Amanda-Users

Re: large filesystem timeouts - is there a better way

2006-10-19 18:45:10
Subject: Re: large filesystem timeouts - is there a better way
From: Pavel Pragin <ppragin AT zmanda DOT com>
To: Steven McDonald <mcdonald AT triumf DOT ca>
Date: Thu, 19 Oct 2006 15:17:02 -0700
Hello,

You can add this line to the dumptype definition that you are using and this will help speed up the estimation.
estimate calcsize

Pavel


Hi

We have a couple of large filesystems that we would like to backup using Amanda. We are currently using it to backup several hundred other smaller systems very successfully, but the two systems that we have struggled to get working are the two largest and most important. Each file system is ~500GB

The error is the timeout,
ev 0  FAILED [Estimate timeout from
which I understand the meaning of, we have increased it (etimeout) several times, we are now up to 3hours and still getting a timeout. My question is the following, is there a way using amcheck or some other means to find out a head of time what this value is, for a particular client filesystem? Or should I just set it really large like 24hours and see what comes back as the estimate.

The next question than is if it does really take many hours just to get an estimate from these two system, Is there a better way to speed up the estimate for these large filesystems, is anybody else using amanda to backup directories trees as large as 500GB.


The Amanda server is 2.4.5p1 and the two clients are 2.4.3 and 2.4.4p1

Thanks in advance
Steve