Thread: LWLockAcquire

LWLockAcquire

From
Edwin Grubbs
Date:
Hello,

We have been running postgresql-7.2.2 for 4 months without problems, but now we
are getting the errors listed below. I am hoping to find out more information
on why this error is being caused now, and wasn't caused earlier. If the only
real fix is an upgrade to pg 7.3.x, then I would like to know how to reduce
these errors until we can test a migration to that version. Are the
LWLockAcquire errors caused by a lack of shared memory or by a lack of system
memory. Could we tweak postgres' buffers to fix this problem?

Thanks,
Edwin Grubbs

System info:
----------------
OS: RH Linux 7.2
libc: glibc-2.2.4-19.3
# uname -a
Linux lhs.rackspace.com 2.4.18-10smp #1 SMP Wed Aug 7 11:17:48 EDT 2002 i686
unknown

Errors:
---------------------------------------------------------------------------
Jan 20 16:50:17 lhs postgres[8517]: [29] FATAL 1:  This connection has been
terminated by the administrator.
Jan 20 16:50:17 lhs postgres[8517]: [30] FATAL 1:  LWLockAcquire: can't wait
without a PROC structure
Jan 20 16:50:17 lhs postgres[22828]: [28] DEBUG:  server process (pid 8517)
exited with exit code 1
Jan 20 16:50:17 lhs postgres[22828]: [29] DEBUG:  terminating any other active
server processes
Jan 20 16:50:17 lhs postgres[8832]: [29-1] NOTICE:  Message from PostgreSQL
backend:
Jan 20 16:50:17 lhs postgres[8832]: [29-2] ^IThe Postmaster has informed me
that some other backend
Jan 20 16:50:17 lhs postgres[8832]: [29-3] ^Idied abnormally and possibly
corrupted shared memory.
Jan 20 16:50:17 lhs postgres[8832]: [29-4] ^II have rolled back the current
transaction and am
Jan 20 16:50:17 lhs postgres[8832]: [29-5] ^Igoing to terminate your database
system connection and exit.
Jan 20 16:50:17 lhs postgres[8841]: [30] FATAL 1:  The database system is in
recovery mode
Jan 20 16:50:17 lhs postgres[22828]: [30] DEBUG:  all server processes
terminated; reinitializing shared memory and semaphores
Jan 20 16:50:18 lhs postgres[8842]: [31] DEBUG:  database system was
interrupted at 2003-01-20 16:44:54 CST
Jan 20 16:50:18 lhs postgres[8842]: [32] DEBUG:  checkpoint record is at 10/
E933F164
Jan 20 16:50:18 lhs postgres[8842]: [33] DEBUG:  redo record is at 10/E92ED5A0;
undo record is at 0/0; shutdown FALSE
Jan 20 16:50:18 lhs postgres[8842]: [34] DEBUG:  next transaction id:
168879810; next oid: 64505583
Jan 20 16:50:18 lhs postgres[8842]: [35] DEBUG:  database system was not
properly shut down; automatic recovery in progress
Jan 20 16:50:18 lhs postgres[8842]: [36] DEBUG:  redo starts at 10/E92ED5A0
Jan 20 16:50:18 lhs postgres[8842]: [37] DEBUG:  ReadRecord: record with zero
length at 10/E972B01C
Jan 20 16:50:18 lhs postgres[8842]: [38] DEBUG:  redo done at 10/E972AFF8
Jan 20 16:50:20 lhs postgres[8843]: [31] FATAL 1:  The database system is
starting up
Jan 20 16:50:20 lhs postgres[8844]: [31] FATAL 1:  The database system is
starting up
Jan 20 16:50:22 lhs postgres[8842]: [39] DEBUG:  recycled transaction log file
00000010000000E8
Jan 20 16:50:22 lhs postgres[8842]: [40] DEBUG:  database system is ready


Re: LWLockAcquire

From
Tom Lane
Date:
Edwin Grubbs <egrubbs@rackspace.com> writes:
> We have been running postgresql-7.2.2 for 4 months without problems, but now we
> are getting the errors listed below.

Update to 7.2.3.

            regards, tom lane