Hi all, we see an issue with Commvault 9R2 (sp10) using a Linux i-data agent client on OES2SP3 x64 backing up (clustered) NSS volumes. For speed we, this case, do not want to use a NSS i-data agent, so we just leverage the Linux i-data agent.
The issue is that if we use the Linux i-data agent and the jobresults directory is also hosted on the same NSS volume that's being backed up we get an error on failures.cvf
"Could not open failures file [/media/nss/LBAPOA/galaxy/JobResults/2/268/Failures.cvf] error number [19, No such device]. The backup would continue but this would affect the ability of reattempting the failed file backup during next incremental backup. A full backup is recommended to back up all files. Check the following: 1. The account this product uses has permissions to the product's install directory. 2. Sufficient disk space is available for the Job Results directory. If the problem persists, please upload log files from CommServe, MediaAgent and Client and contact your vendor's support hotline with the job ID of the failed job."
If we (for testing only) move only the jobresult directory to an ext3 volume we do not see the issue.
Has someone seen (and hopefully fixed) this issue before? My guess it's a Commvault issue, but just before we open a SR with them I would like to now if there's a good guess to solve this:)
The issue is that if we use the Linux i-data agent and the jobresults directory is also hosted on the same NSS volume that's being backed up we get an error on failures.cvf
"Could not open failures file [/media/nss/LBAPOA/galaxy/JobResults/2/268/Failures.cvf] error number [19, No such device]. The backup would continue but this would affect the ability of reattempting the failed file backup during next incremental backup. A full backup is recommended to back up all files. Check the following: 1. The account this product uses has permissions to the product's install directory. 2. Sufficient disk space is available for the Job Results directory. If the problem persists, please upload log files from CommServe, MediaAgent and Client and contact your vendor's support hotline with the job ID of the failed job."
If we (for testing only) move only the jobresult directory to an ext3 volume we do not see the issue.
Has someone seen (and hopefully fixed) this issue before? My guess it's a Commvault issue, but just before we open a SR with them I would like to now if there's a good guess to solve this:)