We recently upgraded from v10 sp14 to v11 sp6 and that included migrating to a new CommServe.
We engaged Professional Services and the person assigned to our project performed a full v11 install and then copied in our DB from v10.
Below are some pain points we experienced and some advice to prevent or shorten the impact.
After the upgrade v11 somehow determined my Active Directory account changed and any objects I created in v10 were now listed under (Deleted) in v11 . Scheduled reports under my (Deleted) account would not run. WorkFlows(WF) triggered by an Alert failed all because of the Deleted account status.
My recommendation before you upgrade to v11 is to run the ChangeScheduleOwner script and change all productive schedules to Admin. If you have scheduled WFs be sure to check the ownership was changed with the script. We also had a problem running scripts in v11 (because of GPOs) so I logged in as admin and recreated the schedules both for Reports and WFs but the ChangeScheduleOwner script should work if you run it pre upgrade.
I could not find a similar script to change ownership on Alerts so as a workaround I deleted any Alerts configured to invoke a WF and recreated them under Admin.
In an effort to clean up the Deleted account we ran 2 scripts, ChangeScheduleOwner and later FixDuplicateADUserEntry. The FixDuplicateADUserEntry did clean up my Deleted account issues but it caused any WFs formerally owned by my Deleted account to fail. Actually the WF would show up in the Job Controller with the status of Queued but when you viewed the log file you could see errors such as “UserModel: ERROR: could not find user information for id ”. You could not kill the queued WFs. I tried stopping and starting the WF engine but the queued jobs remained. Opened a ticket with CV support and the resolution was to have a different ID take ownership of the WF, transferring WF ownership is an option in v11, then redeploy the WF. I logged on as admin and changed ownership on all productive WFs and redeployed. Stopping all CommVault services on the CommServe flushed / removed all the queued WFs from Job Controller.
We have Group Policy Objects (GPO) – CommVault updates the PATH environment variable but our GPO sets it back whenever the CommServe reboots so we had to have CV Support tell us how the PATH variable should be set, update the GPO, and then reboot the CommServe. We have had to set the GPO and reboot several times to get it correct as it seems CommVault was particular on the order of the paths.
I originally thought this was a WorkFlow issue but it turned out to be a missing registry key. I could not get any SQL restores to work and after opening a critical ticket was informed the registry key bEnalbeSQLSnapRestoreStreams is required but was not included in the v11 sp6 download.
WorkFlows – In v10 if WorkFlow A called/executed WorkFlow B any values returned by any of the scripts in WorkFlow B could be accessed in WorkFlow A. This is no longer the case in v11, you need to assign Outputs in the WorkFlow. Only the declared Outputs of the WorkFlow will be available to other WorkFlows. I spent approximately a week modifying our v10 WFs because of the change to accessing WF output.
On a positive note, there are nice options I have found for WorkFlows in v11. There is a syntax check and debug option in the designer. I also discovered if you individually pick a WF and export it, it will append the version number to the file name. I was doing that manually in v10 so this is a welcomed enhancement.