Extremely Slow

[TheGoldyGopher]

I opened a topic here about two months ago. Unfortunately I need to revisit the problem.

 

I have upgraded to 7.7.203 and have installed the lastest patch and driver update.

 

Retrospect is the only service on a Windows Server with 4GB RAM, 200 GB locally attached disk and 2 SAN Drives attached via fibre.

 

I have 6GB available on c: and 150GB on D: (where all the catalog files are stored.)

 

I have 200 clients in 19 groups (based upon functional group in the business units.)

 

I run all proactive backups with a recycle backup once a month for disk space reasons.

 

In October I was backing up 150 to 190 clients a day using my proactive backs, today I have backed up 10 so far.

 

The problem appears to be in the polling:

 

In my Proactive Queue I have

1) Laptop is not here

2) Laptop is not here

3) Laptop is not here

4) Laptop is not here

...

11) User sitting at a desktop

12) User sitting at a desktop

..

 

The queue keeps polling the same 5 to 10 computers to see if they are here and ready for a backup. The problem is the 10 people who are in the far reaches of the globe have risen to the top and it keeps polling these people, not the other 200 plus sitting at thier desk.

 

I have set the Options to:

Check For Source every 5 Seconds

Client connect every 45 Minutes

Retry Failure after 60 minutes

 

It takes about 20 to 30 seconds to poll a computer, but after it gets between 5 and 10 computers into the list it returns to the top of the list.

 

When we had issues in Oct/Nov we cleaned up our clients removing about 35 clients that no longer existed. Which speed up up to about 125 clients a day. However we are down to 10 to 25 clients. My users are getting notices that they haven't been backed up in 10 days. I need to resolve this issue, any help will be appreciated.

Edited January 4, 2010 by Guest

[Richy_Boy]

I think your proactive settings have something to do with this..

 

I have set the Options to:

Check For Source every 5 Seconds

Client connect every 45 Minutes

Retry Failure after 60 minutes

 

I think the 'check source every' will poll the top machines again after 5 minutes. If you set this higher to, say, 60 minutes, it shouldn't re-check those top machines again for a while.

 

I think that's how it works anyway...

 

I put in a suggestion a year or so back that you could add weight to client settings to priotise certain machines, which would help in this instance. i.e. you could set the 'away' people to a low priority.

 

Rich

Edited January 5, 2010 by Guest

[TheGoldyGopher]

Changing "Check Source Every " to 60 minutes had no change on the performance.