[prev in list] [next in list] [prev in thread] [next in thread] 

List:       amanda-users
Subject:    Re: Strange bug on Debian buster, v10.3, segmentation fault on client doing amcheck
From:       Jose M Calhariz <jose.calhariz () tecnico ! ulisboa ! pt>
Date:       2020-06-30 11:18:55
Message-ID: 20200630111855.GA25922 () calhariz ! com
[Download RAW message or body]


On Mon, Jun 29, 2020 at 11:06:46AM -0600, Charles Curley wrote:
> On Mon, 29 Jun 2020 16:36:33 +0100
> Jose M Calhariz <jose.calhariz@tecnico.ulisboa.pt> wrote:
> 
> > On my main amanda installation I have a client that gives time out
> > when doing backups.  I have researched and checked out the most common
> > problems.  In the end I have found that:
> > 
> > - "amcheck Config -c client" gives 30 seconds of timeout.
> 
> I do not see that. I checked two clients, one AMD64, the other i386.

On mine main amanda installation and have dozens of Debian clients.
It is one client only that is failing and I do not understand why.

> 
> > 
> > - I do not have any clue on the logs at the client.
> > 
> > - Running the command by hand on the client I get segmentation fault.
> > 
> > /usr/lib/amanda/amandad -auth=ssh amdump
> > Segmentation fault
> 
> I see that, on both machines.
> 

OK, so the problem is another.  More research to do.  Thank you.
I will post here when I find more info.


> 
> > 
> > - Running the command inside gdb I see a NULL pointer.
> > 
> > gdb /usr/lib/amanda/amandad 
> > GNU gdb (Debian 8.2.1-2+b3) 8.2.1
> 
> I do not have symbols installed, so of course I can't examine the
> variables. I do see a segment fault:
> 
> ...
> (gdb) run -auth=ssh amdump
> Starting program: /usr/lib/amanda/amandad -auth=ssh amdump
> [Thread debugging using libthread_db enabled]
> Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".
> 
> Program received signal SIGSEGV, Segmentation fault.
> 0x00007ffff7f3a687 in stream_sendpkt () from \
> /usr/lib/x86_64-linux-gnu/amanda/libamanda-3.5.1.so (gdb)
> ...
> 
> I have no idea what is going on here, not having the source in front of
> me. But I wonder if this is because amanda is trying to use an SSH
> connection that isn't there?
> 
> The first time, I SSHed in as root, did an su - to backup, then ran
> gdb. To test my hypothesis above, I went to my amanda server, su - to
> backup, then sshed to the client. This time I did not get a seg fault,
> and used Ctl-c to end the process:
> 
> (gdb) run -auth=ssh amdump
> Starting program: /usr/lib/amanda/amandad -auth=ssh amdump
> [Thread debugging using libthread_db enabled]
> Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".
> ^C
> Program received signal SIGINT, Interrupt.
> 0x00007ffff76a27e4 in __GI___poll (fds=0x55555559e3d0, nfds=2, timeout=30000) at \
> ../sysdeps/unix/sysv/linux/poll.c:29 29	../sysdeps/unix/sysv/linux/poll.c: No such \
> file or directory. (gdb) 
> 
> 
> 
> This is with the bog standard Buster version of amanda:
> 
> root@dzur:~# pre amanda
> amanda-client	1:3.5.1-2+b2		amd64
> amanda-common	1:3.5.1-2+b2		amd64
> root@dzur:~# 
> 
> 

Kind regards
Jose M Calhariz


-- 
--

A ciência aumenta nosso poder na mesma proporção que diminui o nosso orgulho

--Claude Bernard


["signature.asc" (application/pgp-signature)]

[prev in list] [next in list] [prev in thread] [next in thread]