[rsyslog] abort in 4.2.1

david at lang.hm david at lang.hm
Wed Aug 26 13:04:00 CEST 2009


I have a 10 hour meeting today, but I should be able to get the info you 
need from the core file.

the problem is very rare compared to the number of logs recieved, when I 
was testing on monday it took quite a while for 5.1.4+ to die, but when I 
duplicated things yesterday it was sometimes dieing so fast that I 
couldn't do a start and then a ps and see it running (and since it dumped 
the core file in / instead of where I ran it from  I did this multiple 
times)

I'll test whatever commit you want me to. I would suggest that instead of 
directing me to a specific branch, just direct me to a commit so that 
there is no abiguity over what is being tested.

David Lang

  On Wed, 26 Aug 2009, Rainer Gerhards wrote:

> Date: Wed, 26 Aug 2009 12:20:09 +0200
> From: Rainer Gerhards <rgerhards at hq.adiscon.com>
> Reply-To: rsyslog-users <rsyslog at lists.adiscon.com>
> To: rsyslog-users <rsyslog at lists.adiscon.com>
> Subject: Re: [rsyslog] abort in 4.2.1
> 
> David, an update:
>
> it still would be good if you could obtain the info I asked for below, but it
> would also useful to know (in addition) if the current v4-stable does
> experience the problem, too. v4-stable (and most probably the version you
> had) has code that is different in some key sections. So a test if it fails,
> too, actually tells more than I initially thought.
>
> Please also note that I have seen a potential bug inside the new sanitation
> code, but I think it is very, very unlikely to be causing the problem. I'll
> address this starting in v4-devel so a re-test of that version would also be
> useful once it is there. I will post when I am done with the fixing.
>
> Rainer
>
>> -----Original Message-----
>> From: rsyslog-bounces at lists.adiscon.com [mailto:rsyslog-
>> bounces at lists.adiscon.com] On Behalf Of Rainer Gerhards
>> Sent: Wednesday, August 26, 2009 11:50 AM
>> To: rsyslog-users
>> Subject: Re: [rsyslog] abort in 4.2.1
>>
>> David,
>>
>> one more thing. If you still have the core file, could you start up gdb
>> again
>> and then do
>>
>> (gdb) thread 1
>> (gdb) print sanitizeMessage::pszMsg
>> (gdb) print sanitizeMessage::szSanBuf
>> (gdb) print sanitizeMessage::pMsg
>> (gdb) print *sanitizeMessage::pMsg  # note the asterisk!
>> (gdb) print sanitizeMessage::iMaxLine
>> (gdb) print sanitizeMessage::maxDest
>>
>> The following ones likely will yield to no result as they are usually
>> optimized out (moved into registers):
>> (gdb) print sanitizeMessage::iSrc
>> (gdb) print sanitizeMessage::iDst
>> (gdb) print sanitizeMessage::pDst
>>
>> That will tell me if the pointers are ok, and what they actually point
>> to.
>> Based on the addresses I see, I guess that the message object pointer
>> provided is already invalid. But it is hard to verify without the
>> context...
>>
>> Rainer
>> _______________________________________________
>> rsyslog mailing list
>> http://lists.adiscon.net/mailman/listinfo/rsyslog
>> http://www.rsyslog.com
> _______________________________________________
> rsyslog mailing list
> http://lists.adiscon.net/mailman/listinfo/rsyslog
> http://www.rsyslog.com
>



More information about the rsyslog mailing list