0

We are trying to trace an issue that's occurring on our SUSE 12 server - twice this week we have had a particular NFS share become un-mounted, and we can't currently determine the cause.

Our server...

  • Only installed a few weeks ago
  • Only runs a single SAP instance, no other applications other than standard SUSE 12
  • Only 3 users defined in the operating system - all Unix administrators, no end users
  • No cron jobs, only 4 custom shell scripts (which we've checked and given the all-clear)
  • No NFS exports, no interfaces coming in to this server (eg no inbound ftp), SUSE 12 firewall in place blocking all ports other than those we have defined explicitly.

With all that said, we can't determine what/how 2 particular NFS mounts are being unmounted.

Things we have looked at...

  • No entries of relevance in the system logs
  • No umount commands in the history file of any users
  • The unmounts occurred a few days apart, and at different times
  • The 2 mounted filesystems are on the same target server, however we have 2 other mounts from the same target server that are still connected, so it hasn't unmounted everything.
  • The mounted filesystem has a name made up by us - ie it's not a common name, and is not a filesystem used/known by the SAP application
  • No auto-unmount configured
  • We have another identical SUSE server in our test environment, and it has never experienced this issue

Is there anything we can do to help us capture a umount command being executed, or something like that, so we can determine what is causing the file system to become unmounted? Or is there any reasons why NFS/SUSE might be unmounting these particular filesystems?

Thanks.

0

2 Answers 2

0

NFS protocol has an idle timeout after which the NFS mount will be unmounted.

Probably there network issue , a firewall cutting the connection after a period of inactivity. Try to monitor the network connection.

0

Suse ended up acknowledging this as a bug, and have since rolled a fix in to the stable kernel from late 2017. We have successfully applied it to our environment and have not seen the issue occur since this time.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .