Re: [h2] Re: Strange long INSER : One INSERT out of 18889 is very long

Hi, I'm using version 1.4.193. What do you mean by create a test case? I'm not sure I understand. By the way I ran several test on linux rather than windows but on the same computer and it went better. Thx, Nicolas Le dimanche 12 février 2017 14:00:51 UTC+1, Thomas Mueller Graf a écrit : > >

Re: [h2] Re: Strange long INSER : One INSERT out of 18889 is very long

Hi, collectReferencedChunks used to be a problem in the past, but recent version of H2 should be better. Which version do you use? If the latest version, could you create a simple test case? Regards, Thomas On Fri, Jan 27, 2017 at 12:48 PM, Noel Grandin wrote: > yeah,

[h2] Re: Strange long INSER : One INSERT out of 18889 is very long

Thanks, indeed it was easy, here the stack which is taking a few seconds : MainH2DataArchiAlone [Java Application] h2.MainH2DataArchiAlone at localhost:51761 Thread [pool-1-thread-1] (Suspended) owns: Object (id=50) owns: MVStore (id=51) owns: TransactionStore (id=52) owns: Database

Re: [h2] Re: Strange long INSER : One INSERT out of 18889 is very long

What I would do is set the breakpoint to stop all threads in the VM. Then I would see what my thread is blocked on - probably the main database lock object. Then I would see what thread is holding that lock, and print out that stack trace (or take a screenshot). In Eclipse this is fairly

[h2] Re: Strange long INSER : One INSERT out of 18889 is very long

First thank you for the help. I couldn't print stack trace of what's happening in H2. When I stop at the right point with a debugger and try to print the stack trace it only show me what's happening in my thread not in H2. Could you tell me more precisely what I can do to in order to see

Re: [h2] Re: Strange long INSER : One INSERT out of 18889 is very long

Hmm, that is being less useful than I thought, because that's all on the background thread, but what we want to know is what the current thread is up to. I regularly run large db's (several G) but I never do inserts of that size, so I'm not sure what is causing that. Could you maybe run it under

[h2] Re: Strange long INSER : One INSERT out of 18889 is very long

Here is what the profiler gives me : Profiler: top 3 stack trace(s) of of 59010 ms of 29386 thread dumps: 29222/29386 (99%): at org.h2.store.fs.FileNio.read(FilePathNio.java:74) at org.h2.mvstore.DataUtils.readFully(DataUtils.java:421) at org.h2.mvstore.FileStore.readFully(FileStore.java:98) at