Need to find reason for Server Hung

rakesheee · 11-12-2017, 03:03 AM

Hi All,

In our environment, we are often encountered with server hung issue..

we verified sar, no errors in log files and dmesg.

what else should we check to get exact reason for server hung ..?

Keruskerfuerst · 11-12-2017, 03:17 AM

ondoho · 11-12-2017, 03:18 AM

we need much more information than that.
system information, both hard- and software etc. basic stuff.
and then you need to define what "hung" means.
and did you look at the logs(*)? what did you find?

(*) i'm being deliberately vague here.
there's dozens of possibilities because we know nothing about your system, and even when you know, there's still a half dozen options. all of which need to be looked at.

edit:
keruskerfürst managed to put it in 3 letters

Keruskerfuerst · 11-12-2017, 04:03 AM

Quote:

keruskerfürst managed to put it in 3 letters

Today is sunday...

Keruskerfuerst · 11-12-2017, 06:09 AM

If the server hangs, the hardware may be defective, especially the processor, the mainboard, the RAM.

Keruskerfuerst · 11-12-2017, 01:22 PM

What does help to recover the server from "hung" state?

Soft reset, hard reset or remote login with ssh?

!!! · 11-12-2017, 03:39 PM

Welcome to LQ, BUT... IF a friend said "car stalled",
how would you proceed? Specific symptoms?

Next 'soft skill' is web-researching: DDGoo: troubleshoot linux hang

Also do this www-search: wiki sysrq
Do any do anything?

Looking forward to hearing more 'clues'. Best wishes. Let us know...

rakesheee · 11-20-2017, 06:16 AM

Quote:

Originally Posted by Keruskerfuerst

What does help to recover the server from "hung" state?

Soft reset, hard reset or remote login with ssh?

Hi,

We have soft rebooted the server .. that resolved our problem

rakesheee · 11-20-2017, 06:20 AM

Quote:

Originally Posted by Keruskerfuerst

If the server hangs, the hardware may be defective, especially the processor, the mainboard, the RAM.

Hi,

To our surprise no where the problem is .. I have already mentioned no errors were detected on dmesg , to my knowledge if there were any hardware related errors it should be logged someehere either in dmesg or /var/log/messages.

Thanks for looking into in my problem .. It would be helpful to me if you can explain about Audit back log limit exceed

sundialsvcs · 11-20-2017, 07:31 AM

There are usually a few things to look for first:

(1) A problem with network connectivity: The process in question is not actually hung, but it is constantly "timing out" because some distant computer isn't answering the phone or because the network connection is failed or failing.

(2) Failed or failing hardware: Do you hear clicking sounds?

(3) Memory overcommit ("thrashing"): The easiest indicator of this is near-constant disk activity.

giis · 11-20-2017, 09:45 AM

Quote:

Originally Posted by rakesheee

Hi,

We have soft rebooted the server .. that resolved our problem

imho, reboot is not solution. You need to dig more into it, there is quite good chance it will hang again.

Try and use smartctl tools and perform health checks.