Re: BUG #3245: PANIC: failed to re-find shared loc k o b j ect
От | Tom Lane |
---|---|
Тема | Re: BUG #3245: PANIC: failed to re-find shared loc k o b j ect |
Дата | |
Msg-id | 19442.1177382558@sss.pgh.pa.us обсуждение исходный текст |
Ответ на | Re: BUG #3245: PANIC: failed to re-find shared loc k o b j ect (Heikki Linnakangas <heikki@enterprisedb.com>) |
Ответы |
Re: BUG #3245: PANIC: failed to re-find shared loc k o b j ect
|
Список | pgsql-bugs |
Well hot dang, I managed to reproduce this: $ ./followlog "dbname = tschema" <log.sql command failed at log line 30522: COMMIT PREPARED '257_dm0td2luMjAwMHByby8yNTM2AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA==_MQAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA==' PANIC: failed to re-find shared lock object server closed the connection unexpectedly This probably means the server terminated abnormally before or while processing the request. $ So at least now we know it's not a Windows-specific issue, which is something I was getting more and more worried about after two days of not being able to duplicate the crash. I went so far as to make a program that reads the log excerpt Michel sent, and execute the commands in sequence on as many different connections as he had active ... and darn if it doesn't fail. It seems fully reproducible on both 8.2 and HEAD. Now to start debugging. We can stop pestering Michel to run various strange binaries, at least. regards, tom lane
В списке pgsql-bugs по дате отправления: