Amanda-Users

Repeating Level 0

2006-04-03 09:49:08
Subject: Repeating Level 0
From: Jim Summers <jsummers AT bachman.cs.ou DOT edu>
To: amanda-users <amanda-users AT amanda DOT org>
Date: Mon, 03 Apr 2006 08:46:05 -0500
Hello List,

I am experiencing an odd behavior from amanda. I have a client that I have been backing up for a couple of years now. A while back, I noticed that it was doing multiple level 0's on one DLE within one backup cycle. It stopped for a while but then began doing it again a couple of months ago. The pattern seems to be, when it does a level 0, then it will do another level 0 the next night, possibly a third one the night after the second. Then it will go back to incrementals. It does take a while to complete the level 0 and I am wondering if that may be throwing things off for the planner.

I checked it a little bit ago with amstatus and here is what it is saying for that host/dle:

192095k dumping   108864k ( 56.67%) (0:21:55)

Is the 21:55 21hours and 55minutes? If I run a second amstatus several minutes later the numbers don't change as expected.

I have looked through the sendsize and selfcheck on the client and did not spot any problems.

Still using amanda-2.4.5
gtar version = tar (GNU tar) 1.13.19


The dle for the affected client/filesystem:

vader  /var/ds          comp-high-sf

The dumptype is:
define dumptype comp-high-sf {
    global
    program "GNUTAR"
    comment "Non-root partitions with server compression"
    compress server fast
    priority high

The dumpcycle is 2weeks and runspercycle is 14.

Here is the planner/ponder from the current still ongoing run:

pondering vader:/var/ds... next_level0 13 last_level 0 (not due for a full dump, picking an incr level)
   picklev: last night 0, so tonight level 1

Previous amdump log:

pondering vader:/var/ds... next_level0 12 last_level 1 (not due for a full dump, picking an incr level)
   pick: size 235212 level 1 days 1 (thresh 20480K, 1 days)
   pick: next size 150864... BUMPED
  curr level 2 size 9617 total size 202729 total_lev0 0 balanced-lev0size 317923

And by the time the actual schedule is generated the ds filesystem is getting a level 0.

None of the other dle's for this client or others exhibit this behavior.

Any ideas or suggestions on what may be going on?

TIA
--
Jim Summers
School of Computer Science-University of Oklahoma
-------------------------------------------------

<Prev in Thread] Current Thread [Next in Thread>