MS15-078 windows update issues

[mswei]

Hello, recently applied https://technet.microsoft.com/library/s ... /MS15-078/

Per the update, I needed to shut down my machine to complete the installation. I clicked Start > Install updates and Shutdown, and the update process started. When I checked the backup this morning, I noticed that a backup had failed around the time that I installed the new update, and then ran normally at its scheduled time. During the scheduled time, it took a full backup instead of an incremental.

Has anyone else experienced this? I have already opened a support case, just wanted to see if the community has any shared experience with this. Thanks!

Veeam Support - Case # 00975046

[Dima P.]

Hello Michael,
That seems to be unexpected (unless the volume structure has changed, which I believe is not the root cause). To tell the truth, I suffered from this update this morning too – my Endpoint tray component got disconnected from the service. Lets await for the log investigation results.

[mswei]

Thanks Dima. For what it's worth, the endpoint backed up successfully on the 18th of July (first full of scheduled rotation), taking the expected incremental backup at the scheduled time each day until this patch last night.

I only discovered this before work today, so I haven't had a chance to try a restore yet and see what is available for recovery.

[Dima P.]

so I haven't had a chance to try a restore yet and see what is available for recovery.

Can you run a quick test and perform a file level recovery to check the backup file to be operational?

[mswei]

Here's a screenshot of FLR wizard with all the backup jobs so far:



The good news is that I can restore files from the old and most current fulls, as well as the incrementals. Any ideas how this would handle a volume restore? Will all points still be valid? I am hoping to backup 2 weeks of data before the schedule starts over, but after doing the math with two full backups on disk, I won't have the necessary space (backing up to external USB hdd).

[Dima P.]

The restore points look good and valid – no worries. The said part is - it looks like you have to delete old backup files manually in case the previous chain is no longer rotated by the Endpoint.

[mswei]

So something broke the chain. Can we determine what broke it? Was it the shutdown event without pause for backup to finish or something else perhaps that can be avoided in the future? I am assuming at this point its my fault

Just to confirm, deleting the Saturday Full would make the incrementals from Saturday, Sunday, Monday, and Tuesday invalid, correct? I'd *like* to keep them around until I get a few more incrementals under the Wednesday full, but don't want to mess up the new chain either. I guess what I'm dancing around here is: do I need to start from scratch or is any of this salvageable?

Thanks for your support Dima!

[Dima P.]

Can we determine what broke it?

QA team is just about to do it

deleting the Saturday Full would make the incrementals from Saturday, Sunday, Monday, and Tuesday invalid, correct?

Yes, deleting the full backup (.vbk) or metadata file (.vbm) makes the incremental files (.vib) completely useless. If I were you – I’d just keep the whole chain for the needed period of time and then delete all the files at once.

[mswei]

If I were you – I’d just keep the whole chain for the needed period of time and then delete all the files at once.

Does that mean I can keep the full from early this AM as my new full or should I start a new job?

[Dima P.]

Yup - you can keep the latest full.

[mswei]

So I've kept the latest full and removed the previous full and it's associated change. Is there a way to update the metadata so that it no longer thinks those backups exist or do I just need to wait until they're rolled out (and suggest a feature for the next release)

[Dima P.]

The .vbm (metadata file) should be automatically updated.

[mswei]

Doesn't appear to be updating automatically. I have rebooted since, and backup has occurred multiple times since the instance where I deleted the full and incremental chains from the drive, but they still showup when attempting to restore a full volume.

[Dima P.]

Michael,
The root of the issue was indeed the shutdown caused by update – we are going to address this issue in 1.1 upgrade. Regarding the .vbm I am checking with the QA.