Fb 4.0.1 - Operating system call pthread_mutex_lock failed. Error code 22

[Lucas Schatz]

Hello, I'm running a CentOS Linux release 7.9.2009 (Core), 16 cores, 48Gb RAM, 200-300 concurrent users, and recently migrated from Fb 3.0 to Fb 4.0.1, since migration almost everyday once a day I'm getting the error from firebird.log:

srvbd-red       Tue May 17 08:57:51 2022
        Operating system call pthread_mutex_lock failed. Error code 22
srvbd-red       Tue May 17 08:57:51 2022
        Operating system call pthread_mutex_lock failed. Error code 22
srvbd-red       Tue May 17 08:57:53 2022
        /opt/firebird/bin/fbguard: /opt/firebird/bin/firebird terminated abnormally (-1)
srvbd-red       Tue May 17 08:57:53 2022
        /opt/firebird/bin/fbguard: guardian starting /opt/firebird/bin/firebird

Also, at /var/log/messages I got (also from the previous days):

cat /var/log/messages* | grep segfault | sort
Apr 29 14:30:50 srvbd-red kernel: firebird[67198]: segfault at 2e8 ip 00007f1b89ffe650 sp 00007f16af1ed218 error 4 in libEngine13.so[7f1b89f70000+94e000]
May 10 10:47:12 srvbd-red kernel: firebird[237088]: segfault at 2e8 ip 00007feca8490650 sp 00007fe7cc43f218 error 4 in libEngine13.so[7feca8402000+94e000]
May 10 11:49:15 srvbd-red kernel: firebird[81150]: segfault at 2e8 ip 00007f05044b3650 sp 00007f0066fd1218 error 4 in libEngine13.so[7f0504425000+94e000]
May 11 14:19:21 srvbd-red kernel: firebird[32906]: segfault at 2e8 ip 00007f22b14d7650 sp 00007f1d9bdf9218 error 4 in libEngine13.so[7f22b1449000+94e000]
May 16 09:31:04 srvbd-red kernel: firebird[116210]: segfault at 2e8 ip 00007f8ba38aa650 sp 00007f85d3f75218 error 4 in libEngine13.so[7f8ba381c000+94e000]
May 17 08:57:51 srvbd-red kernel: firebird[171301]: segfault at 2e8 ip 00007f60db424650 sp 00007f5b9d7f0218 error 4 in libEngine13.so[7f60db396000+94e000]
May  2 14:27:22 srvbd-red kernel: firebird[33851]: segfault at 2e8 ip 00007f4d25b14650 sp 00007f4b169c8218 error 4 in libEngine13.so[7f4d25a86000+94e000]
May  2 17:19:39 srvbd-red kernel: firebird[171887]: segfault at 2e8 ip 00007fe0dbcd2650 sp 00007fdb0deb6218 error 4 in libEngine13.so[7fe0dbc44000+94e000]
May  5 20:29:35 srvbd-red kernel: firebird[109314]: segfault at 50 ip 00007f0e47b4abb3 sp 00007f0a7904eb70 error 4 in libEngine13.so[7f0e478bc000+94e000]
May  6 20:46:24 srvbd-red kernel: firebird[165287]: segfault at 50 ip 00007f7e14e65bb3 sp 00007f7955decb70 error 4
May  6 20:46:24 srvbd-red kernel: firebird[166562]: segfault at 50 ip 00007f7e14e65bb3 sp 00007f79592ebb70 error 4
May  6 20:46:24 srvbd-red kernel: firebird[167512]: segfault at 50 ip 00007f7e14e65bb3 sp 00007f79573a4b70 error 4
May  9 16:44:49 srvbd-red kernel: firebird[44280]: segfault at 2e8 ip 00007fed814e9650 sp 00007fe7c31ef218 error 4 in libEngine13.so[7fed8145b000+94e000]

Any idea of what could it be?

Note: I have the same system but fewer users running with Fb 4.0.0 without problems

Thanks

[Dimitry Sibiryakov]

Lucas Schatz wrote 17.05.2022 17:04:
> Any idea of what could it be?

Almost for sure it is a bug. You should set up your system to create core
dumps and install Firebird debug information package to make them useful.
Some how to: http://www.ibphoenix.com/resources/documents/development/doc_36

--
WBR, SD.

[Lucas Schatz]

ok, as it's a production server, can you tell me if there is a big performance loss?

Thanks

[Dimitry Sibiryakov]

Lucas Schatz wrote 17.05.2022 17:59:
> ok, as it's a production server, can you tell me if there is a big performance
> loss?

No performance losses. Debug info is used only by debugger during core dump
analyze.

--
WBR, SD.

[Lucas Schatz]

got the stack trace, where/how can I send it?

Thanks

[Dimitry Sibiryakov]

Lucas Schatz wrote 20.05.2022 23:37:
> got the stack trace, where/how can I send it?

Firebird-devel mail list is preferable for that:
https://firebirdsql.org/en/mailing-lists/

--
WBR, SD.

[Lucas Schatz]

Looking at the stack-trace, looks something related to a web server with PHP7.3 FPM and Fb V3.0.4.33054.

Just updated Firebird to 3.0.9, let's see if fixes the problem.