Unsolved
This post is more than 5 years old
19 Posts
0
8828
Avamar 7 and replication timeout setting
I searched ECN but I could not find an Avamar 7 topic that covered this, sorry if I missed it. I'm trying to modify the default setting of 72000 for the timeout on a replication job that was setup in the Avamar GUI/Replication window. In previous versions of Avamar this could be done by modifying files in /usr/local/avamar but I cannot find the repl_cron.cfg file on this version 7 grid. Is there a different file that needs to be modified to change the default timeout or a way to override the timeout?
ionthegeek
2K Posts
0
November 21st, 2014 12:00
For plugin-based replication, you can update the schedule associated with the replication job to change the timeout.
ScottDB1
19 Posts
0
November 21st, 2014 12:00
I could see the 72000 in the settings when I viewed them (advanced options or settings), but there wasn't a way to change the value, only add other attributes. Is that the correct way? Add an attribute/setting ([replscript]timeout = somenumber, currently is shows [replscript]timeout = 72000) with a new, higher value?
ionthegeek
2K Posts
1
November 21st, 2014 13:00
Okay, I had a chat with the support guys. It is possible to enable overtime but it is not very intuitive. If you open the Policy pane and go into the /MC_SYSTEM domain, you can enable overtime on the client inside to enable overtime for replication. According to the gentleman I spoke with, this works starting in Avamar 7.0.1. This setting enables overtime for all replication jobs GLOBALLY so use it sparingly.
I was not able to confirm whether "Overtime until successful backup" works or not but you can force replication to use overtime by setting the dropdown to "Always allow overtime".
I would only recommend enabling overtime for initial replications.
ScottDB1
19 Posts
1
November 21st, 2014 13:00
I was going through option b, but it seems the max is 24 hours. I'd like to let the replication run for several days if possible - at least initially.
ScottDB1
19 Posts
0
November 21st, 2014 13:00
Thank you as always for your responses. Is what I'm asking for unreasonable? I know normally that in cases where there are large data sets to replicate you can set the systems up next to each other and "seed" the destination but that wasn't possible or planned in this case. Any other recommendations? One I can think of is to find the offending large client and run replication from the CLI on a client by client basis when the scheduled job isn't running. Just a thought but I would like your thoughts or recommendations.
ionthegeek
2K Posts
0
November 21st, 2014 13:00
My pleasure! I don't think what you're asking for is unreasonable. One other possible approach you could take is to put these large clients into their own replication stream(s), that way they don't interfere with the other replication jobs or vice-versa. You'd still be limited to 24 hours unless you enabled overtime but you wouldn't disrupt the regular pace of replication for the regular replication stream. You could also run the replication from the CLI but I usually recommend against that unless there isn't another option -- it's a bit fiddly and mistakes can be hard to correct.
ionthegeek
2K Posts
1
November 21st, 2014 13:00
Unfortunately I don't see a way to enable overtime in the Replication configuration. Fortunately, the next time replication starts up, it will "catch up" any new backups for the clients it has already processed, then resume where it left off.
I'm going to follow up with some of my contacts on the support team to make sure I haven't just missed something. If it's not possible to enable overtime for replication in the current version of the product, I'll submit a request for enhancement on your behalf.
ionthegeek
2K Posts
0
November 21st, 2014 13:00
It sounds like you're looking at the dataset Advanced options.
You need to either a) Change the replication group to use a different schedule:
or b) Click the Modify button to edit the schedule:
ScottDB1
19 Posts
0
November 21st, 2014 13:00
I received your other reply, sorry I don't know how to reply to that specific email, it's not in this thread. Would you have examples of where to make the changes. Like you did earlier in this thread.
Avamar Exorcist
462 Posts
0
November 24th, 2014 00:00
There's a KB here on this topic although I believe it's still pending final publishing before it will be globally visible at http://support.emc.com. When published, the article will be accessible with the following link:-
192507 - How to configure Avamar plugin based replication to allow clients to run longer than 24hrs (overtime)
ionthegeek
2K Posts
0
November 24th, 2014 06:00
Sorry about the delay, I didn't have a chance to get back to this on Friday. I generated some GUI screenshots. I've had to blank out the name of the system but it will be the same as the Avamar server name.
Open the Policy pane, select the Clients Tab, then MC_SYSTEM, then select the client and choose Edit:
From there, you can change the Overtime setting:
ScottDB1
19 Posts
1
November 24th, 2014 12:00
Ok, that's what I did. Thank you for confirming.
mosohail
35 Posts
0
January 5th, 2016 23:00
Hello, There is another option which you can enable during the configuration of the replication group. Go to option number 6 overview and click on more options----- then more----- then choose 8 concurrent streams---- However you can also enable the option of starting the replication with the clients that didn't started in the last replication as an option. kind regards, Mohamed