Re: [hlds_linux] Source and HL1 dedicated server Beta available

[Martin Zwickel]

--
On Mon, 30 Jan 2006 22:41:23 -0800
"Alfred Reynolds" <[EMAIL PROTECTED]> bubbled:

> We have a beta release available for both the Source and HL1 engine.
> This beta fixes the steam id swapping problem amongst other small
> fixes. You can apply the beta by running the hldsupdatetool and
> adding "-beta swapbeta" to the command line.
>
> A couple of users in an initial test reported a server hang and
> associated assert (pipes.cpp line 298), if you see this please send me
> your OS configuration and hardware.
>
> - Alfred

Thx Alfred! Hope it works, at least it starts ;)

--
MyExcuse:
It must have been the lightning storm we had (yesterday) (last week)
(last month)

Martin Zwickel <[EMAIL PROTECTED]>
Research & Development

TechnoTrend AG 
--
[ signature.asc of type application/pgp-signature deleted ]
--

___
To unsubscribe, edit your list preferences, or view the list archives, please 
visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux

[hlds_linux] Source and HL1 dedicated server Beta available

[Alfred Reynolds]

We have a beta release available for both the Source and HL1 engine.
This beta fixes the steam id swapping problem amongst other small fixes.
You can apply the beta by running the hldsupdatetool and adding "-beta
swapbeta" to the command line.

A couple of users in an initial test reported a server hang and
associated assert (pipes.cpp line 298), if you see this please send me
your OS configuration and hardware.

- Alfred

___
To unsubscribe, edit your list preferences, or view the list archives, please 
visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux

Re: [hlds_linux] High CPU % Choke on Source

[Rüdiger Meier]

On Tuesday 31 January 2006 00:29, Arris wrote:
> We have only one server with Athlon XP 3000+ with 1GB RAM.
> [..]
> This two server never increase 80-90% CPU
> all run fine with no laggs.

Thats alot CPU usage!
For example on a similar machine  I run 40 occupied slots CoD-1.1 and
CoD-2 on 50% CPU maximum.

CS and CSS are needing far too much CPU IMO.
My servers are mainly restricted by CPU - so in my price calculation CS
needs to be at least 50% more expensive than CoD2 allthough CS-1.6 is a
very old game.
I really cant imagine that there were CPUs who could handle 30 slots
CS-1.6 severs at the time the game came out ...

cu,
rudi

___
To unsubscribe, edit your list preferences, or view the list archives, please 
visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux

Re: [hlds_linux] client dropped by server

[William Warren]

I am using a 10-100 switch.  Everything worked fine until the latest
steam update.  Also it is only srcds having this issue.  I also run a
1.6 server on the same box and it runs w/o a hitch..same for the
teamspeak server.  A bad nic or switch would affect all those processes
not just srcds.  I have all power mangement turned off(this is a dell
pweredge sc430) both in linux and in the bios.

Kevin Ottalini wrote:


One thing to look out for on LANs (if you are seeing problems) is that you
don't have a loop feeding both ends of a router to a common switch.

I've seen odd things happening that cleared up immediately when this was
discovered.

IE: both the in and the out of a router are connected to a switch that the
client it connected to.  (Some high-end switches support private LAN
segments which is ok).

If you are by any chance using a HUB and not a SWITCH then dump the HUB
asap, hubs have serious problems dealing with different LAN rates and also
contribute a lot of unnecessary traffic on all nodes.

I have seen bad NICs scramble data too, switch to a different NIC if you
can
to test, also try forcing all connections on your LAN to be the same rates
(all 100mbit or all 1GB).

One last thing to look out for is O/S power savings for the NIC ... this
should be disabled in the hardware options if it's supported.


- Original Message -
From: "William Warren"
To: 
Sent: Monday, January 30, 2006 5:25 AM
Subject: Re: [hlds_linux] client dropped by server



kevin is referring to NAT routers used by small business or home users
not the kind of routers you are ferrreing to..:)

however this issue is not a nat issue as i get booted by my server and
it is on my lan.


kama wrote:

Ok, let me clarify.

What can cause it in a router the servers are attached too?
Especially in
the case of use of Cisco routers in the 7600-series.

/Bjorn

On Sun, 29 Jan 2006, Kevin Ottalini wrote:


Each router and router firmware version has a unique way of mangling
UDP
packets, apparently UDP is not as rigorously tested as TCP.

The exact cause is unknown but the proof is always the same, if you
bypass
the router the problem goes away.

In fact, there are dozens of different error messages that can be
generated,
not just CL_ParseServerMessage, but the cause is always the router,
here
are
a few of the most typical messages:

Bad Parse in Client Command
Bad Command Character in Client Mod
Bad Command Character in Client Command
CMD_MAXBACKUP HIT
Host_Error: CL_ParseServerMessage: Illegible server message - svc_bad
Host_Error: CL_Parse_Version: Server is protocol -# instead of
47
Host_Error: DispatchUserMsg: User Msg CZCareer/130 sent too much data
(195
bytes), 192 bytes max
svc_bad Wrote erroneous message to buffer.dat Host_Error: UserMsg: Not
Present on Client ###
Memory Could Not Be Read Error (one cause of this)
Host error: Cl_Entitynum:  is an invalid number, Cl_max edicts is

Host error: Bad server message

and the most typical symptom is:
Whenever I connect to a (Counter-Strike) server, I get booted after 30
seconds

I have a FAQ about this in the Steam Support database:
http://support.steampowered.com/cgi-bin/steampowered.cfg/php/enduser/std_adp.php?p_faqid=19



- Original Message -
From: "kama"
To: 
Sent: Sunday, January 29, 2006 3:34 PM
Subject: Re: [hlds_linux] client dropped by server



Ok, if so... What are the reasons that causes these to happens then?

/Bjorn

On Sun, 29 Jan 2006, Kevin Ottalini wrote:


CL_ParseServerMessage is very often caused by routers.

- Original Message -
From: "kama"
To: 
Sent: Sunday, January 29, 2006 3:14 PM
Subject: Re: [hlds_linux] client dropped by server



Hmm.. Those two disconnects where probably not the same problem. I
have
reports from serveral admins and regular that they currently are
being
dropped. I got kicked myself once and this was found in my client
console.

Error: server failed to transmit file 'AY&SY-B^K'
Last 32 messages parsed.
4039 0098 ResetHUD
4039 0099 Money
4039 0105 StatusIcon
4039 0111 ScoreAttrib
4039 0114 RoundTime
4039 0117 TeamScore
4039 0124 TeamScore
4039 0138 HideWeapon
4039 0140 SetFOV
4039 0142 Health
4039 0144 Battery
4039 0147 Damage
4039 0160 FlashBat
4039 0162 Train
4039 0164 CurWeapon
4039 0168 WeapPickup
4039 0170 WeapPickup
4039 0172 AmmoPickup
4039 0175 ScoreAttrib
4039 0178 ScreenFade
4039 0189 RoundTime
4039 0192 ScoreAttrib
4039 0195 ScoreAttrib
4039 0198 svc_event_reliable
4039 0201 AmmoPickup
4039 0204 WeapPickup
4039 0206 StatusIcon
4039 0215 ScoreAttrib
4039 0218 CurWeapon
4039 0222 CurWeapon
4039 0226 svc_filetxferfailed
BAD:  237:svc_bad
Host_Error: CL_ParseServerMessage:

/Bjorn




___
To unsubscribe, edit your list preferences, or view the list archives,
please visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux



--
My "Foundation" verse:
Isa 54:17  No weapon that is formed against thee shall prosper; and
every tongue that shall rise agai

Re: [hlds_linux] High CPU % Choke on Source

[Arris]

We have only one server with Athlon XP 3000+ with 1GB RAM.
Debian sarge
Kernel 2.6.11

a couple weeks ago we run two servers, one 1.6, one source

1.6 server:
CS 1.6 with WC3-mode
amxx
VAC2
12 slots

source server:
CSS
mani_admin 1.1.0zh
cvar_plugin (ESL) 0.2 Online Edition
statsxs 0.5.1c
tickrate 100
18 slots

This two server never increase 80-90% CPU
all run fine with no laggs.

After the latest two updates, particularly the newer maps run out on
CPU-Usage.
Now, only with the CSS-Server, the CPU go up to 85%.

I don't run tests with a clean, fresh server without plugins.
testing with "top"

regards
Arris

___
To unsubscribe, edit your list preferences, or view the list archives, please 
visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux

RE: [hlds_linux] Wallhack Bug?

[John McCollor]

-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Brian "Duck"
Holmes
Sent: Wednesday, January 18, 2006 3:24 AM
To: hlds_linux@list.valvesoftware.com
Subject: RE: [hlds_linux] Wallhack Bug?

Not my site.

The link to the video was pasted a couple places discussing the bug.  I
simply reposted it.

-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of
[EMAIL PROTECTED]
Sent: Tuesday, January 17, 2006 11:27 PM
To: hlds_linux@list.valvesoftware.com
Subject: RE: [hlds_linux] Wallhack Bug?

WAHAHAHAHAAA
Bandwidth Limit Exceeded
The server is temporarily unable to service your request due to the site
owner reaching his/her bandwidth limit. Please try again later.




Apache/1.3.33 Server at www.booom.co.uk Port 80


Sorry bro but when you post something like a video to this list you need to
make sure you can support 100+ guys going to check it out. I missed my
chance but you might want to take care next time.

Rayne

-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Brian "Duck"
Holmes
Sent: Tuesday, January 17, 2006 11:53 PM
To: hlds_linux@list.valvesoftware.com
Subject: RE: [hlds_linux] Wallhack Bug?

Yeah there's another similar one for decals. And one that can make ambient
sounds really loud, not sure if this has any advantage though

before:
http://img235.imageshack.us/img235/4365/dedust00058ds.jpg

after:
http://img235.imageshack.us/img235/6361/dedust00029ts.jpg

Video:
http://www.booom.co.uk/downloads/exploit/

There's a plugin called Zblock that purports to fix it, and I started a
thread to ask mani to make his CVAR block configurable to correct this as
well.



>>Okay I hope I'm not bringing something up that's already been
discussed I
>>went through the list for about 100 emails and didn't see it so if I
am
>>posting something that has already been stated I apologize.
Apparently
>>according to a few clients since the latest update there is a command
via
>>the console that can be given to allow wall hacking?  I don't have
many
>>details on this yet but I am trying to get them.  Has anyone heard
this >>yet
>>or am I totally out of line here.  An example of the command is listed
>>below.




___
To unsubscribe, edit your list preferences, or view the list archives,
please visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux


___
To unsubscribe, edit your list preferences, or view the list archives,
please visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux




___
To unsubscribe, edit your list preferences, or view the list archives,
please visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux


___
To unsubscribe, edit your list preferences, or view the list archives, please 
visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux

Re: [hlds_linux] client dropped by server

[Kevin Ottalini]

One thing to look out for on LANs (if you are seeing problems) is that you
don't have a loop feeding both ends of a router to a common switch.

I've seen odd things happening that cleared up immediately when this was
discovered.

IE: both the in and the out of a router are connected to a switch that the
client it connected to.  (Some high-end switches support private LAN
segments which is ok).

If you are by any chance using a HUB and not a SWITCH then dump the HUB
asap, hubs have serious problems dealing with different LAN rates and also
contribute a lot of unnecessary traffic on all nodes.

I have seen bad NICs scramble data too, switch to a different NIC if you can
to test, also try forcing all connections on your LAN to be the same rates
(all 100mbit or all 1GB).

One last thing to look out for is O/S power savings for the NIC ... this
should be disabled in the hardware options if it's supported.


- Original Message -
From: "William Warren"
To: 
Sent: Monday, January 30, 2006 5:25 AM
Subject: Re: [hlds_linux] client dropped by server



kevin is referring to NAT routers used by small business or home users
not the kind of routers you are ferrreing to..:)

however this issue is not a nat issue as i get booted by my server and
it is on my lan.


kama wrote:

Ok, let me clarify.

What can cause it in a router the servers are attached too? Especially in
the case of use of Cisco routers in the 7600-series.

/Bjorn

On Sun, 29 Jan 2006, Kevin Ottalini wrote:


Each router and router firmware version has a unique way of mangling UDP
packets, apparently UDP is not as rigorously tested as TCP.

The exact cause is unknown but the proof is always the same, if you
bypass
the router the problem goes away.

In fact, there are dozens of different error messages that can be
generated,
not just CL_ParseServerMessage, but the cause is always the router, here
are
a few of the most typical messages:

Bad Parse in Client Command
Bad Command Character in Client Mod
Bad Command Character in Client Command
CMD_MAXBACKUP HIT
Host_Error: CL_ParseServerMessage: Illegible server message - svc_bad
Host_Error: CL_Parse_Version: Server is protocol -# instead of
47
Host_Error: DispatchUserMsg: User Msg CZCareer/130 sent too much data
(195
bytes), 192 bytes max
svc_bad Wrote erroneous message to buffer.dat Host_Error: UserMsg: Not
Present on Client ###
Memory Could Not Be Read Error (one cause of this)
Host error: Cl_Entitynum:  is an invalid number, Cl_max edicts is

Host error: Bad server message

and the most typical symptom is:
Whenever I connect to a (Counter-Strike) server, I get booted after 30
seconds

I have a FAQ about this in the Steam Support database:
http://support.steampowered.com/cgi-bin/steampowered.cfg/php/enduser/std_adp.php?p_faqid=19


- Original Message -
From: "kama"
To: 
Sent: Sunday, January 29, 2006 3:34 PM
Subject: Re: [hlds_linux] client dropped by server



Ok, if so... What are the reasons that causes these to happens then?

/Bjorn

On Sun, 29 Jan 2006, Kevin Ottalini wrote:


CL_ParseServerMessage is very often caused by routers.

- Original Message -
From: "kama"
To: 
Sent: Sunday, January 29, 2006 3:14 PM
Subject: Re: [hlds_linux] client dropped by server



Hmm.. Those two disconnects where probably not the same problem. I
have
reports from serveral admins and regular that they currently are
being
dropped. I got kicked myself once and this was found in my client
console.

Error: server failed to transmit file 'AY&SY-B^K'
Last 32 messages parsed.
4039 0098 ResetHUD
4039 0099 Money
4039 0105 StatusIcon
4039 0111 ScoreAttrib
4039 0114 RoundTime
4039 0117 TeamScore
4039 0124 TeamScore
4039 0138 HideWeapon
4039 0140 SetFOV
4039 0142 Health
4039 0144 Battery
4039 0147 Damage
4039 0160 FlashBat
4039 0162 Train
4039 0164 CurWeapon
4039 0168 WeapPickup
4039 0170 WeapPickup
4039 0172 AmmoPickup
4039 0175 ScoreAttrib
4039 0178 ScreenFade
4039 0189 RoundTime
4039 0192 ScoreAttrib
4039 0195 ScoreAttrib
4039 0198 svc_event_reliable
4039 0201 AmmoPickup
4039 0204 WeapPickup
4039 0206 StatusIcon
4039 0215 ScoreAttrib
4039 0218 CurWeapon
4039 0222 CurWeapon
4039 0226 svc_filetxferfailed
BAD:  237:svc_bad
Host_Error: CL_ParseServerMessage:

/Bjorn




___
To unsubscribe, edit your list preferences, or view the list archives, please 
visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux

Re: [hlds_linux] client dropped by server

[Matt Judge]

The only time I have seen svc_bad messages and clients dropping is when
I have been sending the client too much traffic from my plugin or when
my plugin spends too much time outputting debug information and the
client times out.

Make sure you are not flooding your network with data and that it is all
switched and performing correctly.

Matt.

kama randomly typed the following on 01/30/2006 12:34 AM:

Ok, if so... What are the reasons that causes these to happens then?





___
To unsubscribe, edit your list preferences, or view the list archives, please 
visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux

Re: [hlds_linux] renting a box

[kama]

As he stated private servers, I was in to the influence that he meant
warservers. So all the rates should be set to max, without getting any
lag. sys_tickrate should be set to a value that makes the server respond
quick enough. I normaly have set sys_ticrate somewhere between 200-300.
Setting it higher does not help. I have tried really high values, but
instead got decreased performance and some strange lag even though the CPU
wasn't maxed out.

Even if they are empty 95% of the time, the last 5% may be occupied at the
same time. So you should not run more servers than it can handle at peak.
Overbooking in a gameserver environment is not a good thing. It's
timecritical and needs to be operate smoothly. Its not like a webserver,
where you can wait a couple of extra seconds for the server to respond to
your request.

/Bjorn

On Mon, 30 Jan 2006, Gamefuel Support wrote:

> True, but it still depends on what his settings are (tickrate,
> sv_maxrate, etc, etc) and how busy these servers will be. The only way
> to find this out, is to see for yourself what is the max you can squeeze
> out of it (if squeezing the max out of your server is what you want).
> If they are going to be empty 95% of the time, you can run as many as
> your 3GB mem allows you to.
>
> kama wrote:
>
> >Actually he states that he want to run X amount of 12 slot hlds servers on
> >a specific hardware.
> >
> >I never used amd procs with hlds, so I dont know how many that will fit
> >his hardware. But a guess would range from 3 to 6. I run 4 16 player
> >servers on dual Xeon 2.4GHz.
> >
> >/Bjorn
> >
> >On Mon, 30 Jan 2006, Gamefuel Support wrote:
> >
> >
> >
> >>If you read the email, that was sent last before this email by Scott
> >>Pettit, wou will have some indication of what to expect.
> >>Other then that.. I would think about rephrasing your question if i were
> >>you. This is like asking: How many potatoes fit in a bag?
> >>Depends on the size of the potatoes and the bag etc etc.
> >>
> >>Jean-Philippe Caron wrote:
> >>
> >>
> >>
> >>>--
> >>>[ Picked text/plain from multipart/alternative ]
> >>>I'm about to rent a box in chicago (Equinix)
> >>>
> >>>amd athlon64 3500+
> >>>3 gig ram
> >>>linux centos
> >>>
> >>>how many counter-strike server will I be able to run on it
> >>>im talking about 12slot private
> >>>
> >>>or my box specs wont do anything good for game hosting
> >>>--
> >>>
> >>>___
> >>>To unsubscribe, edit your list preferences, or view the list archives, 
> >>>please visit:
> >>>http://list.valvesoftware.com/mailman/listinfo/hlds_linux
> >>>
> >>>
> >>>
> >>>
> >>--
> >>Gamefuel
> >>[EMAIL PROTECTED]
> >>http://www.gamefuel.nl
> >>
> >>
> >>
> >>___
> >>To unsubscribe, edit your list preferences, or view the list archives, 
> >>please visit:
> >>http://list.valvesoftware.com/mailman/listinfo/hlds_linux
> >>
> >>
> >>
> >
> >___
> >To unsubscribe, edit your list preferences, or view the list archives, 
> >please visit:
> >http://list.valvesoftware.com/mailman/listinfo/hlds_linux
> >
> >
>
>
> --
> Gamefuel
> [EMAIL PROTECTED]
> http://www.gamefuel.nl
>
>
>
> ___
> To unsubscribe, edit your list preferences, or view the list archives, please 
> visit:
> http://list.valvesoftware.com/mailman/listinfo/hlds_linux
>

___
To unsubscribe, edit your list preferences, or view the list archives, please 
visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux

Re: [hlds_linux] renting a box

[Gamefuel Support]

True, but it still depends on what his settings are (tickrate,
sv_maxrate, etc, etc) and how busy these servers will be. The only way
to find this out, is to see for yourself what is the max you can squeeze
out of it (if squeezing the max out of your server is what you want).
If they are going to be empty 95% of the time, you can run as many as
your 3GB mem allows you to.

kama wrote:


Actually he states that he want to run X amount of 12 slot hlds servers on
a specific hardware.

I never used amd procs with hlds, so I dont know how many that will fit
his hardware. But a guess would range from 3 to 6. I run 4 16 player
servers on dual Xeon 2.4GHz.

/Bjorn

On Mon, 30 Jan 2006, Gamefuel Support wrote:




If you read the email, that was sent last before this email by Scott
Pettit, wou will have some indication of what to expect.
Other then that.. I would think about rephrasing your question if i were
you. This is like asking: How many potatoes fit in a bag?
Depends on the size of the potatoes and the bag etc etc.

Jean-Philippe Caron wrote:




--
[ Picked text/plain from multipart/alternative ]
I'm about to rent a box in chicago (Equinix)

amd athlon64 3500+
3 gig ram
linux centos

how many counter-strike server will I be able to run on it
im talking about 12slot private

or my box specs wont do anything good for game hosting
--

___
To unsubscribe, edit your list preferences, or view the list archives, please 
visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux





--
Gamefuel
[EMAIL PROTECTED]
http://www.gamefuel.nl



___
To unsubscribe, edit your list preferences, or view the list archives, please 
visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux





___
To unsubscribe, edit your list preferences, or view the list archives, please 
visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux





--
Gamefuel
[EMAIL PROTECTED]
http://www.gamefuel.nl



___
To unsubscribe, edit your list preferences, or view the list archives, please 
visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux

Re: [hlds_linux] renting a box

[kama]

Actually he states that he want to run X amount of 12 slot hlds servers on
a specific hardware.

I never used amd procs with hlds, so I dont know how many that will fit
his hardware. But a guess would range from 3 to 6. I run 4 16 player
servers on dual Xeon 2.4GHz.

/Bjorn

On Mon, 30 Jan 2006, Gamefuel Support wrote:

> If you read the email, that was sent last before this email by Scott
> Pettit, wou will have some indication of what to expect.
> Other then that.. I would think about rephrasing your question if i were
> you. This is like asking: How many potatoes fit in a bag?
> Depends on the size of the potatoes and the bag etc etc.
>
> Jean-Philippe Caron wrote:
>
> >--
> >[ Picked text/plain from multipart/alternative ]
> >I'm about to rent a box in chicago (Equinix)
> >
> >amd athlon64 3500+
> >3 gig ram
> >linux centos
> >
> >how many counter-strike server will I be able to run on it
> >im talking about 12slot private
> >
> >or my box specs wont do anything good for game hosting
> >--
> >
> >___
> >To unsubscribe, edit your list preferences, or view the list archives, 
> >please visit:
> >http://list.valvesoftware.com/mailman/listinfo/hlds_linux
> >
> >
>
>
> --
> Gamefuel
> [EMAIL PROTECTED]
> http://www.gamefuel.nl
>
>
>
> ___
> To unsubscribe, edit your list preferences, or view the list archives, please 
> visit:
> http://list.valvesoftware.com/mailman/listinfo/hlds_linux
>

___
To unsubscribe, edit your list preferences, or view the list archives, please 
visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux

Re: [hlds_linux] client dropped by server

[William Warren]

kevin is referring to NAT routers used by small business or home users
not the kind of routers you are ferrreing to..:)

however this issue is not a nat issue as i get booted by my server and
it is on my lan.


kama wrote:

Ok, let me clarify.

What can cause it in a router the servers are attached too? Especially in
the case of use of Cisco routers in the 7600-series.

/Bjorn

On Sun, 29 Jan 2006, Kevin Ottalini wrote:


Each router and router firmware version has a unique way of mangling UDP
packets, apparently UDP is not as rigorously tested as TCP.

The exact cause is unknown but the proof is always the same, if you bypass
the router the problem goes away.

In fact, there are dozens of different error messages that can be generated,
not just CL_ParseServerMessage, but the cause is always the router, here are
a few of the most typical messages:

Bad Parse in Client Command
Bad Command Character in Client Mod
Bad Command Character in Client Command
CMD_MAXBACKUP HIT
Host_Error: CL_ParseServerMessage: Illegible server message - svc_bad
Host_Error: CL_Parse_Version: Server is protocol -# instead of 47
Host_Error: DispatchUserMsg: User Msg CZCareer/130 sent too much data (195
bytes), 192 bytes max
svc_bad Wrote erroneous message to buffer.dat Host_Error: UserMsg: Not
Present on Client ###
Memory Could Not Be Read Error (one cause of this)
Host error: Cl_Entitynum:  is an invalid number, Cl_max edicts is 
Host error: Bad server message

and the most typical symptom is:
Whenever I connect to a (Counter-Strike) server, I get booted after 30
seconds

I have a FAQ about this in the Steam Support database:
http://support.steampowered.com/cgi-bin/steampowered.cfg/php/enduser/std_adp.php?p_faqid=19


- Original Message -
From: "kama"
To: 
Sent: Sunday, January 29, 2006 3:34 PM
Subject: Re: [hlds_linux] client dropped by server



Ok, if so... What are the reasons that causes these to happens then?

/Bjorn

On Sun, 29 Jan 2006, Kevin Ottalini wrote:


CL_ParseServerMessage is very often caused by routers.

- Original Message -
From: "kama"
To: 
Sent: Sunday, January 29, 2006 3:14 PM
Subject: Re: [hlds_linux] client dropped by server



Hmm.. Those two disconnects where probably not the same problem. I have
reports from serveral admins and regular that they currently are being
dropped. I got kicked myself once and this was found in my client
console.

Error: server failed to transmit file 'AY&SY-B^K'
Last 32 messages parsed.
4039 0098 ResetHUD
4039 0099 Money
4039 0105 StatusIcon
4039 0111 ScoreAttrib
4039 0114 RoundTime
4039 0117 TeamScore
4039 0124 TeamScore
4039 0138 HideWeapon
4039 0140 SetFOV
4039 0142 Health
4039 0144 Battery
4039 0147 Damage
4039 0160 FlashBat
4039 0162 Train
4039 0164 CurWeapon
4039 0168 WeapPickup
4039 0170 WeapPickup
4039 0172 AmmoPickup
4039 0175 ScoreAttrib
4039 0178 ScreenFade
4039 0189 RoundTime
4039 0192 ScoreAttrib
4039 0195 ScoreAttrib
4039 0198 svc_event_reliable
4039 0201 AmmoPickup
4039 0204 WeapPickup
4039 0206 StatusIcon
4039 0215 ScoreAttrib
4039 0218 CurWeapon
4039 0222 CurWeapon
4039 0226 svc_filetxferfailed
BAD:  237:svc_bad
Host_Error: CL_ParseServerMessage:

/Bjorn



___
To unsubscribe, edit your list preferences, or view the list archives, please 
visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux



___
To unsubscribe, edit your list preferences, or view the list archives, please 
visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux



--
My "Foundation" verse:
Isa 54:17  No weapon that is formed against thee shall prosper; and
every tongue that shall rise against thee in judgment thou shalt
condemn. This is the heritage of the servants of the LORD, and their
righteousness is of me, saith the LORD.

-- carpe ductum -- "Grab the tape"
CDTT (Certified Duct Tape Technician)

Linux user #322099
Machines:
206822
256638
276825
http://counter.li.org/

___
To unsubscribe, edit your list preferences, or view the list archives, please 
visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux

Re: [hlds_linux] High CPU % Choke on Source

[William Warren]

i run a tickrate of 33 BTW.  High tickrates will spike cpu usage.

flags wrote:

This is a multi-part message in MIME format.
--
[ Picked text/plain from multipart/alternative ]
Since the December 5th time frame,  we have had this intermittent choke.
Normally Choke is related to network issues.
We have done extensive testing and know the network is not the case.
Currently we are running on RedHatEnterprise 4 on dual 2.8 Zeon's, with 2
Gig ram.  The choke jumps from 0-65.  We have tested in multiple
datacenters, and we have the same problem.  We have even loaded Windows 2003
and we still get some choke, but the choke is only high when you first join.
It may jump to 10-15 but not as bad as on Linux.

The high CPU also showed about in this same time frame.  We had Enterprise 3
and had high CPU, 12-14 man server would hit 100%  So we then built a custom
kernel and got CPU % down.  Next we loaded Enterprise 4 a major release of
Linux and again CPU % is back up.  A single 20 man server will hit 99% on a
dual 2.8 Zeon.  Now what is annoying is prior to the December 5th time frame
we could easily run a 32 with in the 80% range and zero choke.


Does anyone have a way around these problems?

Does Steam plan on supporting major releases of Linux? Or does steam plan on
hoping more 14 year olds will host servers at the homes on Windows?



--



___
To unsubscribe, edit your list preferences, or view the list archives, please 
visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux



--
My "Foundation" verse:
Isa 54:17  No weapon that is formed against thee shall prosper; and
every tongue that shall rise against thee in judgment thou shalt
condemn. This is the heritage of the servants of the LORD, and their
righteousness is of me, saith the LORD.

-- carpe ductum -- "Grab the tape"
CDTT (Certified Duct Tape Technician)

Linux user #322099
Machines:
206822
256638
276825
http://counter.li.org/

___
To unsubscribe, edit your list preferences, or view the list archives, please 
visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux

Re: [hlds_linux] High CPU % Choke on Source

[William Warren]

I am noticing choke as well.  I run CentOS-4.2.  however i have a 12
person source server but cpu usage never goes above 50% on average.  I
run a celey 2.45 ghz machine with 512 megs of ram.

flags wrote:

This is a multi-part message in MIME format.
--
[ Picked text/plain from multipart/alternative ]
Since the December 5th time frame,  we have had this intermittent choke.
Normally Choke is related to network issues.
We have done extensive testing and know the network is not the case.
Currently we are running on RedHatEnterprise 4 on dual 2.8 Zeon's, with 2
Gig ram.  The choke jumps from 0-65.  We have tested in multiple
datacenters, and we have the same problem.  We have even loaded Windows 2003
and we still get some choke, but the choke is only high when you first join.
It may jump to 10-15 but not as bad as on Linux.

The high CPU also showed about in this same time frame.  We had Enterprise 3
and had high CPU, 12-14 man server would hit 100%  So we then built a custom
kernel and got CPU % down.  Next we loaded Enterprise 4 a major release of
Linux and again CPU % is back up.  A single 20 man server will hit 99% on a
dual 2.8 Zeon.  Now what is annoying is prior to the December 5th time frame
we could easily run a 32 with in the 80% range and zero choke.


Does anyone have a way around these problems?

Does Steam plan on supporting major releases of Linux? Or does steam plan on
hoping more 14 year olds will host servers at the homes on Windows?



--



___
To unsubscribe, edit your list preferences, or view the list archives, please 
visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux



--
My "Foundation" verse:
Isa 54:17  No weapon that is formed against thee shall prosper; and
every tongue that shall rise against thee in judgment thou shalt
condemn. This is the heritage of the servants of the LORD, and their
righteousness is of me, saith the LORD.

-- carpe ductum -- "Grab the tape"
CDTT (Certified Duct Tape Technician)

Linux user #322099
Machines:
206822
256638
276825
http://counter.li.org/

___
To unsubscribe, edit your list preferences, or view the list archives, please 
visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux

Re: [hlds_linux] renting a box

[Gamefuel Support]

If you read the email, that was sent last before this email by Scott
Pettit, wou will have some indication of what to expect.
Other then that.. I would think about rephrasing your question if i were
you. This is like asking: How many potatoes fit in a bag?
Depends on the size of the potatoes and the bag etc etc.

Jean-Philippe Caron wrote:


--
[ Picked text/plain from multipart/alternative ]
I'm about to rent a box in chicago (Equinix)

amd athlon64 3500+
3 gig ram
linux centos

how many counter-strike server will I be able to run on it
im talking about 12slot private

or my box specs wont do anything good for game hosting
--

___
To unsubscribe, edit your list preferences, or view the list archives, please 
visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux





--
Gamefuel
[EMAIL PROTECTED]
http://www.gamefuel.nl



___
To unsubscribe, edit your list preferences, or view the list archives, please 
visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux

[hlds_linux] renting a box

[Jean-Philippe Caron]

--
[ Picked text/plain from multipart/alternative ]
I'm about to rent a box in chicago (Equinix)

amd athlon64 3500+
3 gig ram
linux centos

how many counter-strike server will I be able to run on it
im talking about 12slot private

or my box specs wont do anything good for game hosting
--

___
To unsubscribe, edit your list preferences, or view the list archives, please 
visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux

RE: [hlds_linux] High CPU % Choke on Source

[Scott Pettit]

Your max/min rate is WAY too low. On my Source servers we move around
20-23kB/s PER CLIENT - your settings limit that to 9kB/s per client max
so they'll be choking to death.

Try:

sv_minrate  (I prefer this, but if you have 56k players set
sv_minrate 5000)
sv_maxrate 3
fps_max 500
sv_minupdaterate 30
sv_maxupdate rate should match your tick + 1, so for 66 tick use
sv_maxupdaterate 67.

tick 66 will give better hit registration but 33 will use less CPU. I'd
recommend tick 66 and the rates I gave above.

As a general rule - I have _never_ run more than 2 servers per CPU.
Right now I'm running 2x 18 players per AMD64 3200. So if you're running
a dual Xeon I wouldn't go for more than 2x servers per cpu (if they're
powerful enough).

Looking at my MRTG graphs, our CPU load hasn't changed on the source
servers in the last 4 months so I don't think it's an update related
problem you're seeing. AMD does tend to perform better in my experience
with game servers.

Try the settings above and see how you get on.

On Mon, 2006-01-30 at 18:23 +1000, Joel Dickson wrote:
> Tickrate 66
> sv_minrate 5000
> sv_maxrate 
> fps_max 750
> sv_maxupdaterate 120
> sv_minupdaterate 20
>
> Our 10 plyr server are same but set to maxrate 30k
>
> We are running Zblock Max and mani admin mod.
>
> We don't have choke issue, but we have CPU load issues.
>
> What do think?


___
To unsubscribe, edit your list preferences, or view the list archives, please 
visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux

Re: [hlds_linux] client dropped by server

[kama]

Ok, let me clarify.

What can cause it in a router the servers are attached too? Especially in
the case of use of Cisco routers in the 7600-series.

/Bjorn

On Sun, 29 Jan 2006, Kevin Ottalini wrote:

> Each router and router firmware version has a unique way of mangling UDP
> packets, apparently UDP is not as rigorously tested as TCP.
>
> The exact cause is unknown but the proof is always the same, if you bypass
> the router the problem goes away.
>
> In fact, there are dozens of different error messages that can be generated,
> not just CL_ParseServerMessage, but the cause is always the router, here are
> a few of the most typical messages:
>
> Bad Parse in Client Command
> Bad Command Character in Client Mod
> Bad Command Character in Client Command
> CMD_MAXBACKUP HIT
> Host_Error: CL_ParseServerMessage: Illegible server message - svc_bad
> Host_Error: CL_Parse_Version: Server is protocol -# instead of 47
> Host_Error: DispatchUserMsg: User Msg CZCareer/130 sent too much data (195
> bytes), 192 bytes max
> svc_bad Wrote erroneous message to buffer.dat Host_Error: UserMsg: Not
> Present on Client ###
> Memory Could Not Be Read Error (one cause of this)
> Host error: Cl_Entitynum:  is an invalid number, Cl_max edicts is 
> Host error: Bad server message
>
> and the most typical symptom is:
> Whenever I connect to a (Counter-Strike) server, I get booted after 30
> seconds
>
> I have a FAQ about this in the Steam Support database:
> http://support.steampowered.com/cgi-bin/steampowered.cfg/php/enduser/std_adp.php?p_faqid=19
>
>
> - Original Message -
> From: "kama"
> To: 
> Sent: Sunday, January 29, 2006 3:34 PM
> Subject: Re: [hlds_linux] client dropped by server
>
>
> >
> > Ok, if so... What are the reasons that causes these to happens then?
> >
> > /Bjorn
> >
> > On Sun, 29 Jan 2006, Kevin Ottalini wrote:
> >
> >> CL_ParseServerMessage is very often caused by routers.
> >>
> >> - Original Message -
> >> From: "kama"
> >> To: 
> >> Sent: Sunday, January 29, 2006 3:14 PM
> >> Subject: Re: [hlds_linux] client dropped by server
> >>
> >>
> >> >
> >> > Hmm.. Those two disconnects where probably not the same problem. I have
> >> > reports from serveral admins and regular that they currently are being
> >> > dropped. I got kicked myself once and this was found in my client
> >> > console.
> >> >
> >> > Error: server failed to transmit file 'AY&SY-B^K'
> >> > Last 32 messages parsed.
> >> > 4039 0098 ResetHUD
> >> > 4039 0099 Money
> >> > 4039 0105 StatusIcon
> >> > 4039 0111 ScoreAttrib
> >> > 4039 0114 RoundTime
> >> > 4039 0117 TeamScore
> >> > 4039 0124 TeamScore
> >> > 4039 0138 HideWeapon
> >> > 4039 0140 SetFOV
> >> > 4039 0142 Health
> >> > 4039 0144 Battery
> >> > 4039 0147 Damage
> >> > 4039 0160 FlashBat
> >> > 4039 0162 Train
> >> > 4039 0164 CurWeapon
> >> > 4039 0168 WeapPickup
> >> > 4039 0170 WeapPickup
> >> > 4039 0172 AmmoPickup
> >> > 4039 0175 ScoreAttrib
> >> > 4039 0178 ScreenFade
> >> > 4039 0189 RoundTime
> >> > 4039 0192 ScoreAttrib
> >> > 4039 0195 ScoreAttrib
> >> > 4039 0198 svc_event_reliable
> >> > 4039 0201 AmmoPickup
> >> > 4039 0204 WeapPickup
> >> > 4039 0206 StatusIcon
> >> > 4039 0215 ScoreAttrib
> >> > 4039 0218 CurWeapon
> >> > 4039 0222 CurWeapon
> >> > 4039 0226 svc_filetxferfailed
> >> > BAD:  237:svc_bad
> >> > Host_Error: CL_ParseServerMessage:
> >> >
> >> > /Bjorn
> >> >
>
>
> ___
> To unsubscribe, edit your list preferences, or view the list archives, please 
> visit:
> http://list.valvesoftware.com/mailman/listinfo/hlds_linux
>

___
To unsubscribe, edit your list preferences, or view the list archives, please 
visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux

Re: [hlds_linux] High CPU % Choke on Source

[Joseph Laws]

That's my definition of it now too (which is why I won't host it).  It
didn't use to be that way...we were able to host 100 tickrate quite well
on our AMD platform without any problems.  Then VAC2 came out with it's
multiple threads and lopping call routines



Erik Hollensbe wrote:



Maybe I'm wrong here, but I seem to remember that 100 Tick means "use
as much cpu as possible"?

On Jan 30, 2006, at 12:31 AM, Joseph Laws wrote:


My company has had the same problems with SRCDS.  So bad, in fact,
that
we've discontinued the game entirely.  We don't put more then 5-7
clients to a dual 244 machine with 2 gigs of ram...we aren't able to
even host 1 css client on it without TOP breaking 2 and cpu usage
skyrocketing into the 65% range on the SRCDS server even w/ only 10
players (100 tickrate, fps_max 350 and no mods of any kind).  We
aren't
able to run 100 Tickrate SRCDS with 1.6 (32bit or 64bit), nor the
Quake
engined games without SRCDS finding a reason to eat up any and all cpu
resources available to it.  We do not see choke on our servers...but
perhaps that is because we are not willing to deal w/ the high cpu
usage
of even the smallest SRCDS servers so they don't stay on our boxes
long
enough to develop the symptoms.

Joel Dickson wrote:


Yes, I have similar issues. Not as bad though.

We have same spec system. But 1Gb RAM, RHE 4

We run a 32 plyr server on it and a couple of 10 player servers,
all 66
tic though. CPU load averages 80% with just the 32 plyr, add 2 10
plyr
servers and we have >100% (load avg is 1.5), even idle servers chew a
descent amount of CPU, we have to stop the 10 plyr servers when
not in
use.

But yes, the CPU load is ridiculous compared to what it was last year
when we first setup the box.

Think I herd from someone that disabling VAC solves the problem,
but who
wants to do that.

Regards,

Joel

-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of flags
Sent: Monday, 30 January 2006 5:08 PM
To: hlds_linux@list.valvesoftware.com
Subject: [hlds_linux] High CPU % Choke on Source

This is a multi-part message in MIME format.
--
[ Picked text/plain from multipart/alternative ]
Since the December 5th time frame,  we have had this intermittent
choke.
Normally Choke is related to network issues.
We have done extensive testing and know the network is not the case.
Currently we are running on RedHatEnterprise 4 on dual 2.8 Zeon's,
with
2
Gig ram.  The choke jumps from 0-65.  We have tested in multiple
datacenters, and we have the same problem.  We have even loaded
Windows
2003
and we still get some choke, but the choke is only high when you
first
join.
It may jump to 10-15 but not as bad as on Linux.

The high CPU also showed about in this same time frame.  We had
Enterprise 3
and had high CPU, 12-14 man server would hit 100%  So we then built a
custom
kernel and got CPU % down.  Next we loaded Enterprise 4 a major
release
of
Linux and again CPU % is back up.  A single 20 man server will hit
99%
on a
dual 2.8 Zeon.  Now what is annoying is prior to the December 5th
time
frame
we could easily run a 32 with in the 80% range and zero choke.


Does anyone have a way around these problems?

Does Steam plan on supporting major releases of Linux? Or does steam
plan on
hoping more 14 year olds will host servers at the homes on Windows?



--



___
To unsubscribe, edit your list preferences, or view the list
archives,
please visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux



___
To unsubscribe, edit your list preferences, or view the list
archives, please visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux








___
To unsubscribe, edit your list preferences, or view the list
archives, please visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux



--
Erik Hollensbe
[EMAIL PROTECTED]


___
To unsubscribe, edit your list preferences, or view the list archives,
please visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux






___
To unsubscribe, edit your list preferences, or view the list archives, please 
visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux

Re: [hlds_linux] High CPU % Choke on Source

[Erik Hollensbe]

Maybe I'm wrong here, but I seem to remember that 100 Tick means "use
as much cpu as possible"?

On Jan 30, 2006, at 12:31 AM, Joseph Laws wrote:


My company has had the same problems with SRCDS.  So bad, in fact,
that
we've discontinued the game entirely.  We don't put more then 5-7
clients to a dual 244 machine with 2 gigs of ram...we aren't able to
even host 1 css client on it without TOP breaking 2 and cpu usage
skyrocketing into the 65% range on the SRCDS server even w/ only 10
players (100 tickrate, fps_max 350 and no mods of any kind).  We
aren't
able to run 100 Tickrate SRCDS with 1.6 (32bit or 64bit), nor the
Quake
engined games without SRCDS finding a reason to eat up any and all cpu
resources available to it.  We do not see choke on our servers...but
perhaps that is because we are not willing to deal w/ the high cpu
usage
of even the smallest SRCDS servers so they don't stay on our boxes
long
enough to develop the symptoms.

Joel Dickson wrote:


Yes, I have similar issues. Not as bad though.

We have same spec system. But 1Gb RAM, RHE 4

We run a 32 plyr server on it and a couple of 10 player servers,
all 66
tic though. CPU load averages 80% with just the 32 plyr, add 2 10
plyr
servers and we have >100% (load avg is 1.5), even idle servers chew a
descent amount of CPU, we have to stop the 10 plyr servers when
not in
use.

But yes, the CPU load is ridiculous compared to what it was last year
when we first setup the box.

Think I herd from someone that disabling VAC solves the problem,
but who
wants to do that.

Regards,

Joel

-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of flags
Sent: Monday, 30 January 2006 5:08 PM
To: hlds_linux@list.valvesoftware.com
Subject: [hlds_linux] High CPU % Choke on Source

This is a multi-part message in MIME format.
--
[ Picked text/plain from multipart/alternative ]
Since the December 5th time frame,  we have had this intermittent
choke.
Normally Choke is related to network issues.
We have done extensive testing and know the network is not the case.
Currently we are running on RedHatEnterprise 4 on dual 2.8 Zeon's,
with
2
Gig ram.  The choke jumps from 0-65.  We have tested in multiple
datacenters, and we have the same problem.  We have even loaded
Windows
2003
and we still get some choke, but the choke is only high when you
first
join.
It may jump to 10-15 but not as bad as on Linux.

The high CPU also showed about in this same time frame.  We had
Enterprise 3
and had high CPU, 12-14 man server would hit 100%  So we then built a
custom
kernel and got CPU % down.  Next we loaded Enterprise 4 a major
release
of
Linux and again CPU % is back up.  A single 20 man server will hit
99%
on a
dual 2.8 Zeon.  Now what is annoying is prior to the December 5th
time
frame
we could easily run a 32 with in the 80% range and zero choke.


Does anyone have a way around these problems?

Does Steam plan on supporting major releases of Linux? Or does steam
plan on
hoping more 14 year olds will host servers at the homes on Windows?



--



___
To unsubscribe, edit your list preferences, or view the list
archives,
please visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux



___
To unsubscribe, edit your list preferences, or view the list
archives, please visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux








___
To unsubscribe, edit your list preferences, or view the list
archives, please visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux


--
Erik Hollensbe
[EMAIL PROTECTED]


___
To unsubscribe, edit your list preferences, or view the list archives, please 
visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux

Re: [hlds_linux] High CPU % Choke on Source

[Joseph Laws]

My company has had the same problems with SRCDS.  So bad, in fact, that
we've discontinued the game entirely.  We don't put more then 5-7
clients to a dual 244 machine with 2 gigs of ram...we aren't able to
even host 1 css client on it without TOP breaking 2 and cpu usage
skyrocketing into the 65% range on the SRCDS server even w/ only 10
players (100 tickrate, fps_max 350 and no mods of any kind).  We aren't
able to run 100 Tickrate SRCDS with 1.6 (32bit or 64bit), nor the Quake
engined games without SRCDS finding a reason to eat up any and all cpu
resources available to it.  We do not see choke on our servers...but
perhaps that is because we are not willing to deal w/ the high cpu usage
of even the smallest SRCDS servers so they don't stay on our boxes long
enough to develop the symptoms.

Joel Dickson wrote:


Yes, I have similar issues. Not as bad though.

We have same spec system. But 1Gb RAM, RHE 4

We run a 32 plyr server on it and a couple of 10 player servers, all 66
tic though. CPU load averages 80% with just the 32 plyr, add 2 10 plyr
servers and we have >100% (load avg is 1.5), even idle servers chew a
descent amount of CPU, we have to stop the 10 plyr servers when not in
use.

But yes, the CPU load is ridiculous compared to what it was last year
when we first setup the box.

Think I herd from someone that disabling VAC solves the problem, but who
wants to do that.

Regards,

Joel

-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of flags
Sent: Monday, 30 January 2006 5:08 PM
To: hlds_linux@list.valvesoftware.com
Subject: [hlds_linux] High CPU % Choke on Source

This is a multi-part message in MIME format.
--
[ Picked text/plain from multipart/alternative ]
Since the December 5th time frame,  we have had this intermittent choke.
Normally Choke is related to network issues.
We have done extensive testing and know the network is not the case.
Currently we are running on RedHatEnterprise 4 on dual 2.8 Zeon's, with
2
Gig ram.  The choke jumps from 0-65.  We have tested in multiple
datacenters, and we have the same problem.  We have even loaded Windows
2003
and we still get some choke, but the choke is only high when you first
join.
It may jump to 10-15 but not as bad as on Linux.

The high CPU also showed about in this same time frame.  We had
Enterprise 3
and had high CPU, 12-14 man server would hit 100%  So we then built a
custom
kernel and got CPU % down.  Next we loaded Enterprise 4 a major release
of
Linux and again CPU % is back up.  A single 20 man server will hit 99%
on a
dual 2.8 Zeon.  Now what is annoying is prior to the December 5th time
frame
we could easily run a 32 with in the 80% range and zero choke.


Does anyone have a way around these problems?

Does Steam plan on supporting major releases of Linux? Or does steam
plan on
hoping more 14 year olds will host servers at the homes on Windows?



--



___
To unsubscribe, edit your list preferences, or view the list archives,
please visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux



___
To unsubscribe, edit your list preferences, or view the list archives, please 
visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux








___
To unsubscribe, edit your list preferences, or view the list archives, please 
visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux

RE: [hlds_linux] High CPU % Choke on Source

[Joel Dickson]

Tickrate 66
sv_minrate 5000
sv_maxrate 
fps_max 750
sv_maxupdaterate 120
sv_minupdaterate 20

Our 10 plyr server are same but set to maxrate 30k

We are running Zblock Max and mani admin mod.

We don't have choke issue, but we have CPU load issues.

What do think?

Regards,

Joel



-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Scott
Pettit
Sent: Monday, 30 January 2006 5:50 PM
To: hlds_linux@list.valvesoftware.com
Subject: Re: [hlds_linux] High CPU % Choke on Source

I'm having no problems with choke.

Could you please tell us your settings for:

sv_minrate
sv_maxrate
sv_minupdaterate
sv_maxupdaterate
tickrate

And knowing what kind of connection your average client is using as well
- and what cl_cmdrate / cl_updaterate / rate they are using would also
be handy.

As a general rule - choke is when the client is flooded - and a fix is
usually to increase rate and decrease cl_cmd/update rate.

CPU - I haven't noticed higher CPU usage, but then I don't overload my
machines either which a lot of people on this list seem to do; running
ten billion 32 slot servers per machine.

How many servers are you running per machine, and how many slots do each
have? Are you running mods?

Please try and give specific info if possible when posting.

-Scott

On Mon, 2006-01-30 at 01:08 -0600, flags wrote:
> Since the December 5th time frame,  we have had this intermittent
choke.
> Stuff about high cpu usage.


___
To unsubscribe, edit your list preferences, or view the list archives,
please visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux



___
To unsubscribe, edit your list preferences, or view the list archives, please 
visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux