Re: Tracking down log segment corruption

Поиск
Список
Период
Сортировка
От Gordon Shannon
Тема Re: Tracking down log segment corruption
Дата
Msg-id h2kb2dd93301005021247k3cb308d4v5de10ad77c5b051c@mail.gmail.com
обсуждение исходный текст
Ответ на Re: Tracking down log segment corruption  (Tom Lane <tgl@sss.pgh.pa.us>)
Ответы Re: Tracking down log segment corruption  (Tom Lane <tgl@sss.pgh.pa.us>)
Список pgsql-general
Дерево обсуждения


On Sun, May 2, 2010 at 12:52 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
Gordon Shannon <gordo169@gmail.com> writes:
> Bingo.  Yes it is reasonable.  It was 25 seconds between my altering the
> index in question and the server crash.

Sounds like we have a smoking gun.  Could you show all your non-default
postgresql.conf settings on the master?  I'm wondering about
full_page_writes in particular, but might as well gather all the
relevant data while we're at it.  Easiest way is:

select name,setting from pg_settings where source not in ('default','override');

                       regards, tom lane

 archive_command                 | cp %p /var/lib/pgsql/wal/%f.wrk; mv /var/lib/pgsql/wal/%f.wrk /var/lib/pgsql/wal/%f
 archive_mode                    | on
 authentication_timeout          | 20
 autovacuum_analyze_scale_factor | 0.05
 autovacuum_freeze_max_age       | 1500000000
 autovacuum_vacuum_cost_delay    | -1
 autovacuum_vacuum_scale_factor  | 0.1
 checkpoint_segments             | 128
 checkpoint_warning              | 300
 commit_siblings                 | 1
 cpu_index_tuple_cost            | 0.001
 cpu_operator_cost               | 0.0005
 cpu_tuple_cost                  | 0.003
 DateStyle                       | ISO, MDY
 deadlock_timeout                | 5000
 default_text_search_config      | pg_catalog.english
 effective_cache_size            | 6291456
 lc_messages                     | en_US.UTF-8
 lc_monetary                     | en_US.UTF-8
 lc_numeric                      | en_US.UTF-8
 lc_time                         | en_US.UTF-8
 listen_addresses                | *
 log_autovacuum_min_duration     | 0
 log_destination                 | stderr
 log_directory                   | /var/log/postgres
 log_filename                    | pg%d.log
 log_line_prefix                 | %p %u %r %t [%l]
 log_min_duration_statement      | 180000
 log_rotation_age                | 1440
 log_rotation_size               | 0
 log_temp_files                  | 0
 log_timezone                    | US/Eastern
 log_truncate_on_rotation        | on
 logging_collector               | on
 maintenance_work_mem            | 8388608
 max_connections                 | 500
 max_stack_depth                 | 2048
 port                            | 5432
 search_path                     | public, archive, _slony_cluster
 shared_buffers                  | 2359296
 standard_conforming_strings     | on
 synchronous_commit              | off
 temp_tablespaces                | ts27,ts28,ts29,ts30,ts31,ts32,ts33,ts34,ts35,ts36,ts37
 TimeZone                        | US/Eastern
 timezone_abbreviations          | Default
 track_functions                 | pl
 vacuum_freeze_min_age           | 500000000
 vacuum_freeze_table_age         | 1300000000
 wal_buffers                     | 1024
 work_mem                        | 204800

В списке pgsql-general по дате отправления:

Предыдущее
От: Dave Page
Дата:
Сообщение: Re: postgres crashes - could not reattach to shared memory
Следующее
От: Tom Lane
Дата:
Сообщение: Re: Tracking down log segment corruption