Hi Tom,
thanks for answering me. In fact this is a bit strange thing. I don't know
well
postgresql locking mechanism but it seem this to cause the problem.
I was migrating a very large db with pgadmin porting tool from mssql server.
This failure
occours when, during the migration, I click on some areas of pgadmin, trying
to
refresh the content of the db or viewing the data in a table. In fact,
during the migration,
I was also doing some development and I've tried for two times to work in
parallel with
the migration and the development. I observed that this LockAcquire is a
routine about deadlock
avoiding. I've also done a dump of the db a month ago, all went perfect.
Anyway my db is, I think, very large. I have 700 tables, and someone of
those are larger than 10000 records.
Now I try to migrate the db without do nothing else and expanding this
max_locks_per_transaction.
regards, Samuele Brignoli.
-----Messaggio originale-----
Da: Tom Lane [mailto:tgl@sss.pgh.pa.us]
Inviato: lunedì 15 aprile 2002 16.41
A: samuele.brignoli@darpha.com
Cc: PostgreSQL General List
Oggetto: Re: [GENERAL] I: Migrating a very large db
"Samuele Brignoli" <samuele.brignoli@darpha.com> writes:
> ERROR: LockAcquire: holder table out of memory
You could probably work around that by increasing the postmaster
configuration parameter max_locks_per_transaction (note you will need to
restart postmaster to make it take effect). However, I'm curious to
know what you were doing that made it fail; I've not seen very many
reports of people running into this limit. How many tables are in this
database? Does PGAdmin try to restore them all in one transaction?
Also, do you have problems pg_dumping the database?
regards, tom lane