(First of all, can anyone comment on why, when creating a post, I am unable to select 'version' as 2015; unable to select 'operating system' as Windows Server 2012; nor can can I indicate v-ray edition?)
We have some large backups that need to run reliably over the weekend. They are virtual Full backups (Incrementals during the week) and if snapshots are left open on some of the servers it will cause real performance issues on Mondays. One of these backups is so large it needs most of the weekend to run.
This weekend has been a bit of a disaster. I checked to see if the backups were running smoothly on Saturday (I don't work on Saturdays, not properly), to find that Backup Exec won't start. It says the server needs rebooting first. I think it may have pulled through an update, great! So 16 hour delay to the start of backups. My fault I suppose, I discovered the server is set to automatically install updates on Fridays at 15:00. I changed that to 'alert' me, and ticked the box for the system to email me against all the LiveUpdate types of alert. So assuming it was an update that necessitated a restart, that won't happen again.
Having lost time, I needed things to run smoothly. My experience with Backup Exec Services is that they are just not reliable. I typically have problems at least once a fortnight, sometimes much more.
Checked the backups again today, and find they've all frozen, several of the largest at well over 50%. I restart the BackupExec services. For another post perhaps - but can anyone tell me why the 'restart all services' feature is so incapable? You are only going to use it if you have a problem with the services - so it needs to be able to cope with hung services. But it can't. EVERY SINGLE TIME the 'restart all services' feature stops on the Job Engine (so I wait then use the Task Manager to kill it), then the other service restarts continue and stop again on the Remote Agent (again I have to use the Task Manager to kill that too), AND THEN the services will restart.
....at which point, the jobs were listed as 'recovered'. Which I've seen before.
Now our longest and most important jobs are scheduled to run first at the weekend (not surprisingly), but after service restart Backup Exec DOES NOT NECESSARILY start the recovered jobs first. It leaves them sitting there wasting precious time, while it starts other less important jobs, and then sits the recovered jobs stating 'ready, no writable idle devices available'. Frustrating. I suppose I should have 'held' all the many waiting jobs BEFORE I restarted the services, but that's fiddly if there are a lot of them. Incidentally I find 'hold' and 'cancel' seriously lacking. If the system is past a certain point (quite early on) in starting a job, it will just ignore your command, stating 'hold pending' while it carries on opening snapshots and doing all kinds of things you didn't want it to do. Meanwhile 'cancel' normally just doesn't work at all if a job is frozen (which is the main time I want to use it), it will just say 'cancel pending' forever, until you restart the services.
To my dismay, I saw that the 'recovered' jobs were starting at the beginning again. So I looked for the online PDF administrators manual and read up on the job recovery funtionality.
Does anyone agree with me - I don't think of this as 'recovery' and feel it has been badly named. I had a crucial 3TB job that had reached 2.43TB before freezing along with some others. I would love this to be genuinely 'recovered', but I am not happy to see it being started again at the beginning, and Backup Exec call that 'recovered'. Would it be too much to ask to develop a system that could start where it left off?
I will almost certainly have to cancel this job - Backup Exec has messed me around and lost its chance to complete this job and close the snapshot before Monday, so cancellation will be the only option. That means incrementals for another week. I would love to see a real 'job recovery' function, but as far as I am concerned THIS IS NOT IT!
there is an option but not for Virtual machine backups https://www.veritas.com/support/en_US/article.000076111
Workaround - In your case, you could have opened the Job Log for 2.43 TB Job and then editted the job to include only the VMs not backed up. Later could have changed back the selection to original
Thanks Gurv. It would be nice if there WAS an option for Virtual Machine backups, wouldn't it. Oh well.
Not sure I quite understand your workaround. This was a single VM (ours is setup with a job for each machine), although it did include multiple VMDKs. Surely you need the job to finish cleanly without fiddling with it to ensure that BackupExec's records aren't messed up and restores would happen smoothly? Anyway, can you edit the job log? Did you really mean that, or did you mean to edit the job?
I thought it was multiple VMs. If its single VM and backup job does this then its better to rerun the backup. Anyways, the backup should not hang so that needs to be checked. request you to open a case with support if this is a recurring issue.