💥 TRENDING: Groups/pikacsrd/discussions// - Complete Album!

At the time of this post, we are experiencing intermittent connectivity and latency issues. The site and game service may or may not be reachable.

There is no indication in my VPS control panel that it is an attack on this server specifically; I'll assume the networking gear in front of the server is being overwhelmed.

It appears that the connectivity issues have been resolved as of now.

The server will be down for an undetermined amount of time while I ensure a smooth transition to the updated server toolchain[forums.alliedmods.net].

Server is now back up; thank you for your patience.
Had to do some questionable package upgrades, but the server toolchain update tests were successful.

I'm doing some additional testing right now to ensure general stability, so the server may be restarted a few times.

Tests are complete; had an issue with workshop maps not downloading and that has been resolved. The TF2 server is currently on the stable release branch since there may be unexpected SourceMod breakage (though it seems to work fine), but transitioning shouldn't be too difficult now.

Experiencing some weird shutdown problems where the server begins to shut down but doesn't actually terminate. Seems to have started since the branch update.

Looking into it now.

Looks like there's some weird quirk with the main branch of the Source Dedicated Server that, with the system's particular combination of libraries, just forces the server to hang for whatever reason. Instead of segfaulting like it normally does.

I've updated it completely to the beta branch, and it seems to be shutting down like it normally does. So we're on the beta branch for now.

Something's screwy with the CPU usage on the server all of a sudden.

Vanilla installation is fine; something in the addons is breaking things. Which is weird, since I can't think of any plugin that would be an issue.

Server is down pending fixes.

Edit: Scratch that. CPU got blasted all of a sudden. 6 bots on vanilla is taking up full CPU when it normally used ~30%, even apt is incurring unusually high CPU load.

Not sure if I'll be able to get to the bottom of this before I sleep, unfortunately. Might be an extended downtime. For SRCDS, anyways. The site will still be available and all other services shouldn't be affected.

Started SRCDS again after ~17 hours and we seem to be running with the original CPU usage. Looks like we're back!

What a pain. Had to do some impromptu server debugging, since there were two bad combinations of load order:

1. Server hangs on quit.
   
2. Server does not hang on quit, but on the other hand, SteamTools (handles HTTP requests through Steam libraries) doesn't load.

Figured out that -nobreakpad worked on the server, so I decided to disable crash dumps.

Server is up now, though it was restarted a number of times over the past half hour.

Looks like I'm getting the ol' "hang on quit" stuff again, even though I didn't even touch software packages yesterday. Server was down for maybe 10 hours. Hard to say.

In any case, disabled crash reporting; we should be all good again.

Server will be down while I do an early upgrade to Debian Stretch.
Let's hope nothing catches on fire!

The update has been completed; we should be back to our regularly scheduled programming.

The server provider is currently making improvements to their control panel and billing system, and so it's down… right before I could throw my money at them to host us for another year.

Unlucky.

I hope I'm able to throw my money at them by Monday, but if not, the server may be temporarily suspended. I'm in contact with them, and I've been told that they'll lift the suspension should that end up being the case.

We are set for another year.

I removed -nobreakpad and enabled autoloading on the Accelerator extension (which does crash reporting) just to see if they're still causing issues after the major system update.

If the server hangs while attempting to restart itself (or hangs on a crash in some other fashion), it may be a bit before I wake up and find out, so apologies for the inconvenience if that happens to be the case.

Restarted the server, aaaaaaaaand now none of the services are accessible -- can't SSH nor VNC.

Not the best thing to happen right before I decide to head to bed.

Got it sorted out pretty quickly! After filing a ticket and remembering that there was a separate control panel for VNC services, I managed to get myself signed in.

Apparently Debian 9 changed the interface prefix to "ens" from "eth" to be in line with Predictable Network Interface Names[www.freedesktop.org]. The server was expecting to configure a network connection with the latter interface, so there was no network connectivity due to that.

Reverted it back through a kernel command option and we're back up. Since there aren't any hardware changes on a virtual machine, we should be fine with the legacy naming scheme for the time being, though any future machines will be configured with that in mind.

According to the hosting provider, the VPS is being migrated at the moment. Did not get any advance notice on this, unfortunately.

Apologies for the inconvenience.

Looks like it came back up around 45 minutes after I posted, when I was out running groceries. Back to business!

Valve updates their game?!

Apologies for the delay; the update didn't take due to a lack of disk space.

Server was down for about 7.5 hours; had to boot it up in the manager again just now.
Thanks for waiting.

Did a quick server reboot to update dbus-daemon; took about 5 minutes of downtime.

In the process of upgrading the server to Debian 9 → 10.
Incredibly painless compared to my experience managing a different server with Ubuntu LTS, except I do have to fix up pip.

Took about an hour to get everything sorted out (15 minutes for the apt upgrade, 45 for getting everything up and running on Python 3.7).