Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 08 Jul 2005 01:13:22 +0100
From:      Alex Zbyslaw <xfb52@dial.pipex.com>
To:        Tuc at T-B-O-H <ml@t-b-o-h.net>
Cc:        freebsd-questions@freebsd.org
Subject:   Re: Problems since 5.3-RELEASE-p15
Message-ID:  <42CDC522.5010309@dial.pipex.com>
In-Reply-To: <200507072314.j67NENDG020566@himinbjorg.tucs-beachin-obx-house.com>
References:  <200507072314.j67NENDG020566@himinbjorg.tucs-beachin-obx-house.com>

next in thread | previous in thread | raw e-mail | index | archive | help
Tuc at T-B-O-H wrote:

>The only pattern I see is that it starts all of a sudden, goes for
>a little while, EVERYTHING you try to do fails, I get the 
>/libexec/ld-elf.so.1: top: Shared object has no run-time symbol table
>type errors, and then all of a sudden it goes about its merry way.
>
>	I'm not sure what a "run-time symbol table" is, and if it gets
>rebuilt at any time, or is something thats part of the program, or what....
>But its not like one time it behaves one way, then another time it behaves
>another. 
>
My description would be hazy at best, so hopefully someone more au fait 
will chime in.  But I can definitively state that, no it isn't something 
that gets rebuilt every so often.

However, if you run ktrace on a simple program like ls: ktrace ls
then do a kdump | less, you will see that after finding ls, 
/libexec/ld-elf.so.1 is the first thing accessed.  So, when things start 
working again, /libexec/ld-elf.so.1 is magically fixed, which just makes 
no sense.

I assume there are no other messages obvious errors in /var/log/messages?

One final thought is that it could be the disk.  You could try 
installing smartmontools and see if the disk thinks it is OK -- though 
of course it could be the controller.  But in such a case I might expect 
other errors.

>I'd think if it was a memory issue, it'd be something other than 
>the same "Shared object..." thing each time. One time it'd be a bus error,
>another time a "this isn't executable", etc....
>  
>
The classic bad memory symptoms are periodic reboots, and random 
segmentation faults.  The latter is particular to bad memory, I would 
say.  The fomer can have other causes.

It's the intermittent nature of the fault that really makes me think 
hardware.  Am I right in remembering that you upgraded to 5.4 and still 
had the same problems?

--Alex






Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?42CDC522.5010309>