2 minutes to kill server

  • Filter
  • Time
  • Show
Clear All
new posts

  • #16
    How regular is the "only every X minutes"?

    Do you have any cron jobs or does the application has any re-indexing jobs that it runs regularily?

    And when you get it the next time, run the SHOW PROCESSLIST command and read thru _all_ the rows. Or better yet post them her also in an attachment. And please make it a text file since it is a much easier format to filter.

    Because you have something that is taking a lot of CPU and it isn't the amount of processes since most of them was waiting for opening a table, I think that it is just a secondary effect since they can't open the table they want to and then the mysql threads are delayed and that is when they start to pile up.


    • #17
      Yes, there are some cronjobs, but they just insert some rows (up to 100 rows per run, usually about 20-50 rows) to database and exit. I'll check these cronjobs, but they are really simple - a few lines of code to insert several rows of data.


      • #18
        This sounds like an IO bottleneck, are you running off of SAN disks?... nm just saw you posted its SATA drives...