You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I create/alter/drop tables in parallel and in high frequency (stress and concurrency test)
During this test I run gbak on the database, and I get some "table id <number> is not defined" errors from gbak.
I tought gbak runs in snapshot mode, so this should never happen. (or snapshot is just for data and not for metadata?)
Also sometimes the reported error is wrong, because the reported table id is existed before the gbak start, and still existed after the gbak finish and not dropped/altered druing the run and still get reported as not defined.
The text was updated successfully, but these errors were encountered:
Yes, gbak run backup in snapshot transaction, but engine see metadata despite of user transaction mode.
BTW, it is not recommended (since IB times) to run DDL in concurrent environment.
Do you have reproducible test case ?
No promise to fully fix it in current versions, but some improvements could be made, I hope.
Submitted by: Attila Molnár (e_pluribus_unum)
Hi!
I create/alter/drop tables in parallel and in high frequency (stress and concurrency test)
During this test I run gbak on the database, and I get some "table id <number> is not defined" errors from gbak.
I tought gbak runs in snapshot mode, so this should never happen. (or snapshot is just for data and not for metadata?)
Also sometimes the reported error is wrong, because the reported table id is existed before the gbak start, and still existed after the gbak finish and not dropped/altered druing the run and still get reported as not defined.
The text was updated successfully, but these errors were encountered: