Subject Firebird 2.1 crashing
Author mivi71dk

We have a DB running firebird 2.1.
Its version LI-V6.3.1.17910 Firebird 2.1.

Its running on a Debian Linux with 12 Gb RAM.
Its on 23 Gb in size.

Usually there are some 100-150 users connected.

All clients connect with FB 2.1 client.

Well - periodically FB stops.
When this happens I get this error on the client side:

Unable to complete network request to host "kaufmanndbserver".
Error writing data to the connection.
An established connection was aborted by the software in your host machine.

Last time we had this was just 3 days ago.
We retrieved this from linux:

14:36:07 localhost kernel: fbserver[18628]: segfault at 00002aaacbf24af2 rip 00000000005150f3 rsp 0000000041800d70 error 4

Core was generated by `/opt/firebird/bin/fbserver'.
Program terminated with signal 11, Segmentation fault.
#0 0x00000000005150f3 in Firebird::IntlUtil::cvtUtf16ToAscii ()

(gdb) bt
#0 0x00000000005150f3 in Firebird::IntlUtil::cvtUtf16ToAscii ()
#1 0x00000000005157a9 in Firebird::IntlUtil::cvtUtf16ToAscii ()
#2 0x000000000043877a in (anonymous namespace)::threadStart ()
#3 0x00000000005acbcf in SVC_output ()
#4 0x00000000005a9d30 in SVC_output ()
#5 0x00000000004379f3 in (anonymous namespace)::threadStart ()
#6 0x000000000041424d in putchar ()
#7 0x00000000004149a6 in putchar ()
#8 0x0000000000414452 in putchar ()
#9 0x0000000000416e25 in putchar ()
#10 0x000000000042d634 in (anonymous namespace)::threadStart ()
#11 0x00002ac1e051af1a in start_thread () from /lib/
#12 0x00002ac1e06f35d2 in clone () from /lib/
#13 0x0000000000000000 in ?? ()


I don't know if this can be of any use.

This linux server is running in our house, the clients connecting via a Windows Terminal Server. Its the only DB running on this server.

Besides we have 4 other windows servers running FB2.1 with at least 8 DB on each server. Users connect the same way using terminal server clients. Those 4 servers never does this.

Information regarding our linux server is:

Linux DBSRV06 2.6.18-6-amd64 #1 SMP Tue Aug 19 04:30:56 UTC 2008 x86_64 GNU/Linux

Tasks: 107 total, 2 running, 105 sleeping, 0 stopped, 0 zombie
Cpu(s): 0.7%us, 0.1%sy, 0.0%ni, 98.8%id, 0.3%wa, 0.1%hi, 0.1%si, 0.0%st
Mem: 12301116k total, 12213104k used, 88012k free, 113504k buffers
Swap: 6835616k total, 56k used, 6835560k free, 11595032k cached


It seems as if it happens, when there is rather heavy use on the system. But this I am not sure of.

In average it happens once a month.
Some months ago we disabled the sweep function on run it once every nigth. Besides users now never experience a performance drop due to sweep it seems as if it happens less now. Eventhough it still happens.

Usually the gab between transactions is between 2.000 and 100.000.
And always when the last users disconnects back to 1.

We never looses data. Just stopping the DB and starting it again and all is well.
But a customer is a bit annoyed by this.

Anyone with any suggestions to what might cause this?
And any suggestions to any solution.