| 00:55.34 | *** join/#tomcat puff (i=puff@66.45.34.102) |
| 00:56.13 | puff | Evening. |
| 01:01.17 | jasonb | Hi puff. |
| 01:10.07 | *** join/#tomcat octoberdan (n=octoberd@pool-141-154-235-226.bos.east.verizon.net) |
| 01:11.10 | octoberdan | Is it possible to access Realm role information from a servlet/ |
| 01:12.31 | octoberdan | and is it possible to have a certain session bean be atatched depending on the user's Database info...? |
| 02:23.58 | *** join/#tomcat vinse (n=Administ@208.253.223.146) |
| 02:40.58 | puff | jasonb: Still having problems. I doubled the mysql max_connections from 500 to 1000. Now we're seeing a lot of "java.net.SocketException: Too many open files" |
| 02:41.51 | puff | jasonb: My theory is that previosuly, mysqld was running out of connections, and eventuallyk tomcat got stuck waiting for a connection. |
| 02:42.33 | puff | jasonb: And now, since mysql has twice as many connections, I suspect that tomcat is exceeding some resource limit. |
| 02:43.43 | puff | jasonb: Open files, obviously, being the chief suspect. I have a "ulimit -n 4096" line in the shell script that starts up tomcat, but I'm not certain that value gets propagated. |
| 02:46.42 | puff | Needless to say, this is costing me sleep :-(. |
| 02:47.28 | puff | 41 webapps, each with its own dbcp connection pool, running in the same process space. |
| 03:41.29 | lyken | puff, what about doing it interactively? |
| 03:41.29 | lyken | ulimit -n 4000 |
| 03:41.29 | lyken | then calling tomcat script? |
| 03:45.35 | lyken | puff are you sure that all the webapps are using the connection pool? |
| 03:54.40 | jasonb | lyken: I believe that his ulimit setting is in effect, but if there's any doubt, one can simply write it right into the catalina.sh script (or whichever script is being used to start Tomcat). |
| 03:54.58 | lyken | yer |
| 03:54.59 | jasonb | puff: Another suggestion: Do not allocate one connection pool per webapp. That's bad. |
| 03:55.10 | lyken | yer.. |
| 03:55.11 | jasonb | puff: Especially if you have that many webapps. |
| 03:55.24 | lyken | cause that would possibly reach the limit.. |
| 03:56.27 | jasonb | It could, depending on his settings (how big of a pool is each one, max, for instance? And, how many max spare?). But, also, it doesn't help to have lots of pools, each with some spare sitting around. It's way better for them all to share one pool. |
| 03:56.43 | lyken | yer |
| 03:58.15 | jasonb | Now, knowing that, to allocate up to 4000 file descriptors, they don't all have to be sockets open to MySQL. Each incoming request socket is a file descriptor, each file that is open, including jar files with classes in them, each JSP, HTML, etc. is a file descriptor.. plus the mysql connections. |
| 03:58.39 | jasonb | Plus things like a descriptor open to /dev/null and /dev/urandom and others that you don't know you're allocating but you are. :) |
| 04:01.16 | jasonb | lyken: How are things going for you? Keeping up with the teaching load? :) |
| 04:01.29 | lyken | in class atm |
| 04:01.31 | lyken | very tired |
| 04:01.35 | lyken | i had a big weekend |
| 04:01.38 | lyken | my first dj gig |
| 04:06.11 | jasonb | Sounds fun. |
| 04:06.15 | jasonb | Congrats on that. |
| 04:06.24 | jasonb | Now you're going to be tired all week though! |
| 04:06.42 | lyken | nah |
| 04:06.44 | lyken | just today |
| 04:07.35 | vinse | i'm tired cuz i spent all weekend trying to work from home |
| 04:07.39 | vinse | only to give up |
| 04:07.44 | lyken | ick |
| 04:07.45 | vinse | and now i'm in the office on sunday night :( |
| 04:07.49 | lyken | bugga |
| 04:07.56 | lyken | i try to keep work and home seperate |
| 04:08.03 | lyken | otherwise the line gets blurry nd i work too much |
| 04:08.08 | vinse | i love being able to work from home |
| 04:08.08 | lyken | and i get grumpy easily |
| 04:08.22 | lyken | ohi answer emails and newsgroup stuff from my students |
| 04:08.32 | lyken | and throw together slides for lectures |
| 04:08.37 | lyken | but no coding or anything |
| 04:08.38 | vinse | using eclipse wiht a remote workspace over vpn is useless though i discovered |
| 04:08.47 | vinse | everything takes so long it's a joke |
| 04:08.49 | lyken | i get too many |
| 04:08.50 | lyken | :( |
| 04:08.53 | lyken | at home that is |
| 04:09.02 | jasonb | vinse: You should use rsync. |
| 04:09.12 | lyken | 2 room mates always making noise, and turntables that scream for me to play them :) |
| 04:09.28 | vinse | jasonb: yeah, the remote workspace part is just not practical |
| 04:09.55 | vinse | jasonb: but i got to the point where i couldnt try to solve that anymore and had to actually do work |
| 04:09.58 | vinse | so i had to come in :/ |
| 04:10.58 | jasonb | vinse: Keep an rsync'd copy of your whole workspace at home, and then when you need to work on the latest stuff from home just rsync the diffs over to your place. |
| 04:11.16 | jasonb | vinse: Then, when you're done working at home, rsync the changes back over to your work (while you sleep). |
| 04:35.53 | *** join/#tomcat unlord_ (n=nathan@va-69-69-251-168.dyn.sprint-hsd.net) |
| 04:36.36 | jasonb | vinse: Oh, and I can give you a script to help with that if you'd like one. |
| 04:45.09 | lyken | lol |
| 04:45.14 | lyken | script the world! |
| 04:47.06 | vinse | jasonb: a script that will work for MS? |
| 04:47.44 | vinse | (OS is not up to me, if i can preempt comments from the peanut gallery) |
| 04:49.34 | jasonb | vinse: The script will work fine if you install cygwin on your MS OS. |
| 04:50.03 | vinse | of course i have cygwin ;p |
| 04:50.13 | vinse | jasonb: hmm it seems to not have rsync by default |
| 04:50.17 | vinse | but i'm sure that's fixable |
| 04:50.35 | jasonb | Then you have rsync.. My script also uses ssh, and depending on how you use it you may need to run an sshd on your windoze side. |
| 04:50.54 | jasonb | Yeah, you can make it download and install its own rsync package. |
| 04:54.50 | jasonb | The cool thing is that you can run rsync on either side of the connection, you just have to script that differently on each side. It's still pretty easy. |
| 04:55.55 | vinse | wow yeah, that'd be useful in spades |
| 04:56.56 | vinse | i'll let you know when i'm ready to try it out |
| 04:57.00 | vinse | thx in advance ;) |
| 04:57.10 | jasonb | No prob. |
| 05:00.13 | lyken | i use dd to make exact copies hehe |
| 05:23.26 | jasonb | lyken: That would take an awfully long time to sync up a whole laptop to a backup server that way. Heh. |
| 05:23.58 | lyken | lol yer |
| 05:24.43 | jasonb | lyken: For quite some time I've been trying to figure out a relatively easy way to use subversion to do laptop backups, since it will do binary diffs and send just those to the server, instead of sending 100% of each file that changed (this is what rsync does). |
| 05:25.32 | lyken | yer |
| 05:27.34 | lyken | couldnt you use diff + svn ? |
| 05:27.48 | lyken | or does diff only work on text files? |
| 05:28.42 | jasonb | svn already has diff. |
| 05:28.55 | jasonb | And, yes, diff only works for text, not binary. |
| 05:29.09 | jasonb | svn has a binary diff engine built in. |
| 05:29.19 | lyken | well, whats the problem if it can do binary diffs taht stops you from using it |
| 05:30.09 | jasonb | How do you tell it that every file on a machine is now checked into the repository? It would write a .svn dir into every dir on the machine (most of them would work, but some would get in the way a bit).. |
| 05:31.06 | lyken | including applications? |
| 05:31.08 | lyken | hmms |
| 05:31.11 | jasonb | I want to tell it that everything is owned by the repo, but not to write .svn dirs everywhere. I don't think the client supports that.. the default client I think. |
| 05:31.54 | lyken | hack it ;) |
| 05:31.58 | jasonb | There are other svn clients, including a Java one.. but I haven't tried (yet) to modify it to do this. It could create an alternate tree somewhere to keep track of the .svn content.. |
| 05:31.59 | jasonb | Yeah. |
| 05:32.06 | jasonb | I haven't spent time to do it yet. |
| 05:32.23 | jasonb | Plus, I don't know how long it would take to commit one's entire laptop each time.. :) |
| 05:32.39 | lyken | hmm |
| 05:32.46 | lyken | sounds like a fun little project |
| 05:32.47 | jasonb | But, even if it takes a while, that might be okay. |
| 05:32.49 | lyken | i might give it a shot :) |
| 05:33.19 | jasonb | If someone could get that working, backups for Linux would not only be easy, they'd be way, WAY efficient, plus the entire history of what changed would be available for perusing. |
| 05:33.48 | lyken | id also throw it into a pg db |
| 05:33.48 | lyken | hehe |
| 05:33.56 | lyken | then you can sql search your filesystem :) |
| 05:34.02 | jasonb | SVN is great for this because it versions directories themselves, plus it recognizes symlinks on non-windoze boxes. |
| 05:34.08 | lyken | like hardcore locate |
| 05:34.14 | jasonb | But, also, when a dir gets renamed, you'd have to tell svn that. |
| 05:34.24 | lyken | nope |
| 05:34.27 | jasonb | Oh, and files too. |
| 05:34.30 | lyken | just make it to a check on file locations |
| 05:34.41 | jasonb | huh> |
| 05:34.43 | jasonb | huh? |
| 05:34.52 | lyken | if filex is in diry |
| 05:34.58 | lyken | and filex is now in dirw |
| 05:35.20 | lyken | maybe its the same dir |
| 05:35.21 | lyken | heh |
| 05:35.32 | jasonb | "maybe"? hehe! |
| 05:35.43 | lyken | its an assumption you can make :) |
| 05:35.55 | jasonb | I suppose one could watch which inode it came from, and track it that way. |
| 05:36.01 | lyken | yer |
| 05:36.06 | lyken | that was my other suggestion :) |
| 05:36.09 | jasonb | But, native code to do that I think. |
| 05:36.31 | jasonb | Anyway, one day there will be such a thing. |
| 05:36.31 | lyken | yer just patch svn to track inodes also |
| 05:36.51 | lyken | it;d just be another attribute for the file/dir |
| 05:36.57 | lyken | that it reads when adding it |
| 05:38.01 | jasonb | I wanted the server to be a revision control system so that I could see revisions of the files between backups. |
| 05:38.44 | vinse | jasonb: yeah i was reading about something like that |
| 05:38.49 | jasonb | So, for example I could check out last monday's version of my laptop's content (or just a subdir of that) to get something specific back. Talk about great incremental backups! |
| 05:38.55 | vinse | cant for the life of me remember where |
| 05:38.59 | jasonb | vinse: Something that already exists? |
| 05:39.09 | vinse | no |
| 05:39.44 | vinse | a file system that stores revisions by default |
| 05:39.59 | jasonb | JFS? |
| 05:40.05 | jasonb | Journal File System? |
| 05:40.16 | vinse | i really cant remember |
| 05:40.44 | vinse | basically though, it almost didnt have a concept of "saving", only "labeling" |
| 05:41.22 | jasonb | There are more than one journaling file system. But, I don't know how much data that really stores. What I'm talking about would not only store all changes (at backup snapshot intervals only), but it would offer it up like any decent revision control system would. |
| 05:41.47 | lyken | resier does lots of cool stuff like this too |
| 05:41.54 | jasonb | And, on the server side, it wouldn't store lots of copies of whole files. It would just store intra-file diffs for everything, which is way smaller. |
| 05:42.06 | lyken | yer |
| 05:42.21 | vinse | jasonb: right, that was the concept that i was reading, and it was supposed to use svn on the back end |
| 05:43.04 | vinse | but regardless, one day that will be the expected behavior of all file systems |
| 05:43.07 | vinse | you're right about that |
| 05:43.24 | jasonb | And, to top it all off, SVN uses "shared nodes". Say you have ten copies of a 2G file (yeah, just an example here), each with the exact same content. Check that into CVS and you just stored 20G on the server. Check it into SVN and you just stored 2G plus a few bytes to convey that it's stored ten times, and in what paths. |
| 05:44.03 | jasonb | vinse: Really? I'd love to find out who else had this idea, and I'd like to watch them to see if they ever actually implement it.. |
| 05:44.21 | vinse | jasonb: i wish i could remember where i read it |
| 05:44.37 | jasonb | vinse: Keep trying.. |
| 05:44.46 | vinse | but you know how sometimes something is on the tip of your tongue and you know you'll remeber it soon |
| 05:44.49 | vinse | this isnt like that ... |
| 05:44.51 | vinse | i have no idea |
| 05:46.03 | jasonb | bummer. |
| 05:46.42 | jasonb | Did you guys see the IBM JDK 1.5 beta is out? |
| 05:46.51 | lyken | yer you told me |
| 05:46.52 | lyken | lol |
| 05:47.45 | vinse | jasonb: for all i know, it could have been you talking about it ... you ever discussed it in this channel before? |
| 05:47.46 | lyken | is svn from tigris ? |
| 05:48.20 | jasonb | vinse: Nope. |
| 05:48.24 | jasonb | lyken: Yep. |
| 05:48.34 | vinse | the only reason it stands out in my mind is because i remember wondering how they decided when to save, like how many changes makes up a revision? |
| 05:48.43 | vinse | session based? time increments? |
| 05:48.44 | lyken | jasonb: svn tracks file/dir renames |
| 05:49.14 | vinse | every keystroke tracked? |
| 05:49.18 | jasonb | lyken: Yeah, you have to do "svn rename fromfile tofile".. |
| 05:49.42 | lyken | arh |
| 05:49.53 | jasonb | Or, "svn mv fromfile tofile" |
| 05:50.50 | jasonb | One could write a script to crawl the fs and find all new files/dirs and do an svn add on each one, and also find stuff that isn't on the file system anymore and do an svn rm on each one.. but then you lose some history. |
| 05:51.10 | jasonb | And, you couldn't take advantage of binary diffs when you move a file or a dir. |
| 05:51.19 | jasonb | So that would make it inefficient. |
| 05:51.26 | jasonb | (on the network bandwidth) |
| 05:52.12 | jasonb | One could write a kernel module that intercepts system calls to rename, delete, etc.. but then writing that in a way that is easy to install on random Linux boxen is not an easy thing to do. |
| 05:52.24 | lyken | hmmz |
| 05:52.37 | jasonb | See why I haven't already done this? :) |
| 06:02.35 | lyken | http://subversion.tigris.org/faq.html#in-place-import |
| 06:04.50 | jasonb | lyken: Yes. It's possible to make a whole laptop into one big svn working copy, but then to track changes isn't easy because the OS will move files and dirs around without knowing that svn is supposed to know about those changes. |
| 06:05.04 | lyken | arh |
| 06:05.13 | lyken | i wonder if reiser4 could suport that :) |
| 06:06.18 | jasonb | If I could only get the kernel to just run a script whenever a file delete, rename, write, etc. happens.. without writing a custom kernel module (that isn't part of the kernel regularly). |
| 06:09.02 | vinse | hmm |
| 06:09.24 | vinse | wrap the whole file system in something that can fire events |
| 06:09.37 | lyken | that be some overhead though. |
| 06:09.51 | lyken | vinse: heh, if you're gonna wrap the FS |
| 06:10.13 | lyken | just postgresql backend |
| 06:10.21 | lyken | then you can do a select on the fs :) |
| 06:54.53 | *** join/#tomcat vinse (n=Administ@208.253.223.146) |
| 07:11.34 | *** join/#tomcat harpoon (i=stueber@ultra20.rbg.informatik.tu-darmstadt.de) |
| 07:19.58 | *** join/#tomcat Zangdar (n=jordimcl@ASt-Lambert-109-1-5-166.w193-253.abo.wanadoo.fr) |
| 07:31.50 | *** join/#tomcat jsisso1 (n=sissonj@210-0-83-196.nexnet.net.au) |
| 07:32.15 | *** part/#tomcat jsisso1 (n=sissonj@210-0-83-196.nexnet.net.au) |
| 07:33.53 | lyken | whats the jsp/falsh thingy ? |
| 07:33.56 | lyken | open lazerus ? |
| 07:47.31 | *** join/#tomcat pucko_ (n=andreas@h232n2fls31o809.telia.com) |
| 07:55.08 | *** join/#tomcat mh_le (n=mads@port328.ds1-sg.adsl.cybercity.dk) |
| 07:55.20 | mh_le | how do I set the admin password for tomcat? |
| 08:03.44 | vinse | mh_le: i think it's in a config file at first ... tomcat-users.xml? something like that |
| 08:03.56 | mh_le | cool |
| 08:04.24 | *** part/#tomcat vinse (n=Administ@208.253.223.146) |
| 08:05.23 | mh_le | can't find it.. |
| 08:56.17 | GedasRx | hi, i have problem with tomcat 5.5.12. I have 2 jsp files, index.jsp and login.jsp, first is login form, when user enters his name and pass it forwards to login.jsp where authentication starts. Problem is that i can't get username and pass from request |
| 08:56.49 | GedasRx | sometimes i can do it but just sometimes |
| 08:57.07 | GedasRx | i don't understand where is the problem |
| 08:57.16 | GedasRx | can anyone help me? |
| 09:00.51 | *** join/#tomcat geemark (n=geemark@port12.ds1-vey.adsl.cybercity.dk) |
| 09:47.40 | mh_le | it seems that my webserver is trying to redirect everything though JK, and not just /servlet, does anyone know how that can be? |
| 10:05.29 | *** join/#tomcat Zangdar`Away (n=jordimcl@ASt-Lambert-109-1-10-238.w193-251.abo.wanadoo.fr) |
| 11:35.45 | *** join/#tomcat jsisson (n=sissonj@59.167.32.8) |
| 12:12.46 | *** join/#tomcat xr4y (i=xr4y@c-4c2ce055.72-1-64736c11.cust.bredbandsbolaget.se) |
| 12:41.18 | puff | jasonb, lyken: I have three scripts (one in each tomcat dir, one in /etc/tomcat) that invoke tomcat with memory-setting variables, I've added the "ulimit -n 4096" command to each. |
| 12:42.13 | puff | lyken: All 41 apps are using connection pools, I'm beginning to suspect that's one of the problems :-). If each app has 10 connections pooled, 41 apps * 10 connections * 2 tomcat installations = 820 pooled connections. |
| 12:43.14 | puff | jasonb: so yeah, I'm looking at consolidating them all into one big pool. |
| 12:43.51 | puff | jasonb: Got any pointers to tutorials on this sort of thing? I guess I should start by cracking open Tomcat: The Definitive Guide again :-). |
| 13:11.49 | *** join/#tomcat ralg (n=ralg@194.65.138.120) |
| 13:20.40 | ralg | hi all...one question...accessing URL = https://someurl;JSESSIONID=SESSION_ID... does the SESSION_ID travels in plain text over the network? |
| 13:27.20 | *** join/#tomcat Kasica (n=Kasica@aad63.neoplus.adsl.tpnet.pl) |
| 13:27.21 | Kasica | hi |
| 13:27.25 | Kasica | i got error : Cannot create JDBC driver of class '' for connect URL 'null |
| 13:27.32 | Kasica | while loading index.jps |
| 13:27.33 | Kasica | jsp |
| 13:27.37 | Kasica | what is that error |
| 13:31.57 | *** join/#tomcat GedasRx (i=usr1515@c1.edrana.lt) |
| 13:33.13 | *** join/#tomcat Cow_woC (i=Cow_woC@i216-58-12-104.cybersurf.com) |
| 13:33.14 | Cow_woC | hi |
| 13:33.17 | Cow_woC | jasonb: awake? |
| 13:33.24 | Kasica | hi |
| 13:33.29 | Kasica | i got error : Cannot create JDBC driver of class '' for connect URL 'null |
| 13:33.30 | Kasica | what it is |
| 13:33.47 | Cow_woC | Kasica: your JDBC configuration is wrong somewhere |
| 13:34.05 | Kasica | Cow_woC in what ? servlet or server configuration? |
| 13:34.07 | Cow_woC | you're probably not providing the JDBC driver name |
| 13:34.19 | Cow_woC | probably in the webapp xml configuration file |
| 13:34.26 | Cow_woC | well actually it depends |
| 13:34.30 | Cow_woC | it could be in the servlet too ... |
| 13:34.46 | Cow_woC | how do you open your JDBC connections? Do you open them in the webapp container or in the webapp itself? |
| 13:35.42 | Kasica | [root@ns7 public_html]# cat WEB-INF/web.xml |grep jdbc |
| 13:35.42 | Kasica | <PROTECTED> |
| 13:36.23 | Cow_woC | Kasica: I suggest you find yourself a simple JDBC tutorial and go from there |
| 13:36.46 | Kasica | i see |
| 13:36.52 | Kasica | it is my client problerm |
| 13:37.09 | Kasica | i have copied that jdbc in common/lib directory |
| 13:40.21 | Kasica | i dont know what can i do to make work |
| 13:40.27 | Kasica | make it work* |
| 13:56.32 | *** join/#tomcat martind (n=martind@67.71.253.163) |
| 13:57.14 | martind | Anybody knows what this message means: "The Apache Portable Runtime which allows optimal performance in production environments was not found on the java.library.path: /usr/java/jdk1.5.0_05/jre/lib/i386/server:/usr/java/jdk1.5.0_05/jre/lib/i386:/usr/java/jdk1.5.0_05/jre/../lib/i386" |
| 14:20.35 | jasonb | Cow_woC: I'm awake.. but have to get ready for work real soon. |
| 14:39.16 | puff | Blah. |
| 14:39.17 | puff | Wotta day. |
| 14:39.44 | *** join/#tomcat linuxfreck (n=freck@fry.fmi.uni-passau.de) |
| 14:39.56 | puff | linuxfreck: Morning. |
| 14:40.22 | linuxfreck | afternoon, puff |
| 14:41.40 | puff | linuxfreck: Question, /proc/sys/fs/file-nr --> 3683 889 209664 |
| 14:42.23 | linuxfreck | puff: man 5 proc |
| 14:43.36 | puff | linuxfreck: Yeah, read that, I'm a little unclear about the distinction between used/free/max |
| 14:44.11 | puff | linuxfreck: I.e. if the number of "free" handles is low and getting lower, is this a problem? max seems an order of magnitude or two larger than used/free |
| 14:44.29 | puff | linuxfreck: And if it _is_ a problem, if there's some system-wide limit on open file handles, can I boost that? |
| 14:44.58 | linuxfreck | free should be the handles that are unused, but memory has been allocated for them |
| 14:45.12 | puff | linuxfreck: So it's more of an open file handle pool? |
| 14:45.14 | linuxfreck | you did read the comment about inode-max? |
| 14:45.31 | linuxfreck | well I have no idea, but that is the only thing that makes sense to me |
| 14:46.37 | puff | Yeah, this is an older kernel, no inode-max file. |
| 14:48.15 | puff | linuxfreck: Now that I've boosted mysql max_connections to 1000, and set ulimit for mysql and tomcat to 4096 each, I'm getting a lot of exceptions that seem to have a root cause of "java.net.SocketException: Too many open files" |
| 14:58.25 | Zangdar`Away | (Isn't connection pool useful in this kind of environment ?) |
| 14:58.46 | linuxfreck | well he apparently uses a connection pool |
| 15:01.33 | linuxfreck | puff: did you try setting logAbandoned to true and check whether your webapp returns connections properly? |
| 15:06.45 | Cow_woC | jasonb: I was just going to bring your attention to http://issues.apache.org/bugzilla/show_bug.cgi?id=37218 |
| 15:07.17 | *** join/#tomcat Vanuatoo (n=vanob@213.200.4.246) |
| 15:07.33 | Cow_woC | jasonb: That ThreadDeath issue that kept on driving me nuts turns out to be a Tomcat issue after all, not a log4j one like Remy kept on saying. Yoav agrees too. The problem is that Remy still won't let the issue through. |
| 15:11.45 | puff | linuxfreck: Yeah, logAbandoned has always been true, every now and then I see the following in my log: http://pastebin.com/404100 |
| 15:12.44 | *** join/#tomcat bugfixer (n=bugfixer@42gis175.gulftel.com) |
| 15:12.46 | linuxfreck | sounds ok though |
| 15:14.01 | puff | linuxfreck: Just did an "fgrep Abandonedetc | wc" on a catalina.out from a fairly busy day, 47 occurrences. |
| 15:14.29 | puff | Hm... http://64.233.187.104/search?q=cache:QvrEUbRKkaEJ:forum.java.sun.com/thread.jspa%3FthreadID%3D665921%26tstart%3D0+lsof+%22can%27t+identify+protocol%22&hl=en |
| 15:19.58 | puff | Hm, /etc/security/limits is all commented out. |
| 15:53.57 | lilalinux | How can I, while developing, tell tomcat to reload a webapp each time it changed? |
| 15:56.01 | linuxfreck | there is an ant task and an eclipse plugin to do that, lilalinux |
| 15:59.46 | *** join/#tomcat gose (n=gose@216-198-89-214.client.cypresscom.net) |
| 16:01.48 | lilalinux | thx |
| 16:06.20 | gose | how do i get tomcat to automatically compile source in a WEB-INF/src directory to the WEB-INF/classes directory? |
| 16:07.52 | linuxfreck | gose: those are not meant to be compiled by tomcat |
| 16:08.10 | linuxfreck | just use ant or your ide to compile them |
| 16:09.03 | gose | i used to use resin and it was nice having them be automatically compiled.. this isn't possible in tomcat? thx for helping btw |
| 16:10.43 | *** join/#tomcat FreemaniaX (n=Freemani@ip67-153-106-58.z106-153-67.customer.algx.net) |
| 16:10.56 | linuxfreck | well I'm not aware of such a thing for tomcat |
| 16:11.22 | gose | ok, interesting. thx |
| 16:13.51 | puff | gose: Me neither. |
| 16:14.11 | linuxfreck | I didn't know that resin supported that, too. |
| 16:14.25 | linuxfreck | does it also check whether source files were updated? |
| 16:14.25 | puff | linuxfreck: resin is, by all accounts, quite an interesting container. |
| 16:14.52 | puff | linuxfreck: It'll do things like optimize compiled jsps. |
| 16:15.08 | puff | linuxfreck: But then again, they're free to depart from the spec. |
| 16:15.45 | gose | yes, with resin, it automatically recompiled things when i made a change to source |
| 16:15.48 | puff | Hm... come to think of it, I wonder what the impact is on filehandles from having the <Context reloadable="true"> |
| 16:16.03 | *** join/#tomcat coRleone (n=j@host121.201-252-107.telecom.net.ar) |
| 16:16.24 | coRleone | hello people. I need to ask you something... i can't figure out how to do this |
| 16:16.39 | linuxfreck | puff: file handles are freed when associated objects are garbage collected |
| 16:16.58 | coRleone | i need to set up an alias... and map the 'welcome-file' to a different index, in a single webapp |
| 16:17.20 | coRleone | for instance... www.localhost.com goes to index.html, and, admin.localhost.com goes to index2.html |
| 16:17.42 | coRleone | the webapp is exactly the same for both of the 'domains'... but no idea how to do this |
| 16:17.53 | coRleone | if anybody can give me a hint i'll be VERY gratefull |
| 16:19.05 | linuxfreck | why don't you just have two different webapps |
| 16:19.10 | gose | coRleone: the only way i can think to do it is with a redirect from another webapp |
| 16:20.18 | coRleone | not two different webapps because it's basically the same java code... but with less 'actions'. |
| 16:20.57 | coRleone | redirect from another webapp... ?... you mean... create a new webapp, with nothing but the index, and map every link / action to the 'other' webapp ? |
| 16:25.07 | puff | coRleone: Have you read up on tomcat virtual hosts? |
| 16:25.54 | *** join/#tomcat lokkju (n=lokkju@ip53.4.dsl-acs2.ptldor2.iinet.com) |
| 16:25.58 | coRleone | yes puff. But the problem i'm facing is that... i don't need a whole different structure. All i need it to map a subdomain to a different INDEX (in a same webapp) |
| 16:26.01 | lilalinux | gose eclipse automatically recompiles everything if changes are made |
| 16:27.17 | puff | coRleone: I haven't looked too deeply into tomcat virtual hosting, but I haven't seen anything that would preclude having two different virtual hosts that forward to the same webapp. |
| 16:27.46 | *** join/#tomcat jasonb (i=noneoyer@adsl-66-124-73-250.dsl.sntc01.pacbell.net) |
| 16:28.02 | puff | coRleone: I'm not sure if you can have them forward to different home pages within the webapp, or perhaps you'd need a custom Filter that checks the domain and forwards. |
| 16:28.33 | coRleone | thanks puff. I'll figure out another way then. Thanks!!!! |
| 16:29.02 | coRleone | a FILTER!! i haven't thought that. THANKS!!! |
| 16:30.14 | martind | Anybody knows what this message means: "The Apache Portable Runtime which allows optimal performance in production environments was not found on the java.library.path: /usr/java/jdk1.5.0_05/jre/lib/i386/server:/usr/java/jdk1.5.0_05/jre/lib/i386:/usr/java/jdk1.5.0_05/jre/../lib/i386"? |
| 16:30.45 | martind | I get this in catalinba.out, right on startup. Using version 5.5.12. |
| 16:38.09 | puff | martind: No idea about the specifics, but in general it sounds like something's missing from your classpath, or your server installation. |
| 16:38.34 | puff | martind: Have you tried googling on the text of the error? |
| 16:40.30 | coRleone | i'm leaving. THanks a lot, bye! |
| 16:42.13 | *** join/#tomcat LawjoskarAway (n=larry@cpe-065-190-205-196.nc.res.rr.com) |
| 16:52.33 | lokkju | I am trying to run Tomcat as a service on windows 2003 server. I have searched the bugs db, and the mailing list, and I can't find any solutions to my specific problem - I downloaded the Tomcat 5.5 install package, and did a custom install, turning on the service, then the tomcat service manage couldn't start the service (no error, just wouldn't start the service). So, I next tried to start the service from the Windows Service Manager |
| 16:52.33 | lokkju | , and every time I actually go to start that service, I get "System Error 5: Access Denied". I know it is not a permissions issue, because I am logged in as the administrator account, and I can stop and start other services just fine. Anyone have any ideas? |
| 16:52.55 | *** part/#tomcat gose (n=gose@216-198-89-214.client.cypresscom.net) |
| 16:53.19 | lyken | make sure you have jvm 1.5 installed |
| 16:53.34 | lyken | and you're classpaths set up right |
| 16:53.43 | lyken | also check the tomcat logs |
| 16:53.49 | lyken | see if its even been attempted to startup |
| 16:54.09 | lokkju | the JVM is fine, and I can run Tomcat form the command prompt just fine |
| 16:54.13 | lyken | in the $CATALINA_HOME\logs\ directory where you install it |
| 16:54.24 | lyken | well you didnt mention that |
| 16:55.22 | lokkju | sorry :( |
| 16:55.28 | lokkju | and there is nothing in the logs |
| 16:55.37 | lyken | id say check the windows logs |
| 16:55.40 | lyken | but i wouldnt have a clue about that |
| 16:55.43 | lokkju | yep, did that |
| 16:55.47 | lyken | im not a window suser |
| 16:56.33 | lokkju | only thing there is the "Could not start the Apache Tomcat service on Local Computer.\n\nError 5: Access is denied." |
| 16:56.47 | lokkju | usually, access is denied means a permissions issue |
| 16:56.50 | lyken | no clue mate |
| 16:56.56 | lokkju | but in this case I don't see how it could be |
| 16:57.02 | lyken | ive gotta get to sleep, its 3am, ill be late for my lectuers in the morning |
| 16:57.09 | lyken | and im the lecturer. |
| 16:57.18 | lokkju | rofl |
| 17:04.31 | martind | puff: I have tried googling the text and found out there was a Apache Portable Runtime project on the Apache Web site. I am wondering if there is a library from Tomcat distribution that now needs to be installed in the JRE directory in 5.5.12. There is a tomcat-apr.jar file now in the server/lib directory. |
| 17:05.06 | linuxfreck | martind: it's possible to use apr in tomcat, but I'd recommend against it |
| 17:06.00 | linuxfreck | it's a native code library which can be used in the http server part of tomcat |
| 17:06.20 | martind | linuxfeck: So you would recommend I not worry about this warning on startup that flags that APR wasn't found in JRE directory? |
| 17:06.24 | linuxfreck | true |
| 17:06.32 | linuxfreck | it's new in .9 or .12 |
| 17:06.54 | martind | Do you think at some point, when it is stable it will be a good thing to use? |
| 17:07.34 | linuxfreck | if you use it you get minor or not measurable performance improvements with the cost of potential buffer overflows in native c code |
| 17:07.48 | linuxfreck | so I'm not really convinced about the whole idea |
| 17:08.21 | martind | linuxfreck: Not worth it then I guess. |
| 17:09.26 | martind | linuxfreck: Thanks for the feedback. |
| 17:09.34 | linuxfreck | np |
| 17:16.40 | lokkju | rofl - how idiotic is that - it was permissions - somehow the SYSTEM account did not have access to anything on the D drive - normally SYSTEM has full rights everything |
| 17:16.44 | *** part/#tomcat lokkju (n=lokkju@ip53.4.dsl-acs2.ptldor2.iinet.com) |
| 17:18.17 | *** join/#tomcat snibril (n=jh@p54A3DFC0.dip.t-dialin.net) |
| 17:36.57 | *** join/#tomcat vessuvius (n=vamiller@12-223-239-156.client.insightbb.com) |
| 17:37.25 | vessuvius | Is there a tutorial anywhere on setting up Tomcat for virtual hosts? |
| 17:38.16 | vito__ | what do you need to do? I have some experience with tomcat and virtual hosts. |
| 17:39.20 | vessuvius | Well I have two virtual hosts running under apache.... say vessuvius.com and vessuvius.net ... and I want to have webapps for vessuvius.com under /var/www/vessuvius.com/webapps and webapps for vessuvius.net under /var/www/vessuvius.net/webapps |
| 17:39.52 | *** join/#tomcat valmont (n=chrishol@germany.pspdev.pas.earthlink.net) |
| 17:40.25 | vito__ | you would setup tomcat with a CATALINA_BASE and have two instances of tomcat running, plugged into apache with mod_jk, you could use a single apache instance |
| 17:41.10 | vito__ | similar to this article ... http://www.linuxjournal.com/article/8561 |
| 17:41.10 | vessuvius | so I need to run two instances of tomcat? There's not a way to have tomcat itself run two virtual hosts? |
| 17:41.29 | vessuvius | I'll check it out :) |
| 17:41.31 | vessuvius | thnx |
| 17:41.53 | vito__ | there might be... but that would probably be more work |
| 17:42.38 | vessuvius | Just from a resources standpoint it'd be worth it to me :) I can't imagine someone hosting 100 domains would need to run 100 instances of tomcat |
| 17:43.31 | vessuvius | but I guess a bad host would only kill its instance of tomcat that way |
| 17:43.42 | *** join/#tomcat ding (n=ding@pcp06953945pcs.nrockv01.md.comcast.net) |
| 17:43.51 | vito__ | well you could also use apache modproxy to proxy specific webapps of a single tomcat instance to different vhosts |
| 17:44.08 | vito__ | yes... that is a big advantage to multiple instances.... |
| 17:44.37 | vessuvius | I thought there might be a way to just set up different workers... with differen host settings |
| 17:44.59 | vessuvius | anyway... I find the tomcat and apache docs generally not so helpful |
| 17:45.19 | vito__ | oh there certainly is, but I think one of those two options is easier |
| 17:45.22 | vessuvius | I had to figure out and write the wiki for mod_jk on gentoo mysel |
| 17:45.22 | vessuvius | f |
| 17:46.04 | vito__ | <-- cent os user for all servers, fedora for all workstations |
| 17:47.46 | vessuvius | yes, vito, this is the sort of thing I was looking for... thank you very much |
| 17:48.05 | vito__ | no problem, hope it works out for you |
| 17:49.31 | *** join/#tomcat vinse (n=Administ@208.253.223.146) |
| 17:50.26 | *** join/#tomcat ocnarfid9 (n=ocnarfid@207.34.36.50) |
| 18:21.07 | *** join/#tomcat Sashlik (n=Godless@83.221.168.101) |
| 18:24.51 | *** topic/#tomcat by Sashlik -> xdfv |
| 18:25.02 | Sashlik | lol |
| 18:25.16 | *** topic/#tomcat by Sashlik -> http://tomcat.apache.org Latest stable versions of Tomcat are 5.5.12, 4.1.31, and 3.3.2 If you are a newbie, use the official distribution on Jakarta's site. Be sure to check the log files before you attempt to search for a solution. SLOW-MOTION CHANNEL: Ask your question, then wait; check back often to see if anybody answered. |
| 18:27.38 | puff | vessuvius: You might want to try out "Tomcat: The Definitive Guide" |
| 18:28.03 | puff | @rn |
| 18:28.52 | *** join/#tomcat ding (n=ding@pcp06953945pcs.nrockv01.md.comcast.net) |
| 18:37.33 | lilalinux | is there a suitable group for jsp? |
| 18:37.39 | lilalinux | eh channel |
| 18:38.36 | vinse | lilalinux: i dont think tehre's a specific jsp channel, probably ##java is your best bet |
| 18:40.42 | puff | lilalinux: Yeah, #java. Or maybe #web. |
| 18:46.58 | lilalinux | I guess #java will kick&ban me, if I ask newbie jsp questions :) |
| 18:49.13 | *** join/#tomcat levon (n=levon@62.48.88.185) |
| 18:51.09 | *** join/#tomcat putzel (n=in@tor/session/x-3d6e245e9f2b615b) |
| 18:51.18 | *** join/#tomcat levon (n=levon@ad96eab9e.dsl.de.colt.net) |
| 18:53.00 | *** part/#tomcat Cow_woC (i=Cow_woC@i216-58-12-104.cybersurf.com) |
| 18:57.15 | *** join/#tomcat RangerRick (n=ranger@dargo.trilug.org) |
| 18:58.04 | *** join/#tomcat PeterB (n=chatzill@milquetoast.dpc.ucar.edu) |
| 18:59.24 | RangerRick | is it possible to spawn a single worker thread at startup for processing a queue created inside a servlet? my first thought was an init() in my httpservlet, but it looks like that's not re-initialized when undeployed/deployed through the manager, only on actual startup of tomcat |
| 19:00.03 | snibril | RangerRick: do you have loadonstartup=true? |
| 19:01.06 | RangerRick | snibril: hm, guess I don't, is that a web.xml thing? context.xml? something like that? |
| 19:02.14 | snibril | yes, a param in teh servlet definition. don't remmber the exact xml tag though |
| 19:02.25 | *** join/#tomcat ding (n=ding@pcp06953945pcs.nrockv01.md.comcast.net) |
| 19:03.20 | RangerRick | ok, thanks, I'll play with that |
| 19:05.36 | snibril | <load-on-startup>1</load-on-startup> |
| 19:11.06 | RangerRick | awesome, that did it, thanks! |
| 19:11.11 | *** join/#tomcat joered (n=chatter@81-208-83-246.fastres.net) |
| 19:11.14 | *** part/#tomcat RangerRick (n=ranger@dargo.trilug.org) |
| 19:11.30 | puff | Dang, missed RangeRick. |
| 19:11.47 | puff | Apparently the "right" way to do this is not to do it, because you're not supposed to start threads, etc. |
| 19:12.37 | puff | But if you're going to do it, build a context listener that will start the thread when the context loads, and kill it when the context shuts down. |
| 19:12.53 | puff | And also, make it a daemon thread if you can, so your tomcat shutdown doesn't hang. |
| 19:15.01 | jasonb | puff: Yeah, you know about the daemon thread issue.. :) |
| 19:15.12 | puff | jasonb: Heh. Now now... |
| 19:15.31 | jasonb | puff: Also, where does the servlet spec say that a webapp isn't supposed to start threads? |
| 19:15.55 | *** join/#tomcat RangerRick (n=ranger@dargo.trilug.org) |
| 19:16.12 | RangerRick | heh, LawjoskarAway just pasted my missed conversation :) |
| 19:16.39 | RangerRick | honestly, I don't even care to spawn a thread from tomcat, it's mostly because I'm loathe to start a new jvm just to do some cleanup/queing stuff |
| 19:16.42 | Lawjoskar | the way I'd do this RangerRick, is to use a ServletLifecycleListener |
| 19:16.45 | RangerRick | but it's really better done with quartz or somesuch |
| 19:17.25 | Lawjoskar | so this thread won't work for the entire lifetime of your webapp? |
| 19:17.29 | RangerRick | if it's dangerous dealing with threading like that in tomcat, I'm happy to just not bother then, was looking for the easy way out :) |
| 19:18.06 | RangerRick | Lawjoskar: well, I'm going to be doing some stuff asynchronously, essentially, making requests and not caring if a response has returned yet |
| 19:18.07 | PeterB | Hi. I'm needing to deploy under Linux an app from a Windows developer. When I start the application with the manager, it immediately throws a |
| 19:18.09 | PeterB | java.lang.ClassNotFoundException: org.apache.catalina.deploy.FilterDef |
| 19:18.11 | PeterB | when parsing the web.xml, and when I comment out the filters part, it throws |
| 19:18.13 | PeterB | SEVERE: Parse error in application web.xml |
| 19:18.14 | PeterB | java.lang.ClassNotFoundException: org.apache.catalina.Container |
| 19:18.30 | PeterB | I've not seen anything like this with Tomcat 5.0 or 4.X or java 1.4.X. |
| 19:18.35 | PeterB | This is TC5.5.12, and java 1.5.0_05-b05 (Sun) on RHEL 3. |
| 19:18.47 | PeterB | Any idea why it's not seeming to load catalina.jar? |
| 19:19.06 | Lawjoskar | PeterB: my first thought would be to check the capitalization of all classes and jars |
| 19:19.54 | Lawjoskar | PeterB: mis-capitalized stuff will work on Windows, but not on Linux, even inside a jar. I've had jars built by Windows developers completely not work because of a mis-capitalized jar or class file. |
| 19:20.33 | PeterB | Ah-ha -- I have come across that before when working with this developer. I'll poke around... |
| 19:21.37 | Lawjoskar | PeterB: also be aware that a ClassNotFoundException is tricky -- it doesn't necessarily mean that the class name in the exception wasn't found, but that maybe while trying to load that class, another class it depends on wasn't found. Sigh. |
| 19:22.30 | puff | jasonb: The daemon thread issue in my case is inside java.util.concurrent backport; Lea is on record as saying he doesn't think queue manager threads should be daemons, since there are few queue operations that you want to have killed in midsttream. |
| 19:24.28 | linuxfreck | puff: wierd argument, how is a queue manager thread any good if all other (normal) Threads using them are already dead? |
| 19:25.47 | jasonb | puff: Well, that argument all by itself makes sense. But, in Tomcat's case, it's either a daemon thread, or it must take responsibility to make itself robust enough to not hang things.. exit gracefully at *some* point when work is either done or not going to get done. |
| 19:26.23 | jasonb | linuxfreck: Well, if the queue thread can perform work independently, the argument is a good argument. |
| 19:27.12 | linuxfreck | if..., that's the question here |
| 19:27.27 | jasonb | Yes. In puff's use case, this is likely not so. |
| 19:28.03 | jasonb | (and, so since he's doing something with the code that the original author isn't envisioning, then puff should make the threads daemon threads. :) |
| 19:30.48 | jasonb | Anyway, over the years I've often heard people say "you're not supposed to create threads from a webapp" (I've even said it, long ago), and the truth is: it's a perfectly okay thing for a webapp to do (nowhere in the spec does it say you're just not supposed to), but it's not necessarily simple to make it robust, and make it not prevent Tomcat from going about its business. |
| 19:32.19 | RangerRick | jasonb: well it sounds like a ServletContextListener would be perfect for what I"m wanting to accomplish, so I think I'm in good shape |
| 19:32.37 | RangerRick | thanks everyone for the help |
| 19:38.52 | *** join/#tomcat ding (n=ding@pcp06953945pcs.nrockv01.md.comcast.net) |
| 19:40.02 | *** join/#tomcat ding (n=ding@pcp06953945pcs.nrockv01.md.comcast.net) |
| 19:43.52 | *** part/#tomcat ding (n=ding@pcp06953945pcs.nrockv01.md.comcast.net) |
| 20:08.29 | *** join/#tomcat omay (n=Oleg@dslb-084-059-222-045.pools.arcor-ip.net) |
| 20:10.16 | puff | jasonb: I resemble that remark :-). |
| 20:10.47 | puff | jasonb: Still, I'm pretty skeptical that the problem is in the queue manager, I think that one thread dump where it was, was because of other issues tripping up the normal shutdown process. |
| 20:11.26 | puff | jasonb: Especially since the queue manager at that point was long done with its work. Still, maybe I should add something that proactively shuts it down once it's empty. |
| 20:29.10 | *** join/#tomcat arsonx (n=arsonx@adsl-64-254-130-224.totalink.net) |
| 20:55.52 | *** join/#tomcat dreivier (n=dreivier@port-212-202-43-57.dynamic.qsc.de) |
| 21:05.03 | *** part/#tomcat jsisson (n=sissonj@59.167.32.8) |
| 21:32.10 | *** join/#tomcat taf2_ (n=taf2@65.207.103.51) |
| 21:32.31 | taf2_ | i'm getting a bind error on startup of tomcat on my fedora core box |
| 22:06.04 | *** join/#tomcat Zangdar (n=jordimcl@ASt-Lambert-109-1-10-238.w193-251.abo.wanadoo.fr) |
| 22:48.24 | *** join/#tomcat snibri1 (n=jh@p54A3CC88.dip.t-dialin.net) |
| 23:05.37 | *** part/#tomcat vessuvius (n=vamiller@12-223-239-156.client.insightbb.com) |
| 23:48.16 | *** join/#tomcat orthogonal (n=chatzill@24.125.196.142) |
| 23:49.18 | orthogonal | I suddenly getting "ava.io.FileNotFoundException: C:\Program Files\Apache Software Foundation\Tomcat 5.5\webapps\TechEx\META-INF\context.xml (Access is denied)" for an app that previously was working. What do I need to do? |