Closed shlapkov closed 9 years ago
@lvca Can you take a look at this? Is this related to your note here: http://www.prjhub.com/#/issues/4633
This seems the classic MVCC exception, why is it an issue?
Because it wasn't the case before.
Could you show us the code responsible of this exception? Did you implement the retry in case of concurrent modification exception?
It wasn't code any code changes in our applications for quite a while. I notice this issue after resent changes in 2.0.14-SNAPSHOT and database 1.7.x. Issue is consistent. You do have our test database.
@lvca Can you recreate?
@shlapkov I have your database, but this is about your code. Can I reproduce this with OrientDBCluster.java? Or last file you sent to me?
No i can't reproduce it with OrientDBCluster.java. The same code is working fine in stand alone mode but constantly generate exception in distributed mode. There is no concurrent modification in our application in that time one client updating one record.
I'd need a way to reproduce this. Please could you send me a test case? Look also at: http://orientdb.com/docs/last/Concurrency.html to see if you managed concurrency correctly (The optimistic approach of OrientDB is the same since the beginning)
I am trying to reproduce issue with test applications
Hello @lvca please run the following test app:
https://github.com/shlapkov/OrientIssues/blob/master/ConcurrentModify/OrientDBModify.java
config files in: https://github.com/shlapkov/OrientIssues/tree/master/ConcurrentModify
first run with command line: "create 0" run in stand alone mode: "execute 0 false" no issue run in distributed alone mode: "execute 0 true" notice exception.
Running your app I was able to reproduce your problem. Fixing it...
Fixed.
@shlapkov We really do appreciate all your work helping us to recreate this. It should be resolved now. Can you confirm that this is resolved in the 2.0.14 snapshot?
yes in 2.0.14 snapshot working fine.
@shlapkov Great! Thank you.