"R" == Rolandas rolnas@gmail.com writes:
Hi Rolandas
R> Hi, On 2019-05-02 17:47, Goswin von Brederlow wrote: >> Hi, >> >> I can't reproduce this error with the same settings so it doesn't >> seem to be related to the choices you made in the installer. I'm >> thinking that something odd must happen before that causing the >> error.
R> Race conditions are difficult to reproduce.
previously you said something about failure in 50% of the time. So are you now saying that you can't reproduce it anymore after a fresh install and running qlustar-initial-config?
If you still can, please send the logs as described. If not, we consider this issue as closed.
Best,
Roland
R> In our case it could be caused by stealing CPU time from parallel R> threads in sqlalchemy code by other users cloud VM's. Quick R> search on internet found this
R> https://stackoverflow.com/questions/11900553/sqlalchemy-table-already-exists
R> and it suggest to use checkfirst=True. I'll try to test with it, R> if I'll find the way how to change schema.py (I'm not good python R> developer).
R> Regards, Rolandas
>> Please capture the whole output of qlustar-initial-config using >> >> qlustar-initial-config 2>&1 | tee log >> >> and attach that. Maybe there is some error that scrolls by too >> fast so you missed it. >> >> Regards, Goswin von Brederlow