Report: firebird: segfault at 7fb0c24c3000 ip 00007fb3cef3fd5a sp 00007fb043222190 error 4 in libEngine13.so[7fb3cec9a000+6c1000]

[Lucas Schatz]

Firebird version: Firebird-5.0.0.130, superserver

OS: Centos 7

So, just now I got a report from one server running Firebird-5.0.0.1306 suddenly got it's Fb service restarted,  in the firebird.log there is:

server        Tue Jul  2 20:00:19 2024
        /opt/firebird/bin/fbguard: /opt/firebird/bin/firebird terminated abnormally (-1)

server        Tue Jul  2 20:00:19 2024
        /opt/firebird/bin/fbguard: guardian starting /opt/firebird/bin/firebird

Also, half an hour before, `dmesg` reported this line, don't know if related:

[Jul 2 18:37] firebird[188526]: segfault at 7fb0c24c3000 ip 00007fb3cef3fd5a sp 00007fb043222190 error 4 in libEngine13.so[7fb3cec9a000+6c1000]

Attached is the stack trace, hope it's helpful

[AlexPeshkoff]

Very interesting stack - first time I see such. Can you place to some downloadable place your core dump? If yes - please compress it using xz utility (it efficiently prevents abuse of disk space by 'holes' in core dump file). Sorry if I repeat well known things but one advice - use switch '-T NN', where NN is number of CPU kernels you want xz to run in parallel, w/o that switch it's rather slow.

среда, 3 июля 2024 г. в 03:43:55 UTC+3, lucas...@gmail.com:

[Dimitry Sibiryakov]

AlexPeshkoff wrote 03.07.2024 10:37:
> Very interesting stack - first time I see such.

It looks like thread 1 access TipCache while thread 31 already destructed it.

--
WBR, SD.

[Alex Peshkoff]

Yes, but I want more details.

[Lucas Schatz]

Hi, to whom can I send the link?

Would like to send PM

--
You received this message because you are subscribed to the Google Groups "firebird-support" group.
To unsubscribe from this group and stop receiving emails from it, send an email to firebird-suppo...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/firebird-support/00f2596d-ff8c-4c21-9ccf-c425b735f1b8%40gmail.com.

[AlexPeshkoff]

Almost sure that this bug is already fixed by https://github.com/FirebirdSQL/firebird/issues/8150 .

@lucas - you may wait for 5.0.1 release or try snapshot build. Bug is definitely rarely reproducible.

среда, 3 июля 2024 г. в 17:03:12 UTC+3, lucas...@gmail.com:

[Lucas Schatz]

OK, I'll wait for 5.0.1

Thank you

[Lucas Schatz]

Hello, last day 12, I got another segfault in another server:

segfault at 50 ip 00007fd09bddadb6 sp 00007fcd0a459190 error 4 in libEngine13.so[7fd09bb35000+6c1000]

can you take a look at the stack trace to see if it's related? (I'm using Fb 5.0.0)

[AlexPeshkoff]

Crashed thread appears identical. What makes me a little surprised is that there is one more almost same thread (executing ON DETACH triggers), which should not be related to supposed bug. May be your Linux saves something like list of open handles for dead process  (I know some do it). If yes it will be interesting to take a look at it.

вторник, 16 июля 2024 г. в 15:45:55 UTC+3, lucas...@gmail.com:

[Lucas Schatz]

So, what you proposes?

We are using Centos 7