Project

General

Profile

Actions

Support #5838

open

/agatadisks

Added by Legay Eric over 8 years ago. Updated over 8 years ago.

Status:
New
Priority:
Normal
Assigned To:
Category:
-
Start date:
01/16/2014
Due date:
% Done:

0%

Estimated time:

Description

This night /agatadisks disappears this night.


Files

bug-16-12-2014.txt (11.8 KB) bug-16-12-2014.txt Aubert Yann, 01/17/2014 06:05 PM

Related issues

Related to AGATA DAQ - Support #5837: anodeds1 lostClosed01/16/201401/17/2014

Actions
Actions #1

Updated by Legay Eric over 8 years ago

Connected to anodeds2 :
  • gpfs off
  • umount /agatadisks
  • gpfs on
  • mount /agatadisks

Always in NFS stale

Actions #2

Updated by Legay Eric over 8 years ago

Anodeds1 rebooted ...

complicated reboot ...

Always NFS stale

Actions #3

Updated by Legay Eric over 8 years ago

Only 250 GB free on sunxfire ...

Need to find a way to free disk if not possible to have agatadisks quickly ...

Actions #4

Updated by Grave Xavier over 8 years ago

after the mount did you restart the nfs-kernel-server ?

Actions #5

Updated by Grave Xavier over 8 years ago

root@anode01:~# ls /agatadisks/
root@anode01:~# umount /agatadisks/
umount: /agatadisks/: not mounted
root@anode01:~# mount /agatadisks
root@anode01:~# ls /agatadisks/
BLOCKED
/etc/init.d/nfs-kernel-server restart
Stopping NFS kernel daemon: mountd nfsd.
Unexporting directories for NFS kernel daemon....
Exporting directories for NFS kernel daemon....
Starting NFS kernel daemon: nfsd mountd.
Deblocked the BLOCKED command on anode01

CSSH is my friend now

I don't close the ticket since I think Yann should investigate and understand what was the problem

Actions #6

Updated by Grave Xavier over 8 years ago

using cssh I have :
umount /agatadisks
mount /agatadisks
DAQ should run now

Actions #7

Updated by Grave Xavier over 8 years ago

I used a umount -lf /agatadisks on visux and analysisx in order to avoid device busy problems

Actions #8

Updated by Ralet Damian over 8 years ago

  • Priority changed from Immediate to Normal

Was the non disk access due to the crash of anodeds1? Or anodeds1 with the event_builder on it induced the crash?

Actions #9

Updated by Legay Eric over 8 years ago

  • Assigned To changed from Aubert Yann to Ralet Damian

Can you detail your question Damian pls ?

Actions #10

Updated by Ralet Damian over 8 years ago

My question is to understand what happend the morning of the crash. What is the reason for this crash.
Since it is the first time we have such a problem, I am wondering if it is a simple crash of the computer, or if the reason is the eventbuilding that crash the anode.

It is not crutial, but if the event builder can crash the anode, we are having a stability problem, and we should avoid to do event building on anodeds1.

Actions #11

Updated by Ralet Damian over 8 years ago

  • Assigned To changed from Ralet Damian to Legay Eric
Actions #12

Updated by Legay Eric over 8 years ago

  • Assigned To changed from Legay Eric to Aubert Yann

I have to call you this afternoon in GSI.

When are you available ?

Actions #13

Updated by Aubert Yann over 8 years ago

Maybe related to this syslog trace. Need more investigation.

Actions

Also available in: Atom PDF