Losing Key packets

_DP_CEO · 2004-05-31 04:59:38 UTC

PB has been kicking everyone from our servers starting today from that error. PBweb both client and serverside arent working. Is anyone else getting this?

Azarael · 2004-05-31 05:13:55 UTC

I noticed this as well today. I updated my pb files and I was still getting kicked for client authentication and key loss.
The only other thing that I can think of that would be causing this is etpro’s cheat detection, but I doubt it.

Slappy · 2004-05-31 05:29:04 UTC

Yes! And it damn near drove me insane for a few hours. Here’s what I did, until PB gets their act together:

Removed all the new files from the /pb folder on the server. The files “pbcls.so” & “pbsv.so” in the main folder should have dupes with “old” appended to them. Move or delete those new files & rename the old ones to the proper names.
Change the permissions on the /pb folder (if applicable) so the PB Master Server won’t be able to send the new update again, rendering you pissed off & back at step one.
Restart & hope everything goes well.

I must make the point that if this in any way, shape or form destroys any personal property, in this dimension or any other, I am in no way responsible.

The only downside I’ve found is that it spams one of the PB warnings about non-standard cahracters, more than usual. Give it a shot though, should work!

Chester · 2004-05-31 05:34:59 UTC

OMG…I’m getting the same problem. WTH!!

_DP_CEO · 2004-05-31 06:14:30 UTC

looks like they fixed it…

Sunday 05.30.2004 [11:30PM]

Version 1.082 of the PB Client for ET has been released to our PB Master Servers for auto-update and to our website download page.

Release Notes for PB Client v1.069:

fixes bug with v1.081 released earlier today that was causing players to get kicked
starting with this version 1.082, Win32 versions of PB now require the game to be run as the administrator user or equivalent

SiliconSlick · 2004-05-31 16:55:21 UTC

AFAIC, they broke it even more.

SiliconSlick (who can’t even get the game to load with 1.082 on Linux)

RivrStyx · 2004-05-31 18:15:35 UTC

servers with etpro should maybe turn PB off since it has its own cvar checking and anti-cheat… What a hastle.

Slappy · 2004-06-01 01:40:18 UTC

You mean if you can trust the information in the ET Pro cheat file, which an admin can’t even read or draw their own conclusions? Nah, I’ll stick with PB until Bani makes a tool for admins to access that information.

SCDS_reyalP · 2004-06-01 02:09:30 UTC

ETPro AC has its share of flaws, and PB has some pretty useful features. Better to convince evenbalance to fix their s*** and use both.

Fenris · 2004-06-01 06:20:38 UTC

Does anyone know whether this is related to the server session locking up at random interval on Linux ?

After my failed attempt last month to switch to Windows I went back to my old Linux installation and the server has ran perfectly for the last two weeks until sometime between thursday-saturday last week when it started locking up completely, (and the “watchdog” no longer kicks in either) forcing me to kill the console shell and start a new one all the time.

RivrStyx · 2004-06-01 15:09:41 UTC

I definately will use both… just right now it isn’t happening.

Fenris · 2004-06-01 20:11:25 UTC

Since disabling PB this morning the server has ran fine all day, but I’ll wait a few days before I will really know whether that is the cause, and then I’ll just have to hope the next version will solve the problem, whenever that is released…

Lekdevil.NL · 2004-06-01 21:09:05 UTC

Fenris,

I’m running ET my FreeBSD (Linux compat mode) server. The ET server process was running at 100% CPU intermittently after the 1.089 PB server update last Sunday, with no players connected. The console was unresponsive during those periods, but strangely enough there wasn’t any apparent lag when connecting to it as a player. Since I’ve disabled PB on the server, the idle load has returned to normal values (2-3%).

I too think there’s a bug in the PB server version that’s causing the high CPU utilisation.

Cheers.

last pid: 12763;  load averages:  1.00,  1.01,  1.00   up 20+10:48:58  07:53:15
89 processes:  4 running, 85 sleeping
CPU states: 18.5% user,  0.0% nice, 31.7% system,  0.0% interrupt, 49.8% idle
Mem: 310M Active, 870M Inact, 237M Wired, 67M Cache, 163M Buf, 20M Free
Swap: 1024M Total, 320K Used, 1024M Free

  PID USERNAME   PRI NICE  SIZE    RES STATE  C   TIME   WCPU    CPU COMMAND
65326 wet         59   0   140M 76684K CPU1   1 548:22 98.83% 98.83% etded.x86
12763 wet         29   0  2036K  1104K CPU0   1   0:00  4.57%  0.83% top
  503 teamspeak    2   0 20856K  1344K select 0  20:18  0.15%  0.15% server_lin
  244 iplog        2   0  5096K  4248K poll   0  70:21  0.00%  0.00% iplog
  574 root         2   0  4440K  2360K select 1  41:32  0.00%  0.00% snmpd
  499 teamspeak    2   0 20856K  1344K RUN    1  28:15  0.00%  0.00% server_lin
  498 teamspeak    2   0 20856K  1344K RUN    1  25:06  0.00%  0.00% server_lin

Fenris · 2004-06-01 22:22:35 UTC

Thanks for the information, nice to know it’s not just me again :D, just fyi, I’m (back to) using RedHat 7.3.4, however I had no problems with CPU usage, as what happened on RedHat was rather that the console “froze” and the server would no longer be accessible (listed as down in ASE for example), only way out was to restart console.

After disabling PB it now seems to work fine again, which leads me to hope for a fixed version taking us back to “normal” soon, and after that I will probably consider write protecting the PB files to prevent autoupdates, since I want to use PB (as the server is used for clanwars) but it seems a controlled manual update is possibly safer than automatic update. That way it’s easier to save a backup copy of the files before upgrading them for easy revert in situations like this, rather than completely disabling PB. Not what I had wanted but I think at least I will prefer that to a general instability

SiliconSlick · 2004-06-02 13:54:36 UTC

Our servers went nuts when the last 2 PB updates came out…
they all went to 100% and the consoles were non-resposive.

Simply restarting them seemed to have corrected the problem
in the short term.

You can see when we had this problem here:

http://www.nd80usa.net/ganglia/?r=week&s=descending&c=

The red peaks on the CPU graphs for the 116 box and the
208 box show when we had these problems… the times
correlate with PB updates.

However, like I said, restarting seems to have fixed the problem.
We are still running PB ATM.

SiliconSlick

Fenris · 2004-06-02 21:51:06 UTC

Nice to hear, guess I’ll give it a try enabling PB again in a few days to see whether there’s any update available but simply restarting the server did not work for me at least as of tuesday it still hangs after a few hours running.

Usayd · 2004-06-02 22:33:58 UTC

hey guys read the first post, i think thats about it (evenbalance gone crazy)

Fenris · 2004-06-03 18:47:31 UTC

I’m already following all the posts on the topic in this forum but thanks anyway