Hi,<br><br>Can someone point me to typical memory sizes they see on their live systems from the MogileFS processes? And whether you restart these processes periodically to work around memory leaks?<br><br><br>On single machine with single disk (which is admittedly not the best configuration to test mogileFS), this is what I get.
<br>with 5 queryworkers, Total size is *200* MB --mogilefsd processes will use 150Meg and<br>mogstored another 50 meg. <br><br>Thanx, HB<br><br>PS<br>(mogilefsd(replicate) bloats in size for me. I am probably running into some obscure bug because there is nothing that can be replicated in this config).
<br><br><br><br>=========when the system is active Resident size is close to the size of the process =======<br>hbist@bestprice:/var/log$ ps -eo rss,vsz,command |egrep 'mog|COMMAND'<br> RSS VSZ COMMAND<br> 7072 34028 mogstored
<br> 1780 4920 mogstored [diskusage]<br> 1652 4920 mogstored [iostat]<br> 2104 6720 mogstored [fidsizes]<br> 5108 16428 mogilefsd<br>107976 218024 mogilefsd [replicate]<br> 5940 16684 mogilefsd [queryworker]<br> 5928 16692 mogilefsd [queryworker]
<br> 5908 16696 mogilefsd [queryworker]<br> 4352 16784 mogilefsd [fsck]<br> 6212 17160 mogilefsd [delete]<br> 7904 18332 mogilefsd [monitor]<br> 8672 16676 mogilefsd [reaper]<br><br>