I've encountered a few cases where policy based replication is taking days to push new backups over a second Avamar & DD pair. I understand that this could take time time to replicate, especially over distances or small pipes. But what can be done to assess where the bottleneck is, especially if the network between sites isn't fully utilized?
It also seems possible to run simultaneous replication jobs. If the networks don't look saturated can I have two or three running at the same time? That way it would, in theory, handle more than one client and backup at a time.
Please note I am an End User is this is my experience
yes you can have more than one Job/Policy
I have one Policy where I do all physical servers (so I select from the top Avamar domain, then exclude my VM domains)
then I have a second Policy that I just select my vm domains
they run at the same time.
I had to do this because the VM's are at the bottom of my Avamar domains and it was Never getting to the bottom
So I set up the two (because my Physical servers are on the Avamar grid and my VM's are on the DataDomain grid)
then I had to work on how many streams each needed and how long.
My Physicals replicate just fine in a few hours, but the VM's from the DD take a long time. I had to go to 23 hours a day, except for 2 days where I have it stop at 10 am so I can do retires, move, or other work.
But you can also just increase the number of jobs in you existing policy's I think by default it might be 2 can you increase that up to 8 (depending on your version of Avamar which you do not tell us).
Yes - it's in the "More Options" for the replication set itself. You can click on the "Edit" button to edit the replication set, the click on More Options. You can increase the number of simultaneous jobs to 8, as I have shown here.
J_H_
2 Intern
•
498 Posts
627
0
Posted February 26th, 2015 12:00
we recently upgraded the DD os and the Avamar and the VM backups from DD are taking longer than they used to.
After 3 months of talking the EMC about that... the only answer was to increase the time to run the replication for the VM's