CFSMount monitor timeout, offline or file system hang on VOM (Veritas Operations Manager) managed hosts.

Article:TECH140357  |  Created: 2010-09-20  |  Updated: 2010-11-05  |  Article URL http://www.symantec.com/docs/TECH140357
Article Type
Technical Solution

Product(s)

Environment

Issue



VOM (VERITAS Operations Manager) LDR discovery will issue "/opt/VRTS/bin/fsckptadm -l list". The command when issued on VERITAS Cluster File Systems may result in monitor timeouts, offline or cluster file system hangs.

 


Error



VCS ERROR V-16-2-13027 (<host>) Resource(vxfsckd) - monitor procedure did not complete within the expected time.

 


Environment



VOM versions: 3.0, 3.0 RU1, 3.0 RP1, 3.1

VERITAS Cluster File System 4.0, 4.1, 5.0, 5.0MP1, 5.0MP3, 5.0MP4, 5.1


Cause



The VOM agent invokes "/opt/VRTS/bin/fsckptadm -l list" during LDR discovery which will cause expensive cluster wide file system freeze through vx_local_statfset(). The freeze/thaw processing may cause the CVM agents to timeout.

This problem will affect VERITAS file systems that are cluster mounted and don't have any clones created.


Solution



  •  Symantec has released a hotfix for VOM 3.0 and 3.1. The VOM hotfix will not require server downtime.

VOM hotfix HF030040200-02
https://sort.symantec.com/patch/detail/4309

This hotfix applies to VERITAS Cluster File System hosts that are managed by VOM, and prevents VOM agent from invoking the "fsckptadm" command during LDR discovery that might cause some VCS agent monitors to timeout. The HF also contains a modified gendeploy.pl script which will also apply the HF before LDR discovery is run when host is being configured to the VOM central server. To ensure that the HF is applied, re-generation of the VOM attach / configure (gendeploy.pl) script will be required.

  • Release 3.1SP1 will permanently fix the problem when released. VOM 3.1SP1 will use vxlist to get LDR info instead of fsckptadm.

Note: The patch supercedes the 3.0 patches . Information about large and mounted file systems will not be displayed in the LDR report after deployment of this hotfix.

Wokaround:

When LDR discovery is disabled, no CFSMount monitoring timeouts are seen.

# /opt/VRTSsfmh/bin/mh_ctl.pl --family LDR --pause

The above commands will temporarily stop LDR discovery that is used for showing license reports in VOM.


Supplemental Materials

SourceDEFECT
ValueHF030040200-02
Description

VOM hotfix that prevents VOM agent from invoking fsckptadm command that can cause VCS monitors on CFS installations to timeout



Article URL http://www.symantec.com/docs/TECH140357


Terms of use for this information are found in Legal Notices