Thread: Backend crashe with signal 11 wher restoring big database.

Your name        :    Krok
Your email address    :    krok@void.ru


System Configuration
---------------------
  Architecture        : Dual PIII-500

  Operating System    : FreeBSD 4.7-RELEASE-p7

  PostgreSQL version    : PostgreSQL-7.3.2

  Compiler used        : gcc 2.95.4


Please enter a FULL description of your problem:
------------------------------------------------

When restoring large database (there are several tables with >1'000'000 rows and somthing about 50 tables with
1'000-10'000rows) using cat <dump_file> | psql -U psql, backend crashes and there are messages in logs : 

Mar 25 15:14:08 frost postgres[338]: [14] LOG:  server process (pid 64665) was terminated by signal 11
Mar 25 15:14:08 frost postgres[338]: [15] LOG:  terminating any other active server processes
Mar 25 15:14:08 frost postgres[338]: [16] LOG:  all server processes terminated; reinitializing shared memory and
semaphores
Mar 25 15:14:08 frost postgres[64921]: [17] LOG:  database system was interrupted at 2003-03-25 15:13:01 MSK
Mar 25 15:14:08 frost postgres[64921]: [18] LOG:  checkpoint record is at 11/451E6D80
Mar 25 15:14:08 frost postgres[64921]: [19] LOG:  redo record is at 11/450304A0; undo record is at 0/0; shutdown FALSE
Mar 25 15:14:08 frost postgres[64921]: [20] LOG:  next transaction id: 94286227; next oid: 2959327
Mar 25 15:14:08 frost postgres[64921]: [21] LOG:  database system was not properly shut down; automatic recovery in
progress
Mar 25 15:14:08 frost postgres[64921]: [22] LOG:  redo starts at 11/450304A0
Mar 25 15:14:10 frost postgres[64921]: [23] LOG:  ReadRecord: record with zero length at 11/4585EB20
Mar 25 15:14:10 frost postgres[64921]: [24] LOG:  redo done at 11/4585EAFC
Mar 25 15:14:15 frost postgres[64921]: [25] LOG:  recycled transaction log file 0000001100000044
Mar 25 15:14:15 frost postgres[64921]: [26] LOG:  database system is ready


My configuration is :

------

max_connections = 256
superuser_reserved_connections = 2

shared_buffers = 4096           # 2*max_connections, min 16, typically 8KB each
max_fsm_relations = 100 # min 10, fsm is free space map, ~40 bytes
max_fsm_pages = 10000           # min 1000, fsm is free space map, ~6 bytes
max_locks_per_transaction = 64  # min 10
wal_buffers = 128               # min 4, typically 8KB each

sort_mem = 8192         # min 64, size in KB
vacuum_mem = 65536              # min 1024, size in KB

checkpoint_segments = 1 # in logfile segments, min 1, 16MB each
checkpoint_timeout = 3600       # range 30-3600, in seconds
commit_delay = 1000             # range 0-100000, in microseconds
commit_siblings = 5             # range 1-1000
wal_debug = 0

effective_cache_size = 16384

show_parser_stats = false
show_planner_stats = false
show_executor_stats = false
show_statement_stats = false

stats_start_collector = true
stats_reset_on_server_start = true
stats_command_string = true
stats_row_level = true
stats_block_level = true

------

Dual PIII-500, 782MB, SCSI


Thanks.

With best regards,
Krok.

On 25 Mar 2003 at 18:13, KroK wrote:

> Mar 25 15:14:08 frost postgres[338]: [14] LOG:  server process (pid 64665) was terminated by signal 11
> Mar 25 15:14:08 frost postgres[338]: [15] LOG:  terminating any other active server processes

My understanding is that signal 11 normally indicates a hardware
problem, such as faulty RAM.  Have you tested the RAM?

NOTE: I'm guessing.
--
Dan Langille : http://www.langille.org/

Yes, RAM is ok.
My server works with very high load and there are no problems at all...

BTW, backend crashes both when used COPY and INSERT in the dump.

I've split dump file into several parts and restore them one by one. Everything was ok.

Krok.

On Tue, Mar 25, 2003 at 12:35:34PM -0500, Dan Langille wrote:
> On 25 Mar 2003 at 18:13, KroK wrote:
>
> > Mar 25 15:14:08 frost postgres[338]: [14] LOG:  server process (pid 64665) was terminated by signal 11
> > Mar 25 15:14:08 frost postgres[338]: [15] LOG:  terminating any other active server processes
>
> My understanding is that signal 11 normally indicates a hardware
> problem, such as faulty RAM.  Have you tested the RAM?
>
> NOTE: I'm guessing.
> --
> Dan Langille : http://www.langille.org/
>
>
> ---------------------------(end of broadcast)---------------------------
> TIP 6: Have you searched our list archives?
>
> http://archives.postgresql.org