[h2] Cannot open database.

2014-09-11 Thread Dani Fernandez
This morning I tried to open my test database to make some tests there and I couldn't open it anymore; even tried the recover tool. When I try to open it using the h2 console I am getting this exception: Unique index or primary key violation: PRIMARY KEY ON .PAGE_INDEX; SQL statement: ALTER

[h2] Is 60 write operations/sec really the limit for H2?

2014-09-11 Thread Anders
Hi, *According to documentation, the system doesn't call FileDescriptor.sync () nor FileChannel.force () nor fsync http://en.wikipedia.org/wiki/Sync_%28Unix%29 for every commit because they significantly degrade system performance: only 60 write operations per second

Re: [h2] Is 60 write operations/sec really the limit for H2?

2014-09-11 Thread Ryan How
I think it means that is why it doesn't sync for every commit, so it can get higher performance than that. If it syncs for every commit, then if you are committing each row and doing a lot of updates for example, then you get bad performance. Probably wouldn't matter much for an SSD, but on a

[h2] H2DB select performance issue :(

2014-09-11 Thread Arun
Hi guys.. I am new in H2DB .i m running system where H2db is having around 3 crore [30 million] data in one table only. [H2DB contains 1 table only] select query normally takes milliseconds only.around 100-200 ms. bt at sudden h2db takes upto 90 second for a single fetch query. this happens

[h2] Commit or rollback is not allowed within a trigger but I'm not rolling back

2014-09-11 Thread Gili
Hi, I am getting Commit or rollback is not allowed within a trigger under H2 1.3.176 but I am not actually rolling anything back. As far as I understand the following sequence of events occurs: 1. Java code invokes: delete from PARTICIPANT where PARTICIPANT.LAST_SEEN_TIME ? 2. A after delete

[h2] Re: Commit or rollback is not allowed within a trigger but I'm not rolling back

2014-09-11 Thread Gili
Just to clarify, it is very important for my code to differentiate between a deadlock and other kinds of failures. My code assumes that in case of a deadlock a client should sleep and retry the operation. Otherwise, the code assumes this is a programming error and the client never retries.