How does Endpoint Backup handle network outage?

[helman]

I didn't find documentation about this, but this is a problem, laptop users might face.

As backup is running in the background and many users don't check if a backup is running, they close the lid and go home. This means any operation VEB might do is interrupted without it being able to clean up properly.

How does VEB handle incomplete incrementals in the repo?
How does VEB handle being interrupted while integrating .vib's into the vbk?

How big is the chance of ending up with a corrupt repository?

[Mike Resseler]

Hi,

When they "leave" or any other reason to fail, the job itself will fail after a while. It is perfectly possible that there is a "leftover" VIB that is in bad shape, but that doesn't matter actually. We have the VBM file that tracks all of this. The moment a new backup can start and succeeds, you will have a new incremental, and the other one (the bad one) will disappear...

For the merge, we need additional storage when integrating and this is one of the reasons (if not the main reason) why we need this. As long if there is not the "final" job, you are always good and VEBF can recover.

Hope it helps

Mike

[LukeAd]

So if an in-progress backup fails (as in the scenario above where a laptop user walks out in the middle of a backup), the next backup that starts is a completely new instance? I just want to confirm that the failed backup doesn't restart and pick up where it left off.

The reason I ask is for first-time backups on remote users (e.g., at home with a 5Mbps effective upload speed) where we are wanting to backup their full machine. That first backup of 100GB (or more) is going to take 2+ full days to complete...

If they interrupt that first full backup at any time, they are going to need to restart from scratch with no way for the failed backup to pick up where it had left off, correct? If so, are there any tips/tricks/hints to get this to succeed on that first big backup (e.g., pre-seeding)?

We've just started testing Endpoint Backup and I'm using my own laptop as the guinea pig (I'm 1000 miles away from our data center and don't have a fat upload pipe). I'm not trying to integrate my Endpoint Backup with our B&R repository yet and am just trying to send it to a normal file server share (the reason I mention is because of the comments in this thread on pre-seeding the backup repo: http://forums.veeam.com/veeam-endpoint- ... 25466.html)

[Mike Resseler]

Luke,

No, there is no seeding directly from the UI, however, if you would take a backup to a removable device first, then ship it that 1000 miles away and store it on a file share. After that (no backup in between so disable the job on your machine) you can repoint your target to the file share and let the backup run again. VEB will detect that the first full exists on the new location (through the metadata file) and then start an incremental.

As stated in that topic you mention, it won't work with a VBR repository but between a usb device and file share it should work

Mike

[LukeAd]

Thanks for the confirmation.

Well I'm 19 hours and 21 minutes into my initial backup and am at 49%. One thing I must say is that it's been running like a champ and there have been no hiccups so far. Also I am amazed that, unlike other products I've used before, there seems to be no performance impact to anything else I'm doing on my machine. It looks like you guys nailed the throttling algorithm. So far so good with everything.

However I don't expect our other remote users to have the patience (or to remember) to leave their machine up and running for 2-3 days straight for their initial backups. Sending them a USB drive and hoping they set it up, run the backup, and ship it back is not highly realistic either (nor cost effective).

Do you know if a feature request has already been submitted for "restart interrupted backups from where they left off"? If not, do I just start a new thread in this forum to submit a feature request?

[Mike Resseler]

Luke,

I believe there is already one, put please create a new one... I'm sure that Dima P. will merge it into the existing one

Every vote counts and keep pushing

Mike

[Dima P.]

Hi Luke,

How does VEB handle incomplete incrementals in the repo?

In case of disconnect or under any other circumstances where repository becomes inaccessible backup job starts a retry cycle (restarts every 10 minutes during the next 23 hours) and will clean up the incomplete restore point once repository is connected. Same behavior for merge process

f so, are there any tips/tricks/hints to get this to succeed on that first big backup (e.g., pre-seeding)?

Mike is right, for local and shared targets you should be able to create a local copy (i.e. on internal drive or USB stick) and then move it to shared folder. Reconfigured job with the changed destination should be able to pick this restore point. Note that this not going to work with Veeam repository

no performance impact to anything else I'm doing on my machine. It looks like you guys nailed the throttling algorithm

Happy to hear it again and again. Thank you

Do you know if a feature request has already been submitted for "restart interrupted backups from where they left off"? If not, do I just start a new thread in this forum to submit a feature request?

Added you to the list