RE: [hlds] RE: Servers losing VAC, gettin beyond a joke...
Top stuff guys, thanks a lot! - Andrew -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Alfred Reynolds Sent: Saturday, 21 January 2006 11:04 AM To: hlds@list.valvesoftware.com Subject: RE: [hlds] RE: Servers losing VAC, gettin beyond a joke... Thanks for everyone that responded to my request, the information that I got allowed me to track down the problem. There is a bug on startup that can cause a server to go insecure that we are working on a fix for (if the server fails to connect to the VAC2 backend on its first attempt then the server is wrongly forced into insecure mode even if it can later connect). - Alfred Ian mu wrote: > -- > [ Picked text/plain from multipart/alternative ] > Too true! I doubt there's be much to see anyway apart from the same > message > when it can't connect on startup though, at least thats "more" > replicable as > I seem to get that quite frequent. > > Like others, happy to help with any tracing/debugging if needed. > > > On 1/20/06, Whisper <[EMAIL PROTECTED]> wrote: >> >> -- >> [ Picked text/plain from multipart/alternative ] >> A watched VAC enabled server never goes insecure :( >> >> On 1/20/06, Ian mu <[EMAIL PROTECTED]> wrote: >>> >>> -- >>> [ Picked text/plain from multipart/alternative ] >>> Have this mainly on windows (pretty sure we've had it on linux as >>> well, just less css servers on it). >>> >>> Notice it more on a restart, server starts up, goes secure for a >>> few secs, then message unable to contact (steam/valve/auth servers >>> whatever it is), then goes secure 0. Normally on another restart >>> its the same for me, but after a few variations in time on restart >>> it ends up fine. It's different servers each time and different >>> boxes, so no specifics and I doubt config etc. >>> >>> Maybe there's just too much traffic hitting certain boxes at >>> certain times and it times out or something? If there's a pool, can >>> it try one server, if no response try a different one or something >>> ? Thats the way it feels anyway. >>> >>> >>> On Fri, 20 Jan 2006 00:55:47, Dan Sorenson <[EMAIL PROTECTED]> wrote: >>>> >>>> At 09:06 PM 1/19/2006 -0500, Ray wrote: >>>> >>>>> Exactly my point, nothing is in the logs and the server isn't >>>>> crashed so their's no dump. The bottom line is it's impossible >>>>> for us to help you fix what's broken. >>>> >>>>Oh, it's not impossible. It's just difficult. There's >>>> probably a command-line flag you can set to put the server into >>>> dev mode and increase the debug level. Who knows? Perhaps there >>>> being nothing in the logs is the clue that's needed -- something >>>> that should be logged not being there is just as important as >>>> something being logged with an error. >>>> >>>>The other thought is that server startup only takes a few >>>> seconds, maybe a couple of megs worth of traffic to get registered >>>> and grab VAC. Fire up MS's Netmon or maybe Ethereal and do a >>>> packet capture and send that off. Alfred should be able to take >>>> that, filter on his VAC servers, and see exactly what's going out >>>> and coming back. I don't know Valve's network infrastructure, but >>>> chances are it's all a switched network. Send Alfred a private >>>> e-mail with your server IP and perhaps a schedule of when you'll >>>> restart and I bet Alfred will be able to get some logging enabled, >>>> perhaps mirror a server port >>>> and do his own packet capture. >>>> >>>>That's how I'd approach it, anyway. Alfred's a busy guy >>>> and I don't want to say he can or will do this, but it seems to me >>>> that this approach gives him something more solid to work with. >>>> >>>>- Dan >>>> >>>> * Dan Sorenson DoD #1066 A.H.M.C. #35 >>>> [EMAIL PROTECTED] * >>>> * Vikings? There ain't no vikings here. Just us honest farmers. >>>> * >>>> * The town was burning, the villagers were dead. They didn't need >>>> * >>>> * those sheep anyway. That's our story and we're sticking to it. >>>> * >>>> >>>> >>>
RE: [hlds] RE: Servers losing VAC, gettin beyond a joke...
wow screaming and moaning actually worked! :) ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds
Re: [hlds] RE: Servers losing VAC, gettin beyond a joke...
-- [ Picked text/plain from multipart/alternative ] I think constructive screaming and moaning often is ok ;). Will as always thank Alfred, no idea if the fix will work (as there's maybe some people still with mid game drops, but it will at least separate issues if so and help those to be isolated and fixed), but I do sincerely thank Alfred for his input into this forum. At least we get an idea even if something we expect isn't being fixed we get feedback on this so we can give more info so to speak. It's almost like a void when there's a problem and Alfred isn't about to at least give an idea of the standing. On 1/21/06, Napier, Kevin <[EMAIL PROTECTED]> wrote: > > wow screaming and moaning actually worked! :) > > > ___ > To unsubscribe, edit your list preferences, or view the list archives, > please visit: > http://list.valvesoftware.com/mailman/listinfo/hlds > -- ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds
RE: [hlds] RE: Servers losing VAC, gettin beyond a joke...
Thanks for everyone that responded to my request, the information that I got allowed me to track down the problem. There is a bug on startup that can cause a server to go insecure that we are working on a fix for (if the server fails to connect to the VAC2 backend on its first attempt then the server is wrongly forced into insecure mode even if it can later connect). - Alfred Ian mu wrote: > -- > [ Picked text/plain from multipart/alternative ] > Too true! I doubt there's be much to see anyway apart from the same > message > when it can't connect on startup though, at least thats "more" > replicable as > I seem to get that quite frequent. > > Like others, happy to help with any tracing/debugging if needed. > > > On 1/20/06, Whisper <[EMAIL PROTECTED]> wrote: >> >> -- >> [ Picked text/plain from multipart/alternative ] >> A watched VAC enabled server never goes insecure :( >> >> On 1/20/06, Ian mu <[EMAIL PROTECTED]> wrote: >>> >>> -- >>> [ Picked text/plain from multipart/alternative ] >>> Have this mainly on windows (pretty sure we've had it on linux as >>> well, just less css servers on it). >>> >>> Notice it more on a restart, server starts up, goes secure for a >>> few secs, then message unable to contact (steam/valve/auth servers >>> whatever it is), then goes secure 0. Normally on another restart >>> its the same for me, but after a few variations in time on restart >>> it ends up fine. It's different servers each time and different >>> boxes, so no specifics and I doubt config etc. >>> >>> Maybe there's just too much traffic hitting certain boxes at >>> certain times and it times out or something? If there's a pool, can >>> it try one server, if no response try a different one or something >>> ? Thats the way it feels anyway. >>> >>> >>> On Fri, 20 Jan 2006 00:55:47, Dan Sorenson <[EMAIL PROTECTED]> wrote: At 09:06 PM 1/19/2006 -0500, Ray wrote: > Exactly my point, nothing is in the logs and the server isn't > crashed so their's no dump. The bottom line is it's impossible > for us to help you fix what's broken. Oh, it's not impossible. It's just difficult. There's probably a command-line flag you can set to put the server into dev mode and increase the debug level. Who knows? Perhaps there being nothing in the logs is the clue that's needed -- something that should be logged not being there is just as important as something being logged with an error. The other thought is that server startup only takes a few seconds, maybe a couple of megs worth of traffic to get registered and grab VAC. Fire up MS's Netmon or maybe Ethereal and do a packet capture and send that off. Alfred should be able to take that, filter on his VAC servers, and see exactly what's going out and coming back. I don't know Valve's network infrastructure, but chances are it's all a switched network. Send Alfred a private e-mail with your server IP and perhaps a schedule of when you'll restart and I bet Alfred will be able to get some logging enabled, perhaps mirror a server port and do his own packet capture. That's how I'd approach it, anyway. Alfred's a busy guy and I don't want to say he can or will do this, but it seems to me that this approach gives him something more solid to work with. - Dan * Dan Sorenson DoD #1066 A.H.M.C. #35 [EMAIL PROTECTED] * * Vikings? There ain't no vikings here. Just us honest farmers. * * The town was burning, the villagers were dead. They didn't need * * those sheep anyway. That's our story and we're sticking to it. * ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds >>> -- >>> >>> ___ >>> To unsubscribe, edit your list preferences, or view the list >>> archives, please visit: >>> http://list.valvesoftware.com/mailman/listinfo/hlds >>> >> -- >> >> ___ >> To unsubscribe, edit your list preferences, or view the list >> archives, please visit: >> http://list.valvesoftware.com/mailman/listinfo/hlds >> > -- > > ___ > To unsubscribe, edit your list preferences, or view the list > archives, please visit: > http://list.valvesoftware.com/mailman/listinfo/hlds ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds
Re: [hlds] RE: Servers losing VAC, gettin beyond a joke...
OK, I wrote a crappy little compilebasic program that will check every 10 seconds and monitor a source server for secure status and log it. If you have a server that is chronic in losing VAC status while still running (Case 2 only) then email me your server IP address and port and I'll start monitoring it. Once I clean up the code a bit I'll release it to anyone who wants it (just don't laugh ok? it works). - Original Message - From: "Whisper" To: Sent: Friday, January 20, 2006 2:56 AM Subject: Re: [hlds] RE: Servers losing VAC, gettin beyond a joke... We get both Kevin We reboot our servers every 24 hours at around 5am, and some of them won't go secure. Later on we will reboot the ones that did not come up as VAC Secure, and they wil come up as VAC Secure. At some random time in the future before the usual 5am restart process, and not necessarily the same servers that did not go VAC Secure during the reboot process, will go VAC INSecure, without a server reboot or crash. I am watching our servers diligently at the moment and they refuse to now go VAC INSecure. If I left them for 6 hours (I don't want to do this) no doubts some will go VAC INSecure while I am not watching. - Original Message - From: "Kevin Ottalini" To: Sent: Friday, January 20, 2006 2:33 AM Subject: Re: [hlds] RE: Servers losing VAC, gettin beyond a joke... There is a tremendous difference between: Case 1: a server being unable to acquire VAC on startup and Case 2: a server "losing VAC" while running. The title of this thread might be misleading which makes it hard to talk about what the real problem is. If you have consistant problems getting your server to acquire VAC on startup then say so. If you have any evidence that your server (without restarting) dynamically becomes insecure, please say so. If your server crashes and is restarted or programmatically restarts (and does not acquire VAC at that time) then you need to identify this as Case 1. ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds
Re: [hlds] RE: Servers losing VAC, gettin beyond a joke...
-- [ Picked text/plain from multipart/alternative ] We get both Kevin We reboot our servers every 24 hours at around 5am, and some of them won't go secure. Later on we will reboot the ones that did not come up as VAC Secure, and they wil come up as VAC Secure. At some random time in the future before the usual 5am restart process, and not necessarily the same servers that did not go VAC Secure during the reboot process, will go VAC INSecure, without a server reboot or crash. I am watching our servers diligently at the moment and they refuse to now go VAC INSecure. If I left them for 6 hours (I don't want to do this) no doubts some will go VAC INSecure while I am not watching. On 1/20/06, Ian mu <[EMAIL PROTECTED]> wrote: > > -- > [ Picked text/plain from multipart/alternative ] > Too true! I doubt there's be much to see anyway apart from the same > message > when it can't connect on startup though, at least thats "more" replicable > as > I seem to get that quite frequent. > > Like others, happy to help with any tracing/debugging if needed. > > > On 1/20/06, Whisper <[EMAIL PROTECTED]> wrote: > > > > -- > > [ Picked text/plain from multipart/alternative ] > > A watched VAC enabled server never goes insecure :( > > > > On 1/20/06, Ian mu <[EMAIL PROTECTED]> wrote: > > > > > > -- > > > [ Picked text/plain from multipart/alternative ] > > > Have this mainly on windows (pretty sure we've had it on linux as > well, > > > just > > > less css servers on it). > > > > > > Notice it more on a restart, server starts up, goes secure for a few > > secs, > > > then message unable to contact (steam/valve/auth servers whatever it > > is), > > > then goes secure 0. Normally on another restart its the same for me, > but > > > after a few variations in time on restart it ends up fine. It's > > different > > > servers each time and different boxes, so no specifics and I doubt > > config > > > etc. > > > > > > Maybe there's just too much traffic hitting certain boxes at certain > > times > > > and it times out or something? If there's a pool, can it try one > server, > > > if > > > no response try a different one or something ? Thats the way it feels > > > anyway. > > > > > > > > > On Fri, 20 Jan 2006 00:55:47, Dan Sorenson <[EMAIL PROTECTED]> wrote: > > > > > > > > At 09:06 PM 1/19/2006 -0500, Ray wrote: > > > > > > > > >Exactly my point, nothing is in the logs and the server isn't > crashed > > > so > > > > >their's no dump. The bottom line is it's impossible for us to help > > you > > > > fix > > > > >what's broken. > > > > > > > >Oh, it's not impossible. It's just difficult. There's > > probably > > > > a command-line flag you can set to put the server into dev mode and > > > > increase the debug level. Who knows? Perhaps there being nothing > in > > > > the logs is the clue that's needed -- something that should be > logged > > > > not being there is just as important as something being logged with > > > > an error. > > > > > > > >The other thought is that server startup only takes a few > > > > seconds, maybe a couple of megs worth of traffic to get registered > > > > and grab VAC. Fire up MS's Netmon or maybe Ethereal and do a packet > > > > capture and send that off. Alfred should be able to take that, > filter > > > > on his VAC servers, and see exactly what's going out and coming > back. > > > > I don't know Valve's network infrastructure, but chances are it's > all > > > > a switched network. Send Alfred a private e-mail with your server > IP > > > > and perhaps a schedule of when you'll restart and I bet Alfred will > > > > be able to get some logging enabled, perhaps mirror a server port > > > > and do his own packet capture. > > > > > > > >That's how I'd approach it, anyway. Alfred's a busy guy > > > > and I don't want to say he can or will do this, but it seems to me > > > > that this approach gives him something more solid to work with. > > > > > > > >- Dan > > > > > > > > * Dan Sorenson DoD #1066 A.H.M.C. #35 [EMAIL PROTECTED] > > > > * Vikings? There ain't no vikings here. Just us honest farmers. > * > > > > * The town was burning, the villagers were dead. They didn't > need * > > > > * those sheep anyway. That's our story and we're sticking to it. > * > > > > > > > > > > > > ___ > > > > To unsubscribe, edit your list preferences, or view the list > archives, > > > > please visit: > > > > http://list.valvesoftware.com/mailman/listinfo/hlds > > > > > > > -- > > > > > > ___ > > > To unsubscribe, edit your list preferences, or view the list archives, > > > please visit: > > > http://list.valvesoftware.com/mailman/listinfo/hlds > > > > > -- > > > > ___ > > To unsubscribe, edit your list preferences, or view the list archives, > > please visit: > > http://list.valvesoftware.com/mailman/listinfo/hlds > > > -- > > _
Re: [hlds] RE: Servers losing VAC, gettin beyond a joke...
-- [ Picked text/plain from multipart/alternative ] Too true! I doubt there's be much to see anyway apart from the same message when it can't connect on startup though, at least thats "more" replicable as I seem to get that quite frequent. Like others, happy to help with any tracing/debugging if needed. On 1/20/06, Whisper <[EMAIL PROTECTED]> wrote: > > -- > [ Picked text/plain from multipart/alternative ] > A watched VAC enabled server never goes insecure :( > > On 1/20/06, Ian mu <[EMAIL PROTECTED]> wrote: > > > > -- > > [ Picked text/plain from multipart/alternative ] > > Have this mainly on windows (pretty sure we've had it on linux as well, > > just > > less css servers on it). > > > > Notice it more on a restart, server starts up, goes secure for a few > secs, > > then message unable to contact (steam/valve/auth servers whatever it > is), > > then goes secure 0. Normally on another restart its the same for me, but > > after a few variations in time on restart it ends up fine. It's > different > > servers each time and different boxes, so no specifics and I doubt > config > > etc. > > > > Maybe there's just too much traffic hitting certain boxes at certain > times > > and it times out or something? If there's a pool, can it try one server, > > if > > no response try a different one or something ? Thats the way it feels > > anyway. > > > > > > On Fri, 20 Jan 2006 00:55:47, Dan Sorenson <[EMAIL PROTECTED]> wrote: > > > > > > At 09:06 PM 1/19/2006 -0500, Ray wrote: > > > > > > >Exactly my point, nothing is in the logs and the server isn't crashed > > so > > > >their's no dump. The bottom line is it's impossible for us to help > you > > > fix > > > >what's broken. > > > > > >Oh, it's not impossible. It's just difficult. There's > probably > > > a command-line flag you can set to put the server into dev mode and > > > increase the debug level. Who knows? Perhaps there being nothing in > > > the logs is the clue that's needed -- something that should be logged > > > not being there is just as important as something being logged with > > > an error. > > > > > >The other thought is that server startup only takes a few > > > seconds, maybe a couple of megs worth of traffic to get registered > > > and grab VAC. Fire up MS's Netmon or maybe Ethereal and do a packet > > > capture and send that off. Alfred should be able to take that, filter > > > on his VAC servers, and see exactly what's going out and coming back. > > > I don't know Valve's network infrastructure, but chances are it's all > > > a switched network. Send Alfred a private e-mail with your server IP > > > and perhaps a schedule of when you'll restart and I bet Alfred will > > > be able to get some logging enabled, perhaps mirror a server port > > > and do his own packet capture. > > > > > >That's how I'd approach it, anyway. Alfred's a busy guy > > > and I don't want to say he can or will do this, but it seems to me > > > that this approach gives him something more solid to work with. > > > > > >- Dan > > > > > > * Dan Sorenson DoD #1066 A.H.M.C. #35 [EMAIL PROTECTED] * > > > * Vikings? There ain't no vikings here. Just us honest farmers. * > > > * The town was burning, the villagers were dead. They didn't need * > > > * those sheep anyway. That's our story and we're sticking to it. * > > > > > > > > > ___ > > > To unsubscribe, edit your list preferences, or view the list archives, > > > please visit: > > > http://list.valvesoftware.com/mailman/listinfo/hlds > > > > > -- > > > > ___ > > To unsubscribe, edit your list preferences, or view the list archives, > > please visit: > > http://list.valvesoftware.com/mailman/listinfo/hlds > > > -- > > ___ > To unsubscribe, edit your list preferences, or view the list archives, > please visit: > http://list.valvesoftware.com/mailman/listinfo/hlds > -- ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds
Re: [hlds] RE: Servers losing VAC, gettin beyond a joke...
There is a tremendous difference between: Case 1: a server being unable to acquire VAC on startup and Case 2: a server "losing VAC" while running. The title of this thread might be misleading which makes it hard to talk about what the real problem is. If you have consistant problems getting your server to acquire VAC on startup then say so. If you have any evidence that your server (without restarting) dynamically becomes insecure, please say so. If your server crashes and is restarted or programmatically restarts (and does not acquire VAC at that time) then you need to identify this as Case 1. - Original Message - From: "Whisper" To: Sent: Friday, January 20, 2006 2:18 AM Subject: Re: [hlds] RE: Servers losing VAC, gettin beyond a joke... A watched VAC enabled server never goes insecure :( On 1/20/06, Ian mu wrote: Have this mainly on windows (pretty sure we've had it on linux as well, just less css servers on it). Notice it more on a restart, server starts up, goes secure for a few secs, then message unable to contact (steam/valve/auth servers whatever it is), then goes secure 0. Normally on another restart its the same for me, but after a few variations in time on restart it ends up fine. It's different servers each time and different boxes, so no specifics and I doubt config etc. Maybe there's just too much traffic hitting certain boxes at certain times and it times out or something? If there's a pool, can it try one server, if no response try a different one or something ? Thats the way it feels anyway. On Fri, 20 Jan 2006 00:55:47, Dan Sorenson wrote: > At 09:06 PM 1/19/2006 -0500, Ray wrote: > >Exactly my point, nothing is in the logs and the server isn't crashed so > >their's no dump. The bottom line is it's impossible for us to help you > fix > >what's broken. > >Oh, it's not impossible. It's just difficult. There's probably > a command-line flag you can set to put the server into dev mode and > increase the debug level. Who knows? Perhaps there being nothing in > the logs is the clue that's needed -- something that should be logged > not being there is just as important as something being logged with > an error. > >The other thought is that server startup only takes a few > seconds, maybe a couple of megs worth of traffic to get registered > and grab VAC. Fire up MS's Netmon or maybe Ethereal and do a packet > capture and send that off. Alfred should be able to take that, filter > on his VAC servers, and see exactly what's going out and coming back. > I don't know Valve's network infrastructure, but chances are it's all > a switched network. Send Alfred a private e-mail with your server IP > and perhaps a schedule of when you'll restart and I bet Alfred will > be able to get some logging enabled, perhaps mirror a server port > and do his own packet capture. > >That's how I'd approach it, anyway. Alfred's a busy guy > and I don't want to say he can or will do this, but it seems to me > that this approach gives him something more solid to work with. > >- Dan > > * Dan Sorenson DoD #1066 A.H.M.C. #35 ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds
Re: [hlds] RE: Servers losing VAC, gettin beyond a joke...
-- [ Picked text/plain from multipart/alternative ] A watched VAC enabled server never goes insecure :( On 1/20/06, Ian mu <[EMAIL PROTECTED]> wrote: > > -- > [ Picked text/plain from multipart/alternative ] > Have this mainly on windows (pretty sure we've had it on linux as well, > just > less css servers on it). > > Notice it more on a restart, server starts up, goes secure for a few secs, > then message unable to contact (steam/valve/auth servers whatever it is), > then goes secure 0. Normally on another restart its the same for me, but > after a few variations in time on restart it ends up fine. It's different > servers each time and different boxes, so no specifics and I doubt config > etc. > > Maybe there's just too much traffic hitting certain boxes at certain times > and it times out or something? If there's a pool, can it try one server, > if > no response try a different one or something ? Thats the way it feels > anyway. > > > On Fri, 20 Jan 2006 00:55:47, Dan Sorenson <[EMAIL PROTECTED]> wrote: > > > > At 09:06 PM 1/19/2006 -0500, Ray wrote: > > > > >Exactly my point, nothing is in the logs and the server isn't crashed > so > > >their's no dump. The bottom line is it's impossible for us to help you > > fix > > >what's broken. > > > >Oh, it's not impossible. It's just difficult. There's probably > > a command-line flag you can set to put the server into dev mode and > > increase the debug level. Who knows? Perhaps there being nothing in > > the logs is the clue that's needed -- something that should be logged > > not being there is just as important as something being logged with > > an error. > > > >The other thought is that server startup only takes a few > > seconds, maybe a couple of megs worth of traffic to get registered > > and grab VAC. Fire up MS's Netmon or maybe Ethereal and do a packet > > capture and send that off. Alfred should be able to take that, filter > > on his VAC servers, and see exactly what's going out and coming back. > > I don't know Valve's network infrastructure, but chances are it's all > > a switched network. Send Alfred a private e-mail with your server IP > > and perhaps a schedule of when you'll restart and I bet Alfred will > > be able to get some logging enabled, perhaps mirror a server port > > and do his own packet capture. > > > >That's how I'd approach it, anyway. Alfred's a busy guy > > and I don't want to say he can or will do this, but it seems to me > > that this approach gives him something more solid to work with. > > > >- Dan > > > > * Dan Sorenson DoD #1066 A.H.M.C. #35 [EMAIL PROTECTED] * > > * Vikings? There ain't no vikings here. Just us honest farmers. * > > * The town was burning, the villagers were dead. They didn't need * > > * those sheep anyway. That's our story and we're sticking to it. * > > > > > > ___ > > To unsubscribe, edit your list preferences, or view the list archives, > > please visit: > > http://list.valvesoftware.com/mailman/listinfo/hlds > > > -- > > ___ > To unsubscribe, edit your list preferences, or view the list archives, > please visit: > http://list.valvesoftware.com/mailman/listinfo/hlds > -- ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds
Re: [hlds] RE: Servers losing VAC, gettin beyond a joke...
-- [ Picked text/plain from multipart/alternative ] Have this mainly on windows (pretty sure we've had it on linux as well, just less css servers on it). Notice it more on a restart, server starts up, goes secure for a few secs, then message unable to contact (steam/valve/auth servers whatever it is), then goes secure 0. Normally on another restart its the same for me, but after a few variations in time on restart it ends up fine. It's different servers each time and different boxes, so no specifics and I doubt config etc. Maybe there's just too much traffic hitting certain boxes at certain times and it times out or something? If there's a pool, can it try one server, if no response try a different one or something ? Thats the way it feels anyway. On Fri, 20 Jan 2006 00:55:47, Dan Sorenson <[EMAIL PROTECTED]> wrote: > > At 09:06 PM 1/19/2006 -0500, Ray wrote: > > >Exactly my point, nothing is in the logs and the server isn't crashed so > >their's no dump. The bottom line is it's impossible for us to help you > fix > >what's broken. > >Oh, it's not impossible. It's just difficult. There's probably > a command-line flag you can set to put the server into dev mode and > increase the debug level. Who knows? Perhaps there being nothing in > the logs is the clue that's needed -- something that should be logged > not being there is just as important as something being logged with > an error. > >The other thought is that server startup only takes a few > seconds, maybe a couple of megs worth of traffic to get registered > and grab VAC. Fire up MS's Netmon or maybe Ethereal and do a packet > capture and send that off. Alfred should be able to take that, filter > on his VAC servers, and see exactly what's going out and coming back. > I don't know Valve's network infrastructure, but chances are it's all > a switched network. Send Alfred a private e-mail with your server IP > and perhaps a schedule of when you'll restart and I bet Alfred will > be able to get some logging enabled, perhaps mirror a server port > and do his own packet capture. > >That's how I'd approach it, anyway. Alfred's a busy guy > and I don't want to say he can or will do this, but it seems to me > that this approach gives him something more solid to work with. > >- Dan > > * Dan Sorenson DoD #1066 A.H.M.C. #35 [EMAIL PROTECTED] * > * Vikings? There ain't no vikings here. Just us honest farmers. * > * The town was burning, the villagers were dead. They didn't need * > * those sheep anyway. That's our story and we're sticking to it. * > > > ___ > To unsubscribe, edit your list preferences, or view the list archives, > please visit: > http://list.valvesoftware.com/mailman/listinfo/hlds > -- ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds
[hlds] RE: Servers losing VAC, gettin beyond a joke...
At 09:06 PM 1/19/2006 -0500, Ray wrote: >Exactly my point, nothing is in the logs and the server isn't crashed so >their's no dump. The bottom line is it's impossible for us to help you fix >what's broken. Oh, it's not impossible. It's just difficult. There's probably a command-line flag you can set to put the server into dev mode and increase the debug level. Who knows? Perhaps there being nothing in the logs is the clue that's needed -- something that should be logged not being there is just as important as something being logged with an error. The other thought is that server startup only takes a few seconds, maybe a couple of megs worth of traffic to get registered and grab VAC. Fire up MS's Netmon or maybe Ethereal and do a packet capture and send that off. Alfred should be able to take that, filter on his VAC servers, and see exactly what's going out and coming back. I don't know Valve's network infrastructure, but chances are it's all a switched network. Send Alfred a private e-mail with your server IP and perhaps a schedule of when you'll restart and I bet Alfred will be able to get some logging enabled, perhaps mirror a server port and do his own packet capture. That's how I'd approach it, anyway. Alfred's a busy guy and I don't want to say he can or will do this, but it seems to me that this approach gives him something more solid to work with. - Dan * Dan Sorenson DoD #1066 A.H.M.C. #35 [EMAIL PROTECTED] * * Vikings? There ain't no vikings here. Just us honest farmers. * * The town was burning, the villagers were dead. They didn't need * * those sheep anyway. That's our story and we're sticking to it. * ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds
Re: [hlds] Re: Servers losing VAC, gettin beyond a joke...
At 10:30 AM 1/17/2006 -0600, you wrote: >The problem is that it doesnt put anything into the event log, or logs. >It just quits, creates a mdmp and thats it. Now unless you are a >programmer (and I am not) it is not exactly easy to figure out the mdmp >files. Agreed. The tools with the Windows server are a bit more primitive, or perhaps just not as added to as the Linux tools tend to be, but working in both worlds I find them a lot easier to use for most tasks. Making those memory dumps available to the developers via a webserver or something might be useful though. >So when valve >helps us to help them, maybe we can include more things =/ It is definately a 2-way street. We are one really large quality control center for Valve, using their products in ways they cannot begin to duplicate in a lab. One would think the better the feedback and information we can give them the better their product is going to become. - Dan * Dan Sorenson DoD #1066 A.H.M.C. #35 [EMAIL PROTECTED] * * Vikings? There ain't no vikings here. Just us honest farmers. * * The town was burning, the villagers were dead. They didn't need * * those sheep anyway. That's our story and we're sticking to it. * ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds
RE: [hlds] Re: Servers losing VAC, gettin beyond a joke...
You could also post the mdump and your relevant log file (if you wanna xx the ips and steams go to.) -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Whisper Sent: Tuesday, January 17, 2006 10:36 AM To: hlds@list.valvesoftware.com Subject: Re: [hlds] Re: Servers losing VAC, gettin beyond a joke... -- [ Picked text/plain from multipart/alternative ] You would send the mdmp's to Valve That is what they are there for. On 1/18/06, Hell Phoenix <[EMAIL PROTECTED]> wrote: > > The problem is that it doesnt put anything into the event log, or logs. > It just quits, creates a mdmp and thats it. Now unless you are a > programmer (and I am not) it is not exactly easy to figure out the mdmp > files. Trust me, I have tried. I am MCSA certified (on my way to mcse) > so I am not exactly dumbfounded when it comes to Windows. So when valve > helps us to help them, maybe we can include more things =/ > > Dan Sorenson wrote: > > At 02:15 PM 1/16/2006 +0100, Saint K. wrote: > > > >> Currently we have to reset our servers multiple times a day because > they are > >> losing VAC. [...] > >> We need VAC fixed asap as this is getting out of controle!! > >> > > > > I run both Linux and Windows servers, and hence subscribe > > to both lists. In the Linux list you'll see posted logfile exerpts, > > error messages, core dumps, stuff the dev team can use to fix a > > problem. You'll also see troubleshooting steps taken, various > > library versions that Source depends upon, etc... > > > > In the windows list you see "Fix this." > > > > I suspect the problem has been so long in the offering > > because Valve can't reliably reproduce it. Help them. They're > > a small team, why not put the event logs and any Dr. Watson's and > > server logfiles on a webserver and help them out by providing the > > information they need to start working the problem? > > > > "Fix this" leaves a lot of leeway in isolating the problem. > > Without repeatable results, Valve would have to guess at the cause > > and risk ruining all servers with their attempted fix. I don't > > think any of us want to go down that road. > > > > - Dan > > > > * Dan Sorenson DoD #1066 A.H.M.C. #35 [EMAIL PROTECTED] * > > * Vikings? There ain't no vikings here. Just us honest farmers. * > > * The town was burning, the villagers were dead. They didn't need * > > * those sheep anyway. That's our story and we're sticking to it. * > > > > > > ___ > > To unsubscribe, edit your list preferences, or view the list archives, > please visit: > > http://list.valvesoftware.com/mailman/listinfo/hlds > > > > ___ > To unsubscribe, edit your list preferences, or view the list archives, > please visit: > http://list.valvesoftware.com/mailman/listinfo/hlds > -- ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds
Re: [hlds] Re: Servers losing VAC, gettin beyond a joke...
Thats fine...but how do we know if its a 3rd party problem or a valve problem if we cant see why it crashed? Whisper wrote: -- [ Picked text/plain from multipart/alternative ] You would send the mdmp's to Valve That is what they are there for. On 1/18/06, Hell Phoenix <[EMAIL PROTECTED]> wrote: The problem is that it doesnt put anything into the event log, or logs. It just quits, creates a mdmp and thats it. Now unless you are a programmer (and I am not) it is not exactly easy to figure out the mdmp files. Trust me, I have tried. I am MCSA certified (on my way to mcse) so I am not exactly dumbfounded when it comes to Windows. So when valve helps us to help them, maybe we can include more things =/ Dan Sorenson wrote: At 02:15 PM 1/16/2006 +0100, Saint K. wrote: Currently we have to reset our servers multiple times a day because they are losing VAC. [...] We need VAC fixed asap as this is getting out of controle!! I run both Linux and Windows servers, and hence subscribe to both lists. In the Linux list you'll see posted logfile exerpts, error messages, core dumps, stuff the dev team can use to fix a problem. You'll also see troubleshooting steps taken, various library versions that Source depends upon, etc... In the windows list you see "Fix this." I suspect the problem has been so long in the offering because Valve can't reliably reproduce it. Help them. They're a small team, why not put the event logs and any Dr. Watson's and server logfiles on a webserver and help them out by providing the information they need to start working the problem? "Fix this" leaves a lot of leeway in isolating the problem. Without repeatable results, Valve would have to guess at the cause and risk ruining all servers with their attempted fix. I don't think any of us want to go down that road. - Dan * Dan Sorenson DoD #1066 A.H.M.C. #35 [EMAIL PROTECTED] * * Vikings? There ain't no vikings here. Just us honest farmers. * * The town was burning, the villagers were dead. They didn't need * * those sheep anyway. That's our story and we're sticking to it. * ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds -- ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds
Re: [hlds] Re: Servers losing VAC, gettin beyond a joke...
-- [ Picked text/plain from multipart/alternative ] You would send the mdmp's to Valve That is what they are there for. On 1/18/06, Hell Phoenix <[EMAIL PROTECTED]> wrote: > > The problem is that it doesnt put anything into the event log, or logs. > It just quits, creates a mdmp and thats it. Now unless you are a > programmer (and I am not) it is not exactly easy to figure out the mdmp > files. Trust me, I have tried. I am MCSA certified (on my way to mcse) > so I am not exactly dumbfounded when it comes to Windows. So when valve > helps us to help them, maybe we can include more things =/ > > Dan Sorenson wrote: > > At 02:15 PM 1/16/2006 +0100, Saint K. wrote: > > > >> Currently we have to reset our servers multiple times a day because > they are > >> losing VAC. [...] > >> We need VAC fixed asap as this is getting out of controle!! > >> > > > > I run both Linux and Windows servers, and hence subscribe > > to both lists. In the Linux list you'll see posted logfile exerpts, > > error messages, core dumps, stuff the dev team can use to fix a > > problem. You'll also see troubleshooting steps taken, various > > library versions that Source depends upon, etc... > > > > In the windows list you see "Fix this." > > > > I suspect the problem has been so long in the offering > > because Valve can't reliably reproduce it. Help them. They're > > a small team, why not put the event logs and any Dr. Watson's and > > server logfiles on a webserver and help them out by providing the > > information they need to start working the problem? > > > > "Fix this" leaves a lot of leeway in isolating the problem. > > Without repeatable results, Valve would have to guess at the cause > > and risk ruining all servers with their attempted fix. I don't > > think any of us want to go down that road. > > > > - Dan > > > > * Dan Sorenson DoD #1066 A.H.M.C. #35 [EMAIL PROTECTED] * > > * Vikings? There ain't no vikings here. Just us honest farmers. * > > * The town was burning, the villagers were dead. They didn't need * > > * those sheep anyway. That's our story and we're sticking to it. * > > > > > > ___ > > To unsubscribe, edit your list preferences, or view the list archives, > please visit: > > http://list.valvesoftware.com/mailman/listinfo/hlds > > > > ___ > To unsubscribe, edit your list preferences, or view the list archives, > please visit: > http://list.valvesoftware.com/mailman/listinfo/hlds > -- ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds
Re: [hlds] Re: Servers losing VAC, gettin beyond a joke...
The problem is that it doesnt put anything into the event log, or logs. It just quits, creates a mdmp and thats it. Now unless you are a programmer (and I am not) it is not exactly easy to figure out the mdmp files. Trust me, I have tried. I am MCSA certified (on my way to mcse) so I am not exactly dumbfounded when it comes to Windows. So when valve helps us to help them, maybe we can include more things =/ Dan Sorenson wrote: At 02:15 PM 1/16/2006 +0100, Saint K. wrote: Currently we have to reset our servers multiple times a day because they are losing VAC. [...] We need VAC fixed asap as this is getting out of controle!! I run both Linux and Windows servers, and hence subscribe to both lists. In the Linux list you'll see posted logfile exerpts, error messages, core dumps, stuff the dev team can use to fix a problem. You'll also see troubleshooting steps taken, various library versions that Source depends upon, etc... In the windows list you see "Fix this." I suspect the problem has been so long in the offering because Valve can't reliably reproduce it. Help them. They're a small team, why not put the event logs and any Dr. Watson's and server logfiles on a webserver and help them out by providing the information they need to start working the problem? "Fix this" leaves a lot of leeway in isolating the problem. Without repeatable results, Valve would have to guess at the cause and risk ruining all servers with their attempted fix. I don't think any of us want to go down that road. - Dan * Dan Sorenson DoD #1066 A.H.M.C. #35 [EMAIL PROTECTED] * * Vikings? There ain't no vikings here. Just us honest farmers. * * The town was burning, the villagers were dead. They didn't need * * those sheep anyway. That's our story and we're sticking to it. * ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds
Re: [hlds] Re: Servers losing VAC, gettin beyond a joke...
Vac needs fixing, as well as the cvar that will enable a wallhack.. ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds
[hlds] Re: Servers losing VAC, gettin beyond a joke...
At 02:15 PM 1/16/2006 +0100, Saint K. wrote: >Currently we have to reset our servers multiple times a day because they are >losing VAC. [...] >We need VAC fixed asap as this is getting out of controle!! I run both Linux and Windows servers, and hence subscribe to both lists. In the Linux list you'll see posted logfile exerpts, error messages, core dumps, stuff the dev team can use to fix a problem. You'll also see troubleshooting steps taken, various library versions that Source depends upon, etc... In the windows list you see "Fix this." I suspect the problem has been so long in the offering because Valve can't reliably reproduce it. Help them. They're a small team, why not put the event logs and any Dr. Watson's and server logfiles on a webserver and help them out by providing the information they need to start working the problem? "Fix this" leaves a lot of leeway in isolating the problem. Without repeatable results, Valve would have to guess at the cause and risk ruining all servers with their attempted fix. I don't think any of us want to go down that road. - Dan * Dan Sorenson DoD #1066 A.H.M.C. #35 [EMAIL PROTECTED] * * Vikings? There ain't no vikings here. Just us honest farmers. * * The town was burning, the villagers were dead. They didn't need * * those sheep anyway. That's our story and we're sticking to it. * ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds
[hlds] Re: Servers losing VAC, gettin beyond a joke...
At 09:23 PM 1/16/2006 -0600, Skel wrote: >oh man, people stealing steam ids is annoying.. i had this one person in my >server the other day that stole the steam id of one of my admins... yea that >was fun... cuz then they had full admin through mani. this is like the 3rd >time i have banned this person by ip, they keep comming back with a >different, i know its the same person by how they boast and brag about it. Forward the IP addresses to Alfred directly with an explanation. Most DSL and cable companies use DHCP to hand out addresses, so they'll change every few days if a user resets his modem and lets the lease time expire. However, there are ways to track this and stealing Steam ID's is likely in violation of his ISP's Terms of Use, so Valve will have leverage to use on the ISP to get those logs and find the miscreant. Valve can then ask that a static IP be assigned for a short time to get this user's technique captured via a packet sniffer or the like. The local ISP now has leverage on the miscreant -- disconnection of service followed by a $75 re-connect fee as a new user if he wants new service, and of course there's no discount rate at that time. It's really up to the ISP, but you'd be amazed at how much fun most network engineers consider it to mess with hackers and script kiddies so I suspect Valve will have full cooperation. Dan * Dan Sorenson DoD #1066 A.H.M.C. #35 [EMAIL PROTECTED] * * Vikings? There ain't no vikings here. Just us honest farmers. * * The town was burning, the villagers were dead. They didn't need * * those sheep anyway. That's our story and we're sticking to it. * ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds