Our app server still restarts via cron job every night at 22:30, it’s been 3 years :)

@jespersh got it in one, heap space error when we don’t restart it

@daarkfall wow...

B-but why?

I love how hacks like these become features that you use to impress the client with.

Client: We had a user tell us about a strange problem last night which occurred at... 22...
Me: (Internally panicking.) Please don't say 30, please don't say 30.
C: 30... You know anything about this, can you see anything in your logs?
M: (Panic intensifies.) Yeah we.. ummmmmmmmmmmmmm do that to keep things fresh and snappy so we automated this restart... procedure.. yes, automated restart procedure. Protocol. It's automatic. Very technical. And so good. Totally intentional, yessir, mhhmm we're well-compensated engineers after all! Gosh it's hot in here isn't it? Maybe we should automated that AC, haha... Woweeee.
C: ...
C: Very nice, impressive. Thank you!

Of course they bring it up the next meeting so that everyone will know just how automatic and cool their system is and you just pray to every deity that no one is technical enough to call out the BS. God, I hate hacks. So much anxiety lol

@daarkfall @ihatecomputers
Make it into a feature with a load balancer and graceful handoff to the new instance 😊

I saw a similar solution with an "integration app", it was just an app that reads data from a legacy system (in Fox Pro) and send these data to a REST endpoint... About 23 and a half hours the app stops to send requests, but still reading the legacy system database... A colleague work for a week to make a job to restart the service every 23 hours...

It took 3 hours of debugging to realize that it was a connection leak and fix it...

Yeah, a lot of bad java code out there. I know lots of places where servers are restarted because of this.
Not that my company had such an issue. *Laugh*...

Impressed that you can restart it every night without it failing to start back up

I guess that's the nice thing about php's lifecycle model. No worries about a continually running process with compounding memory issues.

Has it ever failed to restart?

@nathan815 touch wood, no!
the app server isn’t actually bad it’s just got a memory leak that we’ve never had budget to fix