Our database have 6+GB, 1200w+ lines of maas event in database.
I’m suspecting that this is the real cause of machine listing issue (see Updated to 2.6 from 2.5 now UI slow and not usable and https://bugs.launchpad.net/maas/+bug/1835316)
I saw that there was a similar issue and mark resolved https://bugs.launchpad.net/maas/+bug/1403080
not sure it’s a regression or we should let this table grow?
Can the data inside be safely deleted?
Following the @hyuwang question… Is there any garbage collector for the DB? There are tons of old events/data there
We don’t have any garbage collection at the moment but are discussing ideas.
It’s safe to truncate this table, assuming you no longer care about historical events.
IMHO, having such a large number of events shouldn’t be a case of performance issues because the UI shouldn’t even try to load all of them. In fact, it should be only loading:
So I believe this has already been fixed on the UI perspective?
ya, problem gone after merge the UI fix
sorry, let me put my words back. just a few days after the truncation, our table once again grow up to 500M+ rows, and everything become slow again.
the problem still exists after the UI fix, I found some extremely complex query which join multiple table with maasserver_event. See the detail in my other post here Updated to 2.6 from 2.5 now UI slow and not usable
