DATE: July 11, 2023
SUBJECT: Possible defect in HA configuration leading to data corruption
VERSIONS AFFECTED: All versions prior to 23.4.8 utilizing High Availability
It has come to our attention that some customers may have their HA nodes configured in a way that would under certain circumstances allow both nodes to perform a pool import concurrently, a condition that could lead to data corruption and potential data loss.
We are asking that all customers perform the below procedure on both cluster nodes as soon as possible to ensure that this configuration is corrected.
- Extract the attached zip file and upload the fixcachefile.sh script to your BrickStor into /var/tmp using SCP.
- SSH to the node and cd to /var/tmp using PuTTY or another SSH client.
- As root, make the file executable
# chmod +x fixcachefile.sh
- As root, run the file
- The command will provide output of "update" for each pool.
Please email firstname.lastname@example.org to acknowledge that you have performed this procedure or if you would like interactive assistance.
As of the time of this bulletin, the 23.4.8 version has not been released. We recommend that the above procedure be performed immediately. This procedure is non-disruptive and does not require a maintenance window.