>Two letters just came in with the same Subject but (somewhat) different
>From: values.
i apologize. i was having much trouble with the list and tried
sending through egroups (yahoo).
>There are a couple of possibilities. One is that you've added some
>"disks" and the request (or corresponding reply) is now longer than a
>UDP packet (roughly 65 KBytes).
i did add a disk recently, but it is on a remote client, which
doesn't appear to be having problems.
>Look for "Amanda 2.4 REP HANDLE ...". Following that is the reply
>being sent back to the server. Copy from the next line to the "----"
>line to another file and see how big it is.
wc -c reports 2959 bytes, but as i mentioned in one of the two
messages i posted (again, sorry everyone), the problem only
seems to arise when actually backing up, as opposed to running
amcheck (which subsequently overwrites the amandad.debug file,
so the log only reflects the successful amcheck run).
>The other thing you can do is look through the reply and see if anything
>looks goofy. It should just be a bunch of entries like this:
>
> / 0 SIZE 606750
> / 3 SIZE 13030
> /opt 0 SIZE 690390
> /opt 1 SIZE 172
>
>(disk, level, "SIZE" and estimated KBytes)
i am not seeing anything that looks like this (a size report)
in any of the logfiles. sendsize.debug is still timestamped
with the last 'real' backup request time (early this morning).
this may be normal, however (ie, amcheck does not request
sizes).
btw, does amanda normally write sendbackup.debug
by opening and closing on each write? it seems that there is
only one entry (presumably the last) in the sendbackup.debug
log (which corresponds to the last successful backup a couple
days ago.)
>If the problem is the request/response has gotten too large, the current
>242 CVS source tree contains a workaround. You could either upgrade
>your server to that, or I can probably find the patch in the mailing
>list archive and you could apply it (assuming you're running 2.4.2p2).
i probably forgot to mention in the last *two* messages that
sendsize.debug says:
sendsize: debug 1 pid 12995 ruid 34 euid 34 start time Thu Jul 26 00:06:44 2001
/usr/lib/amanda/sendsize: version 2.4.1p1
REQ packet is bogus: bad level
sendsize: pid 12995 finish time Thu Jul 26 00:06:45 2001
which may just be a general parse error on an oversized packet,
if that is the case. unfortunately, i can't run amdump at the
moment, so i will have to wait until tonight's backup to get a
a look at the troublesome amandad.debug.
thanks john, i appreciate your help.
-c
______________________________________________________
Get Your FREE FlashMail Address now at http://www.flashmail.com
It's Free, Easy, & Fun !!!