V 6.2.4 "Cannot allocate memory" in logs and server eventually dies

[LinuxGuyGary]

Hi Folks,

 

Helping out a friend here with a trial key and an entry level server. 

 

Its currently running Version 6.2.4 on a Dell 755  Core 2 Duo @ 2.6GHz.  Using the onboard NIC and sata sockets.

There are two Dockers installed, The Limetech/Plex Media Server and Crashplan.  Plex is running, but Crashplan is not yet as we have not been able to make the system stable

 

The server has 4GB ram install, Memtest passed 8 passes without issue. 

 

The server runs for a few hours or possibly a day or two, and then things seems to deteriorate with multiple lines in the log like :

 

Feb 6 02:06:01 Ninja emhttp: err: _shcmd: system: Cannot allocate memory

Feb 6 02:06:01 Ninja emhttp: err: _shcmd: system: Cannot allocate memory

Feb 6 02:06:01 Ninja emhttp: err: _shcmd: system: Cannot allocate memory

Feb 6 02:06:01 Ninja emhttp: err: _shcmd: system: Cannot allocate memory

 

 

Eventually things crawl to a stop. 

 

SSH to the box and checking memory usually gives numbers similar to this:

 

root@Ninja:~# free

              total        used        free      shared  buff/cache  available

Mem:        3910884      314160      822772      351868    2773952    2848460

Swap:            0          0          0

 

 

Fix Common problems plugin finds no issues except the version update to 6.3.0

 

I have attached the diagnostics zip . 

ninja-diagnostics-20170206-0926.zip

[John_M]

Have you tried adding more memory? unRAID doesn't use swap space and 4 GB of RAM isn't an awful lot.

 

I can't open your diagnostics zip.

 

[LinuxGuyGary]

hmmm... Curious,

 

I can download the zip from the message and see it, so not sure what that's about, but I've attached another copy of the diagnostics files zipped on another linux box.

 

This PC has only 4 memory slots (all filled)  and is an older machine unfortunately I do no have any larger DIMMs to put in it without buying more. 

 

I was curious about this that I also see in the syslog...

 

kernel: pmd_set_huge: Cannot satisfy [mem 0xe0000000-0xe0200000] with a huge-page mapping due to MTRR override.

diagnostics-20170206-0926.zip

[John_M]

I was able to open your second zip file.

 

You could try disabling the VM service (Settings -> VM Manager). That would save you some RAM. I don't use Crashplan but I understand it's a memory hog so be warned if you're tempted to run it.

 

Did you try Googling the MTRR override message? I found this but I'm not a kernel programmer.

[LinuxGuyGary]

John_M

 

I've managed to secure a bit more memory, I'll replace that, and make sure the BIOS is at the latest version ...

 

I've also disabled the VM Service per your suggestion ..

 

We'll see how that all  goes.

 

 

[John_M]

Another place you might be able to save a little RAM is to reduce that reserved by any integrated graphics to the minimum (but non-zero!) amount.