Bug 592126
Summary: | [LSI 5.7 bug] RHCS node gfs failed to mount the filesystems during node bootup [LSI CR178792] | ||
---|---|---|---|
Product: | Red Hat Enterprise Linux 5 | Reporter: | hong.chung |
Component: | cman | Assignee: | Christine Caulfield <ccaulfie> |
Status: | CLOSED CANTFIX | QA Contact: | Cluster QE <mspqa-list> |
Severity: | medium | Docs Contact: | |
Priority: | medium | ||
Version: | 5.5 | CC: | abdel.sadek, abdel.sadek, cluster-maint, coughlan, dl-iop-bugzilla, edamato, mbroz, rwheeler, swhiteho |
Target Milestone: | rc | Keywords: | OtherQA |
Target Release: | 5.7 | ||
Hardware: | ppc64 | ||
OS: | Linux | ||
Whiteboard: | |||
Fixed In Version: | Doc Type: | Bug Fix | |
Doc Text: | Story Points: | --- | |
Clone Of: | Environment: | ||
Last Closed: | 2011-05-25 13:54:33 UTC | Type: | --- |
Regression: | --- | Mount Type: | --- |
Documentation: | --- | CRM: | |
Verified Versions: | Category: | --- | |
oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
Cloudforms Team: | --- | Target Upstream Version: | |
Embargoed: | |||
Bug Depends On: | |||
Bug Blocks: | 557597 |
Description
hong.chung
2010-05-13 22:55:22 UTC
One possible cause of this could be that that the clustered disks take some time to appear on the system, so that clvmd starts up before they are all available. I suppose one way round this would be to put an init script up before clvmd that waits for the drives to be available. clvmd can't wait for disks because it does not know in advance which disks it needs. can you provide steps on how to implement the init script? Just something as simple as a sleep command in a script should be enough to prove the point. Create a script such as #!/bin/sh sleep 30 And then put it i na file in the right directory so that it runs at startup eg: /etc/rc.d/rc3.d/S25waitfordisks and make it executable chmod +x /etc/rc.d/rc3.d/S25waitfordisks This is just a quick proof-of-concept and you might have to tweak the delay a few times to get a good compromise between the disks appearing, and a longer startup time. Chrissie We tried the suggested script and so far, the nodes are consistently remounting the gfs. will there be a fix in RHEL 5.6? Abdel: We are still discussing this internally. Stay tuned. The 'needinfo' is not on LSI. Deferring to RHEL 5.7 since we don't have a solution in place and a change of this size/complexity would need to make the Beta development window, which we've already passed for RHEL 5.6. Sorry for the delay in responding. Is the storage iSCSI? If so, have you specified _netdev in /etc/fstab ? If you would, please re-create the problem (remove the start-up delay), and then post the resulting /var/log/messages output. Let us know which "sd" devices comprise the the GFS clvmd volume, and are apparently the ones being configured too late. Tom Closing as per comments #15 and #17. If there is still an unresolved issue, please reopen this bug. |