Re: problem with locks on head, backend crash

Hello

when I tested ptop, I found some problems

this pgbench is very slow and when after getting table of locks from
ptop I got backend crash.

[pavel(at)prg31 ~]$ /usr/local/pgsql/bin/pgbench -c80 -t1000 postgres
starting vacuum...end.
WARNING: you don't own a lock of type ShareLock
WARNING: you don't own a lock of type ShareLock
WARNING: terminating connection because of crash of another server process
DETAIL: The postmaster has commanded this server process to roll back
the current transaction and exit, because another server process
exited abnormally and possibly corrupted shared memory.
HINT: In a moment you should be able to reconnect to the database and
repeat your command.
WARNING: terminating connection because of crash of another server process
DETAIL: The postmaster has commanded this server process to roll back
the current transaction and exit, because another server process
exited abnormally and possibly corrupted shared memory.

from log:

LOG: unexpected EOF on client connection
LOG: could not send data to client: Přerušena roura (SIGPIPE)
LOG: unexpected EOF on client connection
LOG: could not send data to client: Přerušena roura (SIGPIPE)
LOG: unexpected EOF on client connection
WARNING: problem in alloc set MessageContext: req size > alloc size
for chunk 0x8b6d1e0 in block 0x8b6bb50
WARNING: problem in alloc set MessageContext: req size > alloc size
for chunk 0x8b6d1e0 in block 0x8b6bb50
LOG: server process (PID 29693) was terminated by signal 11: Segmentation fault
LOG: terminating any other active server processes
WARNING: terminating connection because of crash of another server process
DETAIL: The postmaster has commanded this server process to roll back
the current transaction and exit, because another s
HINT: In a moment you should be able to reconnect to the database and
repeat your command.
WARNING: terminating connection because of crash of another server process

Regards
Pavel Stehule

Pavel Stehule wrote:
> when I tested ptop, I found some problems

Which version are you running?

--
Heikki Linnakangas
EnterpriseDB http://www.enterprisedb.com

On 07/04/2008, Heikki Linnakangas <heikki(at)enterprisedb(dot)com> wrote:
> Pavel Stehule wrote:
>
> > when I tested ptop, I found some problems
> >
>
> Which version are you running?
I am sorry, HEAD 8.4 today actualized

Pavel

>
> --
> Heikki Linnakangas
> EnterpriseDB http://www.enterprisedb.com
>

Pavel Stehule escribió:

> WARNING: problem in alloc set MessageContext: req size > alloc size
> for chunk 0x8b6d1e0 in block 0x8b6bb50
> WARNING: problem in alloc set MessageContext: req size > alloc size
> for chunk 0x8b6d1e0 in block 0x8b6bb50

I suggest you make distclean and rebuild the whole thing from scratch --
I have seen these kind of problems when a loadable module is not built
the same as the backend, for example.

--
Alvaro Herrera http://www.CommandPrompt.com/
The PostgreSQL Company - Command Prompt, Inc.

On 07/04/2008, Alvaro Herrera <alvherre(at)commandprompt(dot)com> wrote:
> Pavel Stehule escribió:
>
>
> > WARNING: problem in alloc set MessageContext: req size > alloc size
> > for chunk 0x8b6d1e0 in block 0x8b6bb50
> > WARNING: problem in alloc set MessageContext: req size > alloc size
> > for chunk 0x8b6d1e0 in block 0x8b6bb50
>
>
> I suggest you make distclean and rebuild the whole thing from scratch --
> I have seen these kind of problems when a loadable module is not built
> the same as the backend, for example.
>

I did it, but maybe there was some ??? I can't to repeat it after restart

Pavel

> --
> Alvaro Herrera http://www.CommandPrompt.com/
> The PostgreSQL Company - Command Prompt, Inc.
>