java invoked oom-killer on server with 24Gb of Memory

[1stRaven]

Recently, my server has been crashed and taking down all websites and user shares, leaving only telnet running although this is also 50/50 if it remains. I end up having to run a shutdown -r now command in order to get the server to reboot. I have been having this issue for a few days now but haave managed to get the syslog tonight on the latest crash.

 

In the log, the server crash is listed at 19:42:27 where java invoked oom-killer and continues to do so until the server is rebooted. Does anyone know how to resolve this as my searches on here are not bringing me any answers.

 

The system : Supermicro h8dme-2 with twin 6 core opteron 2419 processors and 24gb of ram.

Drives : 1 x 2tb Parity, 1 x 2tb cache and 20 data drives between 1 and 2tbs. Total space of around 22Tb.

Unraid version : 5.0 Final

Plugins : Sickbeard, Sab, Plex, Crashplan, APC UPS and Cache Directories (this one is disabled at the moment)

 

Thanks for the help.

Syslog is attached.

syslog.zip

[sureguy]

Your server is probably running out of low memory, this is not a problem with 64 bit oses.

 

You may want to investigate unraid 6

 

 

Sent from a mobile device, sorry for any typos.

 

[1stRaven]

I'll take a look at the beta then but with the amount of data stored, ita difficult to fully back up the media elsewhere.

[unevent]

Recently, my server has been crashed and taking down all websites and user shares, leaving only telnet running although this is also 50/50 if it remains. I end up having to run a shutdown -r now command in order to get the server to reboot. I have been having this issue for a few days now but haave managed to get the syslog tonight on the latest crash.

 

In the log, the server crash is listed at 19:42:27 where java invoked oom-killer and continues to do so until the server is rebooted. Does anyone know how to resolve this as my searches on here are not bringing me any answers.

 

The system : Supermicro h8dme-2 with twin 6 core opteron 2419 processors and 24gb of ram.

Drives : 1 x 2tb Parity, 1 x 2tb cache and 20 data drives between 1 and 2tbs. Total space of around 22Tb.

Unraid version : 5.0 Final

Plugins : Sickbeard, Sab, Plex, Crashplan, APC UPS and Cache Directories (this one is disabled at the moment)

 

Thanks for the help.

Syslog is attached.

 

Notice a couple things:

• Looks like /boot/config/passwd and/or shadow may have been deleted at some point, perhaps during an upgrade which is giving the "Failed to find user 'avahi'" error.  Search the forums to see how to correct.
  
• Log is showing numerous mce hardware errors that start to occur after the start of the parity check.  See if there is anything in /var/log/mcelog.  Problem is that strucrture is lost on reboot/reset so there may not be anything until the next time the mce errors are logged.  They usually are hardware errors like memory problems and things like that.
  
• sshd has a configuration issue with port 22 on 0.0.0.0.  Might want to double-check the config
  
• The java looks to be part of Plex and both Java and Plex ate a lot of memory before dying.  May be a result of the hardware errors, see second bullet, or some configuration issue that doesn't get logged.