From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.7.8) Gecko/20050513 Fedora/1.7.8-1.3.1 Description of problem: Shared nfsclient resources do not get status monitor run against them if they are referenced multiple times. This was originally posted by Birger Wathne on the linux-cluster mailing list. In the config file below, only 5 of the 9 exports below get a status monitor, export1fs:h-nfs-root export1fs:h-nfs-insecure-ro export1fs:nis-hosts-ro export2fs:h-nfs-insecure export2fs:nis-hosts The rest do not get checked since they are the second reference to the shared resource. Relevant cluster.conf section--- <resources> <fs fstype="ext3" name="export1fs" mountpoint="/exports/export1" device="/dev/LocalGroup01/Export1Vol00" options="acl"/> <fs fstype="ext3" name="export2fs" mountpoint="/exports/export2" device="/dev/LocalGroup01/Export2Vol00" options="acl"/> <fs fstype="ext3" name="export3fs" mountpoint="/exports/export3" device="/dev/LocalGroup01/Export3Vol00" options="acl"/> <nfsclient name="nis-hosts" target="172.19.30.31" options="rw,sync"/> <nfsclient name="nis-hosts-ro" target="172.19.30.32" options="ro,sync,no_root_squash"/> <nfsclient name="h-nfs-root" target="172.19.30.33" options="rw,sync,no_root_squash"/> <nfsclient name="h-nfs-insecure" target="172.19.30.34" options="rw,sync,insecure"/> <nfsclient name="h-nfs-insecure-ro" target="172.19.30.35" options="ro,sync,no_root_squash,insecure"/> </resources> <service name="nfssvc"> <fs ref="export1fs"> <nfsexport name="export1fs"> <nfsclient ref="h-nfs-root"/> <nfsclient ref="h-nfs-insecure-ro"/> <nfsclient ref="nis-hosts-ro"/> </nfsexport> </fs> <fs ref="export2fs"> <nfsexport name="export2fs"> <nfsclient ref="h-nfs-root"/> <nfsclient ref="h-nfs-insecure"/> <nfsclient ref="nis-hosts"/> </nfsexport> </fs> <fs ref="export3fs"> <nfsexport name="export3fs"> <nfsclient ref="h-nfs-root"/> <nfsclient ref="h-nfs-insecure"/> <nfsclient ref="nis-hosts"/> </nfsexport> </fs> </service> Version-Release number of selected component (if applicable): How reproducible: Always Steps to Reproduce: 1.Reference the same nfsclient shared resource in multiple filesystem nfsexport Actual Results: only 5 monitor commands are executed, instead of 9. Expected Results: All 9 exports are monitored. Additional info:
Created attachment 116552 [details] Proposed fix The proposed patch makes a copy of the resource action list in each resource node that references a resource. It then uses that list when checking the status, so each use of the resource has it's own last checked timestamp for each action.
Patch is also here: https://www.redhat.com/archives/linux-cluster/2005-July/msg00090.html Discussed in this thread: https://www.redhat.com/archives/linux-cluster/2005-July/msg00045.html
I received an email from Birger, stating that the patch did fix the nfs export monitoring problem on his cluster.
Excellent. Putting in STABLE, RHEL4, and HEAD.
I ran a regression test to ensure that there was no memory leaked; there wasn't. Patch is definitely acceptable.
Patches in.
An advisory has been issued which should help the problem described in this bug report. This report is therefore being closed with a resolution of ERRATA. For more information on the solution and/or where to find the updated files, please follow the link below. You may reopen this bug report if the solution does not work for you. http://rhn.redhat.com/errata/RHBA-2005-738.html