When connected to a HikariCP, leaks accumulate and crash occurs.
GeorgyCherkasov opened this issue ยท 15 comments
Huh weird, do you have the full error?
It shouldn't be leaking connections, the connection pool takes care of that.
The [luckperms-scheduler-worker-x]
log prefix in your screenshot does not line up with what I'd expect to see from the latest version - this was changed a while ago. Are you sure you're definitely using the version you say you are?
I'd also expect to at least see some other messages from LuckPerms, but if you say so....
Unless you're willing to cooperate there's very little I can do to actually help.
There is an error at the bottom.
2020-05-23-1.log
We can agree that the DB server does not accept connections. But clearly the crash connection - the fault of LuckPerms.
P.S. our build has a plugin that works with the same DB on the same server using HikariCP. So there this problem is not noticed, which means that LuckPerms is to blame.
"Crash connection"? Sorry, I don't know what you're referring to.
If you can point to a specific Hikari configuration option that you think is causing the problem then I'm happy to try changing things.
However, the current configuration is working perfectly fine for thousands of other servers - if there was such an egregious error in the way that LP was handling connections, you would not be the only person experiencing them.
Which leaves me to suggest reading the wiki article I linked above - that is my exhaustive advice regarding connection issues - I've got nothing else to share.
Alright thanks, I've had a quick look, seems that your database is just not responding to connections - there's nothing else in the log which would indicate it's a problem caused by LP.
There are some more details about these sort of problems here:
https://github.com/lucko/LuckPerms/wiki/Storage-system-errors