dito

Keeping fighting the good fight Frog.

Like llama I've 10 years of Unix admin but have been too long out of that loop to be much use.

I believe THG run a linux dedi and had similar, if not as bad issues I don't know if anyone has contacts over there?

It would be nice of Valve to dish out special hats to those admins who keep tf2 servers afloat.

A penguin hat now that's an idea.

I play there a lot and I can confirm they have the same lag issues as we do. No contacts to speak of though, sorry.

THGs lag problems have never, in my experience, reached the levels seen on HH, and as such the chances of them running an identical, and therefore relevant to this conversation, server set-up to HH are extremely low.

I actually had a, brief, conversation with whoever was manning the THG server manager one evening last week.

They attributed their problems to server hardware, not software.

Of course, I've no idea who I was talking to, nor if they genuinely knew what the problem was, so take this with a grain of salt.

Additionally, reaching out to another server for advice on resolving an issue that they themselves haven't fixed (unless something changed in the past couple of days) seems rather pointless, no?

Of course, frog has already pointed this out.



There might be 1 page, maybe 2, of relevant information in this entire 26 page thread. If you're not reading frog's posts, you're not helping.

That said, the server was much improved last night.

Pleasingly less lag and spikes tonite (friday) so what ever you did , it seemed to work

Although there were a couple of minor spikes reported

Boris

The THG boys have been having operator trouble the past few weeks, but they've never had the lag problems we've had.

07-06-2011 7pm-10pm - HH1 seemed laggy this evening. Did anything change recently?

frog, I'm curious if you're running* CentOS 6.0 yet?

I spent some time trying to figure out if the 6.0 kernel would help improve SRCDS performance out of the box, but I couldn't find much literature on the topic**.

You may have already seen it/disregarded it, but I did find people consistently pointing to this German site http://www.ulrich-block.de/?page_id=156 for help with pre-compiled kernels.

*I'm vaguely aware that you're busy/on holiday at the moment, & very aware that you're under no obligation to update the server as soon as humanly possible, so take your time =)

** the more I read about the subject, the more I see people favouring Debian/Ubuntu over CentOS/Red Hat for running a game server - this is entirely anecdotal.

I'm away a lot. It's festival/out door party season and I have a lot of events to do.

We are still on 5.6. I was hoping there would be a work-around/hack to upgrade to 6.0 without a reinstall, but there isn't any option that won't result in a mess.

There is a period of downtime coming. Please refer to viewtopic.php?f=13&t=7742 , during this downtime we may well switch to Debian. There is no scheduled date for this yet. This link also explains a possible reason for the occasional periods of constant lag.

Also have a read through this very recent thread "[hlds_linux] orangebox cpu usage" at http://www.mail-archive.com/hlds_linux@ ... 62269.html particularly Nephyrin Zey's post, who is a reliable source, knows what he is talking about and is also a valuable contributor to the TF2 community.

It's not just us that are having problems. Other Linux server operators are experiencing the same thing. Valve needs to optimise OB to make better use of newer processors.

Thanks for the reply, frog.

It does appear that Valve have been more than negligent developing/optimising TF2 post-launch on both the client & server sides, with this being a far too common sight:


I hope you know that I only ask because I'm interested in the subject/server situation. I'm not intentionally poking a wound.