Hello list(s),
First of all, sorry for crossposting but I think this is both a
question for -users as well as -hackers. I hope I get some response
since I've been through the archives so thoroughly I get a headache
only from thinking of them...
This is the setup of Amanda in our case:
The amanda backup server is called `backup' and is connected to the
backup client called `mobilkom' via a 100Mbit/s crossover cable.
Both hosts run the services needed for backup operations (amindex,
amidxtape, amandad, etc, etc). We have set the server up for tapeless
operation, using a `small' 600GB IDE disk array instead of tapes. Dumping
works really well, both with dump and with tar. The `tapes' get cycled
through the chg-multi script and all is well.
Both hosts run Amanda 2.4.4p1 and both on RedHat Linux 8.0 with kernel
version 2.4.18.
But a backup is worth nothing without the ability to restore it, hm?
Here begins the problem:
Using amrecover from `mobilkom' to access the index/catalogue on the
backup server (`backup') works fine, we can select and add files for
extraction and recovery. But once the point is reached to actually
extract the files from backup back to mobilkom the connection stalls
and eventually dies out after hours.
Here's what we do:
--------------------
mobilkom% amrecover -C normal -s backup -t backup
AMRECOVER Version 2.4.4p1. Contacting server on backup ...
220 backup AMANDA index server (2.4.4p1) ready.
200 Access OK
Setting restore date to today (2003-07-23)
200 Working date set to 2003-07-23.
200 Config set to normal.
200 Dump host set to mobilkom.
Trying disk /d ...
Trying disk md0 ...
$CWD '/d' is on disk 'md0' mounted at '/d'.
200 Disk set to md0.
/d
amrecover> ls
2003-07-3 testdir1/
amrecover> add testdir1
Added dir /testdir1 at date 2003-07-2
amrecover> extract
Extracting files using tape drive file:/dump/normal03 on host backup.
The following tapes are needed: normal03
Restoring files into directory /d
Continue [?/Y/n]? y
Extracting files using tape drive file:/dump/normal03 on host backup.
Load tape normal03 now
Continue [?/Y/n/s/t]? y
[here we wait for hours]
-------------------
And here it sits forever until it dies. tcpdump shows that backup
tries to reach mobilkom and awaits ACK but gets nothing and only
retransmits the request packets. mobilkom eventually ACKs something,
but definately nothing concerning the ongoing connection. I will attach
the log output of tcpdump, starting at the moment amrecover started on
mobilkom.
The point is that we need amrecover working, or else the system would
never be able to restore 300GB worth of user data in the event of any
kind of emergancy, since extracting the files on backup itself and
then transferring them to mobilkom is not an option.
Can somebody give us any input? This does not seem to be a frequently
encountered problem, or my archive-searching abilities and google have
died, which is not the case (hopefully).
Any help is greatly appreciated. Really.
Thanks in advance!
PS: I think this problem is related to the one discussed in this message
(from the NetBSD-current mailing list though):
http://mail-index.netbsd.org/current-users/2002/07/08/0002.html
He there has the problem that the TCP connection times out via the
lookback interface while it works when doing amrecover remotely. Our
problem here goes vice versa.
PPS: If any other logfiles are needed just tell me and I'll attach them,
we have lots of them here (amidxtaped.timestamp.debug, etc).
--
Andreas "ant" Ntaflos | "A cynic is a man who knows the price of
ant AT overclockers DOT at | everything, and the value of nothing."
Vienna, AUSTRIA | Oscar Wilde
tcpcump.030723_124716.out
Description: Text document
|