!
    Dismiss Notice
  3. Dear Pleskians, I really hope that you will share your opinion in this Special topic for chatter about Plesk in the Clouds. Thank you!
    Dismiss Notice

InnoDB out of memory

Discussion in 'Plesk for Linux - 8.x and Older' started by muffinresearch, Mar 23, 2006.

  1. muffinresearch

    muffinresearch Guest

    I'm seeing the following in the mysqld logs and just recently the server has hung twice.

    060320 16:42:32 mysqld started
    060320 16:42:33 InnoDB: Out of memory in additional memory pool.
    InnoDB: InnoDB will start allocating memory from the OS.
    InnoDB: You may get better performance if you configure a bigger
    InnoDB: value in the MySQL my.cnf file for
    InnoDB: innodb_additional_mem_pool_size.
    060320 16:42:33 InnoDB: Database was not shut down normally.
    InnoDB: Starting recovery from log files...
    InnoDB: Starting log scan based on checkpoint at
    InnoDB: log sequence number 0 89774760
    InnoDB: Doing recovery: scanned up to log sequence number 0 89774760
    060320 16:42:33 InnoDB: Flushing modified pages from the buffer pool...
    060320 16:42:33 InnoDB: Started
    /usr/libexec/mysqld: ready for connections

    This looks like a likely culprit and as Plesk is the only thing that uses InnoDB I reckon I need to increase the memory allocation up from 8MB. My system has 256MB so I was thinking of upping it to 16MB to start. Does this sound like a good plan or is there anything else that could be causing the hangs. Also presumably it is /etc/my.cnf I should change?
  2. wagnerch

    wagnerch Guest

    I would be suprised if this is your problem. Did you happen to look in /var/log/messages for anything that may explain why your server "hangs"? Also when you mention "hangs", does it require someone to power cycle the machine or are you just talking about the http service?
  3. muffinresearch

    muffinresearch Guest

    Yep I checked /var/log/messages and I found kernel messages saying that Mysqld ran out of memory and was killed.

    To clarify the entire server had hung bothtimes and no service was available. The server had to be hardware rebooted both times.

    I would paste them here but I have no access to SSH from work.

    I upped the memory to the recommended levels and will see what happens over the next few days.

    Additionally I found lots of "stuck" messages in the mail queue whchi I have dealt with as they would never be sent.

    They might be unrelated but everything seems to be working much better now.

    I will keep an eye on mysqld.log to see if anything changes.
  4. wagnerch

    wagnerch Guest

    Still skeptical, out of memory wouldn't cause the box to hang. But good luck, hopefully it will solve your problem. BTW, Mysql always complains about InnoDB out of memory but it never affects the server.
  5. muffinresearch

    muffinresearch Guest

    Thanks for your comments, sounds like the MySQL logs were a red herring. I've just noticed that our server which should have 512MB of memory is currently running with 192MB

    Looks like some thing is a bit wrong on the hardware front. So I am getting in touch with the server supplier to get the hardware looked at.

    Hopefully with minimum disruption :)
  6. wagnerch

    wagnerch Guest

    Hmm, yeah that's definately odd. Bad RAM would have unpredictable results (especially server freezing). I have heard of a bank of RAM failing, but I would think you would have 256MB. Possibly 64M is mapped to the video card (ewe).
  7. muffinresearch

    muffinresearch Guest

    My thoughts exactly!