RE: Odd Tomcat 3.3m3 Problem
Routers and fire walls may have settings in them that decide that a connection has been idle too long and close it. We had a problem with an installed product (ours) that kept having problems connecting to other components. (Distributed app). After months of finger wagging and getting chewed out by the customer and getting the threats of pull outs and all that their particular make of Cisco router had this setting set to kill those idle connections by default. When we set the setting to ignore how long a connection was sitting idle, the problem went away. At 11:16 AM 6/26/2001, you wrote: >On Tue, 26 Jun 2001 16:20:27 +0200, GOMEZ Henri wrote: > > >What's you workers.properties file, could you send us a copy ? > > > > >in ajp12 you'll see a performance loss since Apache will loose > >time to reestablish the connection. > >Yeah, but I figure that's better than what I'm getting now. Based on >the logs, I'm guessing it's having to do a lot of that already with ajp13. > > >Also did you have a firewall > >between Apache and Tomcat, something which could drop connections > >after a certain time of inactivity ? > >Actually, you may have hit part of the problem here. I do have firewalls in >there that could be shutting connections down and causing mod_jk to have >to cycle through connections more often than it normally would. I'll look >into >that...thanks. > >Best Regards, > >Jason > > >> > >> > >>On Tue, 26 Jun 2001 08:20:57 +0200, GOMEZ Henri wrote: > >> > In my mod_jk.log file, I'm getting a lot of different error > messages that must > be getting handled properly as this isn't happening as often > as there are > messages in the log. The one that looks to be of the > >>frequency of this > problem follows: > > [Mon Jun 25 12:01:05 2001] [jk_ajp13_worker.c (579)]: Error > sending request try > another pooled connection > [Mon Jun 25 12:01:08 2001] [jk_ajp13_worker.c (579)]: Error > sending request try another pooled connection > >>> > >>>You seems to be out of thread process on the java side. > >>>Try to increase the number of Ajp13 threads (in server.xml > >>but I didn't > >>>remember how :) > >> > >> > >>
RE: Odd Tomcat 3.3m3 Problem
On Tue, 26 Jun 2001 16:20:27 +0200, GOMEZ Henri wrote: >What's you workers.properties file, could you send us a copy ? >in ajp12 you'll see a performance loss since Apache will loose >time to reestablish the connection. Yeah, but I figure that's better than what I'm getting now. Based on the logs, I'm guessing it's having to do a lot of that already with ajp13. >Also did you have a firewall >between Apache and Tomcat, something which could drop connections >after a certain time of inactivity ? Actually, you may have hit part of the problem here. I do have firewalls in there that could be shutting connections down and causing mod_jk to have to cycle through connections more often than it normally would. I'll look into that...thanks. Best Regards, Jason >> >> >>On Tue, 26 Jun 2001 08:20:57 +0200, GOMEZ Henri wrote: >> In my mod_jk.log file, I'm getting a lot of different error messages that must be getting handled properly as this isn't happening as often as there are messages in the log. The one that looks to be of the >>frequency of this problem follows: [Mon Jun 25 12:01:05 2001] [jk_ajp13_worker.c (579)]: Error sending request try another pooled connection [Mon Jun 25 12:01:08 2001] [jk_ajp13_worker.c (579)]: Error sending request try another pooled connection >>> >>>You seems to be out of thread process on the java side. >>>Try to increase the number of Ajp13 threads (in server.xml >>but I didn't >>>remember how :) >> >> >>
RE: Odd Tomcat 3.3m3 Problem
Thanks for the idea, but that's not it. I'm getting data back, it's just the wrong data from the wrong context. That problem usually results in an Apache generated error. As far as 3.3m3 is concerned, that old problem doesn't seem to be there anymore. I haven't had any problems stopping and starting Tomcat and getting pages from Apache. Seems to recover nicely with this version. jck On Tue, 26 Jun 2001 10:05:18 -0400, Randy Layman wrote: > > I'm going to throw my 2 cents worth in here, did you, perhaps, stop >Tomcat after starting Apache? I seem to remember that every time you stop >Tomcat, you must also stop Apache, then start Tomcat and start Apache. > > Randy > > >> -Original Message- >> From: GOMEZ Henri [mailto:[EMAIL PROTECTED]] >> Sent: Tuesday, June 26, 2001 10:20 AM >> To: [EMAIL PROTECTED] >> Subject: RE: Odd Tomcat 3.3m3 Problem >> >> >> >I know where to find the documentation on increasing the >> >threads, but I thought it was set to 100 by default. If that's >> >the case, I'd love to know how 1 user could exhaust 100 >> >threads in a little less than 30 minutes. This machine isn't >> >live yet...I've just been doing testing on our new setup. >> >> 100 max by default but not a startup :) >> >> >Keep in mind, these entries below are from the Apache >> >machine, not Tomcat. My thought was that mod_jk was >> >caching connections to Tomcat, and when they needed >> >to be used again, the connection was gone resulting in >> >the errors below. I could only guess that the error wasn't >> >handled properly resulting in the odd results. >> >> What's you workers.properties file, could you send us a copy ? >> >> >Anyway, thanks for the recommendation. I'll switch my >> >thread settings to make sure I know what they are and >> >will see what happens. If that doesn't work, I guess I'll >> >go to ajp12. Looks like I've got to find time to dig into the >> >mod_jk code and work some of these problems out for >> >myself. >> >> in ajp12 you'll see a performance loss since Apache will loose >> time to reestablish the connection. Also did you have a firewall >> between Apache and Tomcat, something which could drop connections >> after a certain time of inactivity ? >> >> >> >Best Regards, >> > >> >Jason >> > >> > >> >On Tue, 26 Jun 2001 08:20:57 +0200, GOMEZ Henri wrote: >> > >> >>>In my mod_jk.log file, I'm getting a lot of different error >> >>>messages that must >> >>>be getting handled properly as this isn't happening as often >> >>>as there are >> >>>messages in the log. The one that looks to be of the >> >frequency of this >> >>>problem follows: >> >>> >> >>>[Mon Jun 25 12:01:05 2001] [jk_ajp13_worker.c (579)]: Error >> >>>sending request try >> >>> another pooled connection >> >>>[Mon Jun 25 12:01:08 2001] [jk_ajp13_worker.c (579)]: Error >> >>>sending request try another pooled connection >> >> >> >>You seems to be out of thread process on the java side. >> >>Try to increase the number of Ajp13 threads (in server.xml >> >but I didn't >> >>remember how :) >> > >> > >> > >>
RE: Odd Tomcat 3.3m3 Problem
> I'm going to throw my 2 cents worth in here, did you, >perhaps, stop >Tomcat after starting Apache? I seem to remember that every >time you stop >Tomcat, you must also stop Apache, then start Tomcat and start Apache. No more a problem on mod_jk for TC 3.3-m3 :)
RE: Odd Tomcat 3.3m3 Problem
I'm going to throw my 2 cents worth in here, did you, perhaps, stop Tomcat after starting Apache? I seem to remember that every time you stop Tomcat, you must also stop Apache, then start Tomcat and start Apache. Randy > -Original Message- > From: GOMEZ Henri [mailto:[EMAIL PROTECTED]] > Sent: Tuesday, June 26, 2001 10:20 AM > To: [EMAIL PROTECTED] > Subject: RE: Odd Tomcat 3.3m3 Problem > > > >I know where to find the documentation on increasing the > >threads, but I thought it was set to 100 by default. If that's > >the case, I'd love to know how 1 user could exhaust 100 > >threads in a little less than 30 minutes. This machine isn't > >live yet...I've just been doing testing on our new setup. > > 100 max by default but not a startup :) > > >Keep in mind, these entries below are from the Apache > >machine, not Tomcat. My thought was that mod_jk was > >caching connections to Tomcat, and when they needed > >to be used again, the connection was gone resulting in > >the errors below. I could only guess that the error wasn't > >handled properly resulting in the odd results. > > What's you workers.properties file, could you send us a copy ? > > >Anyway, thanks for the recommendation. I'll switch my > >thread settings to make sure I know what they are and > >will see what happens. If that doesn't work, I guess I'll > >go to ajp12. Looks like I've got to find time to dig into the > >mod_jk code and work some of these problems out for > >myself. > > in ajp12 you'll see a performance loss since Apache will loose > time to reestablish the connection. Also did you have a firewall > between Apache and Tomcat, something which could drop connections > after a certain time of inactivity ? > > > >Best Regards, > > > >Jason > > > > > >On Tue, 26 Jun 2001 08:20:57 +0200, GOMEZ Henri wrote: > > > >>>In my mod_jk.log file, I'm getting a lot of different error > >>>messages that must > >>>be getting handled properly as this isn't happening as often > >>>as there are > >>>messages in the log. The one that looks to be of the > >frequency of this > >>>problem follows: > >>> > >>>[Mon Jun 25 12:01:05 2001] [jk_ajp13_worker.c (579)]: Error > >>>sending request try > >>> another pooled connection > >>>[Mon Jun 25 12:01:08 2001] [jk_ajp13_worker.c (579)]: Error > >>>sending request try another pooled connection > >> > >>You seems to be out of thread process on the java side. > >>Try to increase the number of Ajp13 threads (in server.xml > >but I didn't > >>remember how :) > > > > > > >
RE: Odd Tomcat 3.3m3 Problem
>I know where to find the documentation on increasing the >threads, but I thought it was set to 100 by default. If that's >the case, I'd love to know how 1 user could exhaust 100 >threads in a little less than 30 minutes. This machine isn't >live yet...I've just been doing testing on our new setup. 100 max by default but not a startup :) >Keep in mind, these entries below are from the Apache >machine, not Tomcat. My thought was that mod_jk was >caching connections to Tomcat, and when they needed >to be used again, the connection was gone resulting in >the errors below. I could only guess that the error wasn't >handled properly resulting in the odd results. What's you workers.properties file, could you send us a copy ? >Anyway, thanks for the recommendation. I'll switch my >thread settings to make sure I know what they are and >will see what happens. If that doesn't work, I guess I'll >go to ajp12. Looks like I've got to find time to dig into the >mod_jk code and work some of these problems out for >myself. in ajp12 you'll see a performance loss since Apache will loose time to reestablish the connection. Also did you have a firewall between Apache and Tomcat, something which could drop connections after a certain time of inactivity ? >Best Regards, > >Jason > > >On Tue, 26 Jun 2001 08:20:57 +0200, GOMEZ Henri wrote: > >>>In my mod_jk.log file, I'm getting a lot of different error >>>messages that must >>>be getting handled properly as this isn't happening as often >>>as there are >>>messages in the log. The one that looks to be of the >frequency of this >>>problem follows: >>> >>>[Mon Jun 25 12:01:05 2001] [jk_ajp13_worker.c (579)]: Error >>>sending request try >>> another pooled connection >>>[Mon Jun 25 12:01:08 2001] [jk_ajp13_worker.c (579)]: Error >>>sending request try another pooled connection >> >>You seems to be out of thread process on the java side. >>Try to increase the number of Ajp13 threads (in server.xml >but I didn't >>remember how :) > > >
RE: Odd Tomcat 3.3m3 Problem
I know where to find the documentation on increasing the threads, but I thought it was set to 100 by default. If that's the case, I'd love to know how 1 user could exhaust 100 threads in a little less than 30 minutes. This machine isn't live yet...I've just been doing testing on our new setup. Keep in mind, these entries below are from the Apache machine, not Tomcat. My thought was that mod_jk was caching connections to Tomcat, and when they needed to be used again, the connection was gone resulting in the errors below. I could only guess that the error wasn't handled properly resulting in the odd results. Anyway, thanks for the recommendation. I'll switch my thread settings to make sure I know what they are and will see what happens. If that doesn't work, I guess I'll go to ajp12. Looks like I've got to find time to dig into the mod_jk code and work some of these problems out for myself. Best Regards, Jason On Tue, 26 Jun 2001 08:20:57 +0200, GOMEZ Henri wrote: >>In my mod_jk.log file, I'm getting a lot of different error >>messages that must >>be getting handled properly as this isn't happening as often >>as there are >>messages in the log. The one that looks to be of the frequency of this >>problem follows: >> >>[Mon Jun 25 12:01:05 2001] [jk_ajp13_worker.c (579)]: Error >>sending request try >> another pooled connection >>[Mon Jun 25 12:01:08 2001] [jk_ajp13_worker.c (579)]: Error >>sending request try another pooled connection > >You seems to be out of thread process on the java side. >Try to increase the number of Ajp13 threads (in server.xml but I didn't >remember how :)
RE: Odd Tomcat 3.3m3 Problem
>I have one site running three contexts with Apache 1.3.19 on FreeBSD >running against Tomcat 3.3m3 on a Solaris machine. Right now, I'm >using ajp13. > >Everything seems to be going fine except every once in a while, almost >randomly, I get back a directory listing of the first context >instead of the data >requested. All I have to do is hit reload on the page, and >the correct data >comes up. > >In my mod_jk.log file, I'm getting a lot of different error >messages that must >be getting handled properly as this isn't happening as often >as there are >messages in the log. The one that looks to be of the frequency of this >problem follows: > >[Mon Jun 25 12:01:05 2001] [jk_ajp13_worker.c (579)]: Error >sending request try > another pooled connection >[Mon Jun 25 12:01:08 2001] [jk_ajp13_worker.c (579)]: Error >sending request try another pooled connection You seems to be out of thread process on the java side. Try to increase the number of Ajp13 threads (in server.xml but I didn't remember how :)