Hi Marc,
first: I´m new to sequoia.
Welcome to the community !
I just installed version 2.10.3 and 3.0beta2 (with PostgreSQL 8.1), but with
both versions I ran into a deadlock after a few statements (heavy-load, about
20 concurrent threads with transactions, but only simple selects, inserts and
updates). Without sequoia (normal jdbc) the application runs for weeks on the
same system and database without any problems.
Are there any known issues?
From the message you get below, it seems that Sequoia detected a
deadlock based on its table-based deadlock detection. It might happen
that the timeout for the deadlock detection thread is too short (default
is 30 seconds). However, the fact that writes are running slower through
Sequoia than directly on PostgreSQL might produce deadlocks that
couldn't be seen when running with a single database.
It also depends on the kind of query your are executing. If you are
calling stored procedures, they require an exclusive lock on the
database if no semantic information is provided. Therefore this can
easily introduce artificial deadlocks.
If this is a simple scheduling problem, there are options to force
strict table locking or tweak the locking using the proper semantic info.
Could you provide us with more details about your workload and optimally
with a use minimal use case to reproduce the deadlock problem?
Thanks for your feedback,
Emmanuel
--
Emmanuel Cecchet
Chief Scientific Officer, Continuent
Blog: http://emanux.blogspot.com/
Open source: http://www.continuent.org
Corporate: http://www.continuent.com
Skype: emmanuel_cecchet
Cell: +33 687 342 685
_______________________________________________
Sequoia mailing list
[email protected]
https://forge.continuent.org/mailman/listinfo/sequoia
