Unsolved
This post is more than 5 years old
2 Intern
•
132 Posts
0
952
February 2nd, 2018 06:00
Avamar replication appears to have terrible performance for one client
Working with a customer, and I'm seeing an issue I have seen before, but not to this extent.
What appears to be a fairly "innocuous" Avamar backup is taking "forever" to replicate.
Avamar v7.5.0-183, Linux server backup.
From the output below, obviously it has been able to replicate previous backups, since it is now replicating "backup 6" - but it has been doing so for the past few days with progress numbers similar to the following output:
2018-02-01 14:26:45 avtar Info <6654>: Replicating backup 6, Label "BP Oracle DB Schedule-Bayport Oracle Database Se", 218.6 GB
...
2018-02-01 23:46:00 avtar Info <6090>: Restored 7.589 KB in 559.29 minutes: 834 bytes/hour (0 files/hour)
I believe the server does have a relatively high amount of files, and they may not be large, but I don't know whether the file distribution warrants such poor performance as the numbers above, at least over several days. I can see it happening on one day, and then getting better on a subsequent day (which I have seen on other clients in the past), but this client is a whole different story.
There are other backups replicating in the same policy, and they are making much better progress on larger backup sets.
Trying to figure out what the heck is going on. Have done some Support searches, but likely haven't been "asking the right question" so far.
All comments/feedback appreciated - thanks.
connolly
2 Intern
•
132 Posts
0
February 2nd, 2018 07:00
Just realized - this might have some bearing on the performance; still looking for feedback/comments relative to most of us being used to previous replication performance with other methods.
2018-02-01 14:26:58 avtar Info <40200>: id:1 VSR selected because all base files are available (container.1.cdsf)
Will be looking at VSR related KBs to refresh my memory - feel free to post links for others who might be viewing this discussion.
connolly
2 Intern
•
132 Posts
0
February 5th, 2018 11:00
Update - I'm seeing the same kind of poor performance even with the job says it has chosen "Range Replication".
I thought that Range replication was supposed to make replication faster? It sure isn't doing it for what now appears to be 3x clients, one of which is a file server client.
And while a couple of these are still replicating their initial backups, I'm not expecting performance in the TB range at this point - but I certainly wasn't expecting performance in the KB range either.