Archiver Proxy Thread

[damogill]

Hey

Are the O365 proxies limited in their ability to consume resources, single threaded maybe? We have a 4 vCPU VM with 8Gb RAM and the archiver thread appears to be capped to 4GB RAM and 25% CPU, which is 1/4 the available capacity on the VM.

[Polina]

Hi Gill,

VBO proxies are multi-threaded and you can always check the number of threads in the proxy settings and adjust them as needed (by default, it's 64). Resource consumption depends mostly on the load and operations performed by a proxy. For example, if you're backing up a single mailbox with just a few folders, you won't see any significant load on CPU or memory.

[damogill]

OK thanks, very stange that this process appears to be capped at 25%. We have 16 auxillary accounts backing up a 1TB sharepoint site, i'd expect more resources to be used.

Do you know if adding additional auxillary accounts requires a restart of the backup job? or can it dynamically adjust.

[Polina]

It's adjusted dynamically, no need to stop any jobs.

[damogill]

Thanks. We are getting item backup errors which when you investigate each item the permissions are set correctly. We get things like these, however if we clear the repository and re-run the job we get different errors. Is there any way to get extended support for a trial key? The customer is not gonna buy this with these issues occuring, I need some help to resolve them as I can't see why the errors come and go, they are not consistent.

The remote server returned an error: (401) Unauthorized. :: 36:09:43
Item not found in the database (ID: 8ea01cb2-dd6a-4b66-ad0d-d26987bf0e8d, OwsHiddenVersion: 11542). :: 19:42:55
The remote server returned an error: (500) Internal Server Error. :: 47:39:35
The remote server returned an error: (400) Bad Request. :: 76:09:07

[Polina]

Hi Gill,

Evaluation licenses come with support included, and you can open a case if you're within the 30-days trial period.

In general,
- the 401 error may indicate authentication issues for one of the additional backup accounts (in the logs, you'll see something like "Warning: Authentication failed for the account: username@yourorgname.onmicrosoft.com")
- the (500) Internal Server Error often gets fixed during job retries
- the (400) Bad Request can occur, for example, when there're some issues with the item itself and VBO cannot access it.

[damogill]

I've logged a case with support, the issues are random and dont appear to make sense when you investigate them manually. I.e. the account definately has access so the 400 error should not of occurred. What appears to happen is when the job fails, the next jobs is incrimental but only includes files that were previously backed up by the full - it essentially ignores backup up any files that may have came after the last error on the full. Which means the backup set is useless. I can confirm this by the fact that subsequent incrimentals say completed quickly and lots (hundreds of Gbs) of data is missing from the backup setup.

Hopefully support can get to the bottom of it.