Backup Copy + WAN: Creating Fingerprints All The Time

[Technogod]

I have a backup that every time I turn around it's creating fingerprints. Last night there was a disconnect with the WAN and here we go again. I've started fresh with a seeded backup. Still have the same problem. Is anybody else experiencing this type of behavior? Case #04586599

[HannesK]

Hello,
I saw that you ended up in the free product cases support queue while working for a large company... I suggest to open the case with a valid support ID and support permissions.

for me it sounds like expected with unstable connections. But your description of the problem is very short and probably I did not fully understand your setup. As you did not provide any additional information in the case, chances are low that someone can guess your problem. I recommend to give support a more detailed description about your environment (repository type, backup job type, bandwidth, amount of data, how you seeded).

Best regards,
Hannes

[Technogod]

This is a NFR license. I attached the logfile to the support case. Yes, it probably is network instability but the way B&R handles it is the problem.

[Dima P.]

Technogod,

To clarify, fingerprints are for Linux machine and it gets automatically added to the trusted list, right? Thanks!

[Technogod]

No, it's a Windows 10 machine. I deleted the Backup Copy. I had the original 14 day old seeded copy still on the Server. I tried to use it but had the same result.

[Egor Yakovlev]

Can you please be more detailed about those created fingerprints you are talking about.
Maybe a screenshot or description? Are you referring to Backup Copy Job + WAN Accelerator creating digests?

Uploaded logs are full of connection issues with WAN Accelerator service, and it seems you are using public IP there, isnt it? It might be best to go through some sort of VPN instead, as backup processes require a list of ports to be opened...

/Thanks!

[Technogod]

Backup Copy Job + WAN Accelerator creating digests. Why is it so difficult to recover when there is a connection issue? It took 10 hours of data transmission to fix the problem. I've uploaded the logfile.

[Egor Yakovlev]

Well if there is no connection between sites, it does make it harder to transfer data...
It wasn't 10 hours of data transmission, but rather 10 hours of trying to establish connection over the broken link.
Anyhow, please keep working with support to find the root of the drop.

/Thanks!

[Technogod]

It was the WAN Accelerator. B&R was unable to recover properly anytime there was a disconnect with the ISP. I've seen where other users stopped using the WAN Accelerator for the same reason. Once I selected Direct. My problem disappeared.

[Technogod]

There were no disconnects and it still failed. Not to mention it took 10 hours to create fingerprints before it failed. I disabled WAN Accelerator and it was done in 3 hours. Goodbye WAN Accelerator!

[Egor Yakovlev]

Well, quick glance into logs does show service connectivity problem for the entire duration - however I am not sure of the nature of it(like dropped networking or a service failure).
WAN Accelerator does bring a lot of value for offsite data transfers and if you would like to utilize it, feel free to open another ticket - I am 100% positive support will find the root of WAN Accelerator failure.

/Cheers!

[Technogod]

I have experienced connectivity issues during a three hour period from 3pm to 6pm. I'm still trying to find the root cause. The backup copy jobs run at night. I have had no failures and all jobs have run successfully after switching to direct.

[Technogod]

This is the second support ticket I've opened on this issue. I have deleted the WAN Accelerator. Deleted all files in the VeeamWAN folder. Recreated WAN Accelerator and populated Cache. I still have the same problem with failing during creating fingerprints. This seems to occur on large Windows Agent Backups. Today there was more data transferred creating fingerprints than the backups themselves. I have deleted the WAN Accelerator and refuse to use it. Case #04617548

[Dima P.]

Technogod,

Thanks for sharing, I've asked QA team to review the case details. I'll update this thread once I hear back anything. Cheers!

[soncscy]

> I have deleted the WAN Accelerator and refuse to use it.

I promise not snarky, but then what's the point of pursuing it further?

Software WAN accelerators are tricky as I see it because they're based on the local cache, and you're battling not just with the algorithm, but the OS + hardware hosting the cache.

Cache rebuilds for software WAN accelerators are not unusual -- for your post there, sending more data after a rebuild makes total sense -- the WAN accelerator has no idea what data is unique vs already in the cache, so it can't possibly reduce WAN traffic. And in traditional models for such acceleration, if either (or both!) caches have a reason to believe their data is suspect, then the cache needs to be ignored and not trusted!

This is just common sense -- if the choice is more data vs. corrupt data, more data is always preferable. If you're having connection issues, you need to fix those; network connections at the application level should "not" be forced to accommodate issues at the network or hardware level.

If it were my company's infrastructure, we'd be plunging with pcap right about now to understand where these network issues originate from.

[Technogod]

Dima P. You can close the case. I fixed the problem. There was a software package on the Router that was intermittently causing the Internet disconnects. The SUPERCOMPUTER4 backup was corrupted. I created a new backup. I am currently running the new Backup Copy Job using the WAN Accelerator. My other Backup Copy Jobs are running properly using the WAN Accelerator.

[Dima P.]

Glad you nailed it and thanks for sharing the update!