I have an Avamar currently backing up a VMware environment and just upgraded from 7.0 to 7.1 today. I have three clusters with a lot of horsepower and two standalone hosts with, well, not so much in the way of horsepower.
In my old 7.0 environment, the two Avamar VMware proxies were limited to two simultaneous jobs at once. The Avamar proxies in my clusters were allowed to do the full 8 simultaneous jobs.
I'm doing a bit of reading and it seems as though I can't replicate this behaviour in version 7.1 What I'm seeing is that I'd need to edit the mcserver.xml file on my Avamar grid to make a global setting that bumps the number of jobs down to 2. Is this actually the case? Or is there a way that my two standalone hosts can be backed up by 2 simultaneous jobs, while my clusters take advantage of the full bandwidth afforded to me?
There is a basic difference between the Avamar 7.0 and 7.1 proxies. When we register 7.0 proxy we get 8 instances like proxy-1 to proxy-8 and each instance runs a backup/restore job. Starting from 7.1 we only have one instance which can run 8 jobs.
So my suggestion is after upgrading to 7.1 to avoid any problems is to deploy fresh proxies.
Also i don't seem to recall in 7.0 proxy being limited to 2 jobs (i think back in some version of 6.x we were limited to 2 jobs per proxy.)
Yes, this is what I did. I removed all my old 7.0 proxies. In two of my proxy VM's I reduced the number of actual proxies down to proxy-1 and proxy-2. proxy3 through proxy-8 were all disabled.
The remaining proxy VMs all had proxy-1 through proxy-8 configured.
I don't have the storage bandwidth on these two small standalone servers to run all 8 jobs at once.
OK if I understand correctly we basically want to restrict the number of simultaneous jobs on the standalone esx hosts to 2.
If that is the case you can achieve by doing something like dedicating a proxy for each esxi host and then create many groups
each with just 2 clients and stagger the groups enough so that only 2 jobs run at a time.
since the number of jobs per proxy is a global setting this might be a way to tackle.
Not very efficient I agree.
That's funny. I actually thought of this yesterday and implemented it before my backups last night.
I did have a bunch of issues in last night's backups that I'm working through.
Agreed, it's not really efficient. Unfortunately, unless I can fine tune the number of threads the proxy service can run within the VM proxy itself, it's what's going to have to do.