PDA

View Full Version : planner ERROR, timeout waiting for REP



zekkerj
March 28th, 2007, 10:33 AM
Hello,

I'm using amanda 2.5.1p2 in several Suse Linux boxes, here, and a few days ago one of these boxes started to fail in amdump whith this error message:


FAILURE AND STRANGE DUMP SUMMARY:
<host> /media/nss/INTRANET lev 0 FAILED [disk /media/nss/INTRANET, all estimate timed out]
planner: ERROR Request to <host> failed: timeout waiting for REP

It sounds very confusing to me, specially when revising the disklist:


<host> /media/nss/INTRANET normal-tar
<host> /media/nss/INTERNET normal-tar

The error only occurs in "/media/nss/INTRANET", not in "/media/nss/INTERNET".

The same configuration worked very well for about 3 months, and is working, right now, in 6 other similar boxes.

ppragin
March 28th, 2007, 10:40 AM
1. Did the amount of data increase in /media/nss/INTERNET recently
2. What is the "etimeout" set to in amanda.conf
3. are you able to ping or connect the Amanda server from the client
4. have there been changes to the firewall between server and client recently
Pavel

zekkerj
March 28th, 2007, 05:05 PM
1. Did the amount of data increase in /media/nss/INTERNET recently
Yes, but the total amount of data in INTERNET is far below the capacity of tape (72GB DDS5). (*)


2. What is the "etimeout" set to in amanda.conf
No "etimeout" defined.


3. are you able to ping or connect the Amanda server from the client
Same server.


4. have there been changes to the firewall between server and client recently
Same server.


(*) Thanks a lot... I think I found the problem.
Recently, I moved two huge, complex, directories from another server into INTRANET. These directories take about 10GB, with hundreds of sub-directories (thanks God, only one level). I only remembered it when I tried to "du" INTRANET, and it doesn't came out, even after about 15min...

I think I'll have to fiddle with this "etimeout" parameter. :(