forums › forums › Monyog › Monyog: Bugs / Feature Requests › Segfault
- This topic is empty.
-
AuthorPosts
-
-
April 30, 2010 at 11:05 am #11956rob2000Member
Hi Support,
I've installed MONyog-3.7.6-0.glibc-2.5-42.x86_64.rpm for a trial
on
Linux version 2.6.18-164.11.1.el5 ([email protected]) (gcc version 4.1.2 20080704 (Red Hat 4.1.2-46)) #1 SMP Wed Jan 20 07:32:21 EST 2010
and I'm getting the following segfault on analyzing the slow query log:
kernel: MONyog-bin[15941]: segfault at 0000000000000000 rip 0000003b9847bf0b rsp 0000000045b6fda8 error 6
thanks for your help.
regards,
Robert
-
April 30, 2010 at 12:22 pm #30795Sayan ChalihaMember
Hi Rob,
Could you check in MONyog's installation directory for a core dump? It's weird that it's segfaulting when analyzing the slow log: that part of MONyog is very old, and has gone through thorough testing over the years.
You should also check MONyog's log file (you'll find in the installation directory itself).
-
April 30, 2010 at 12:56 pm #30796rob2000Member
Hi Sayan,
thanks for your fast reply.
There is nothing in the log files. There is a core dump in the bin folder (quite large).
regards,
Robert
-
April 30, 2010 at 2:36 pm #30797Sayan ChalihaMember
Hi Rob,
Initially I missed the signal number — 6 — ie, SIGABRT, the signal sent to the process calling the function 'abort()' to make an abnormal exit. There are a couple of instances in MONyog's source where this function is called and all of them are, one way or the other, related to the filesystem and the system's word-size.
Is there any particular characteristic of your system I'm missing — as in any special settings or the like? I have never seen MONyog signal an abort before.
BTW, besides the log analyzer are the other features of MONyog working?
-
April 30, 2010 at 3:17 pm #30798Sayan ChalihaMember
Also, it looks like you're using CentOS 5.4 x86_64 but just to be doubly sure, could you include the output of
Code:uname -a?
-
April 30, 2010 at 3:30 pm #30799rob2000Member
Hi Sayan,
uname returns:
Linux testdb 2.6.18-164.11.1.el5 #1 SMP Wed Jan 20 07:32:21 EST 2010 x86_64 x86_64 x86_64 GNU/Linux
The system is running in vmware.
regards,
Robert
-
May 3, 2010 at 9:19 am #30800sandhya.rMember
Hi,
We are trying to reproduce the issue at our end with similar configuration as yours, but with no luck.
We have received the core dump and we are analyzing the dump.
Regards,
Sandhya
-
May 3, 2010 at 1:39 pm #30801sandhya.rMember
Hi,
We have found the location of the crash and analyzing the possibilities.
We'll update you tomorrow.
Regards,
Sandhya
-
May 3, 2010 at 3:46 pm #30802RohitMember
Thanks for your patience!
From the dump, we can make out that MONyog crashed while parsing the Slow Query Log.
Can you try parsing your Slow Query Logs once again and confirm?
If a particular Slow Query Log is the culprit, please create a ticket at http://www.webyog.com/support and attach the Log File.
Once reproduced, we guarantee a fix within 72 hours.
-
May 4, 2010 at 12:32 pm #30803Sayan ChalihaMember
Hi Rob,
There is a scenario where this same crash can be reproduced if the system is low on memory (such as in a virtualized system). To eliminate that as a possibility, here's a link to a binary: http://www.webyog.com/downloads/betas/not_released/MONyog-3.7.6-1.glibc-2.5-42.x86_64.rpm
This version of MONyog will log an error message in case memory allocation fails (besides the usual abnormal exit procedure).
-
May 4, 2010 at 4:06 pm #30804rob2000Member
Hi Sayan,
thanks for the feedback. I've installed the new version and it still segfaults without a log message. The system has 2GB memory.
regards
Robert
-
May 4, 2010 at 4:30 pm #30805rob2000Member
I'll try flushing the logs, they are quite large.
-
May 5, 2010 at 3:20 am #30806RohitMember
Even if the log file is large, MONyog should never crash.
Let Sayan comment here.
-
May 5, 2010 at 9:11 am #30807Sayan ChalihaMember
Hi Rob,
Sorry for all the trouble, but could you try the binary once again? I made a couple of changes to it. Here's the link again:
http://www.webyog.co…5-42.x86_64.rpm
You can expect the same thing — a message like “Allocation failed…” should be logged to MONyog's log this time.
It is very likely that there is some text in your slow log which MONyog isn't being able to parse correctly. When attempting to parse that particular line it gets caught in an infinite loop which is only broken (by the crash) when it tries to allocate a string which is 2.5GB in size.
It would be very helpful if you could upload your slow log file. If you're worried about privacy, let me assure you that we guarantee full privacy on our ticket system — what goes on there stays between the user and us. Moreover, you may definitely remove any information that you deem sensitive from the log.
-
-
AuthorPosts
- You must be logged in to reply to this topic.