| 00:00.05 | *** join/#tomcat jasonb (i=noneoyer@adsl-66-124-73-250.dsl.sntc01.pacbell.net) |
| 00:19.04 | *** join/#tomcat Zangdar (n=jordimcl@ASt-Lambert-109-1-8-73.w193-252.abo.wanadoo.fr) |
| 00:42.01 | puff | Hm, anyone awake? |
| 00:42.15 | puff | javax.servlet.ServletException: Error allocating a servlet instance |
| 00:45.38 | puff | Root cause is a null pointer exception. |
| 00:45.46 | puff | http://pastebin.com/421245 |
| 00:50.10 | jasonb | puff: I would bet that's a bug that has since been fixed in the Tomcat trunk.. which ends up making it into newer versions of Tomcat than the one you're running. |
| 00:51.00 | jasonb | puff: It even sounds familiar to me, but in my own tc 5.5 changelog I find nothing about it.. so I would guess it was either fixed in Tomcat 5.0.x, or a later version of Tomcat 4.1 than you're running.. I'd guess 5.0.x. |
| 00:51.32 | jasonb | puff: Do a search for FileDirContext in bugzilla.. maybe you'll find a bug about it. |
| 00:54.02 | puff | http://issues.apache.org/bugzilla/show_bug.cgi?id=24588 |
| 00:54.29 | puff | FileDirContext or ProxyDirContext? |
| 00:56.06 | jasonb | Yeah, you're right.. ProxyDirContext. |
| 00:56.11 | jasonb | Sorry. |
| 00:57.10 | puff | No sweat, for a moment I thought you were refering to some secret :-). |
| 00:57.17 | puff | Hm, i need an ASF bugzilla account. |
| 00:57.48 | jasonb | Yeah, you should create one. |
| 00:58.20 | jasonb | I think this is one of those problems that just went away when code was rearranged for the TC 5.0 and then TC 5.5 refactorings. |
| 00:58.31 | jasonb | Are you able to reproduce this bug? |
| 00:58.38 | puff | Right now I can. |
| 00:58.43 | jasonb | Also, which version of Tomcat are you running? |
| 00:58.50 | puff | Some of the pages I've found on it suggest it will go away after a reboot. |
| 00:58.54 | puff | 4.1.31 |
| 00:59.01 | jasonb | That's great if you can reproduce it -- then it's something you could debug and find the cause of. |
| 00:59.08 | puff | It's only on one app, and one which had a fresh WAR file deployed ot it. |
| 00:59.13 | jasonb | Yeah, if it's 4.1.31 they'll likely fix it if you can show them where the problem is. |
| 00:59.21 | puff | I'd like to figure it out tonight, yeah. |
| 00:59.24 | puff | Any idea what next steps? |
| 00:59.24 | jasonb | Hook up the NetBeans debugger. |
| 00:59.32 | puff | Can you do that live? |
| 01:00.15 | jasonb | Well, you can turn on the JPDA debugger in your Tomcat JVM by adding "jpda" before the "start" in your call to catalina.sh. |
| 01:00.22 | jasonb | catalina.sh jpda start |
| 01:01.00 | jasonb | Then, you can, over the net even, hook up any JPDA debugger via a TCP socket. I think you also need some more parameters .. let me get that out of my RPM's config file. :) |
| 01:02.07 | jasonb | set JAVA_OPTS="-Xdebug -Xrunjdwp:transport=dt_socket,address=8000,server=y,suspend=n" |
| 01:02.48 | jasonb | That's really all you need to change in the startup. |
| 01:03.10 | jasonb | Then, you need to be able to set breakpoints in the source. |
| 01:03.51 | jasonb | So, you need to make a new NetBeans project with the source from your version of Tomcat. |
| 01:04.11 | jasonb | (this will help you anyway, because then it becomes far easier to debug problems..) |
| 01:05.22 | puff | jasonb: Sorry, I was off registering for the ASF bugzilla. |
| 01:05.27 | jasonb | That's ok. |
| 01:05.36 | puff | jasonb: Ah, so I can't enable debugger live. |
| 01:06.07 | puff | I'd have to run the production site under JPDA until the problem shows up. |
| 01:07.58 | jasonb | If you had set the jpda options in the first place, you could hook up to it live at any time, but most people don't run their production system that way. There's a small amount of overhead involved in turning it on. |
| 01:08.09 | puff | Small? I keep hearing "large" |
| 01:08.20 | jasonb | Of course those are relative terms.. |
| 01:08.22 | puff | Sounds like it'd be worth it, for us... |
| 01:08.32 | jasonb | But, I think the overhead is around 7-10%. |
| 01:08.33 | puff | Okay, I'm messing with the stress testing framework now, as soon as I can consistently reproduce the problem I'll take a crack at JDPA. |
| 01:08.35 | puff | Ah. |
| 01:08.39 | puff | In CPU, memory? |
| 01:08.51 | jasonb | CPU mainly I believe. |
| 01:08.54 | puff | I think we have gobs of CPU to spare, so it might easily be worth it for us. |
| 01:09.14 | puff | As far as I can tell, the CPU is hardly noticing our app. |
| 01:09.16 | jasonb | Yep, then you may want to turn it on and see how it affects your performance. |
| 01:09.23 | jasonb | That sounds great. |
| 01:09.37 | puff | Ah, crap, 8pm already. |
| 01:10.13 | jasonb | Too late. You might as well stay and play for a while. :) |
| 01:10.40 | puff | Hm, http://marc.theaimsgroup.com/?l=tomcat-user&m=104378432821451&w=2 |
| 01:10.55 | puff | Says he has reloadable set true, if he sets it to false he gets "too many open files" instead. |
| 01:11.01 | yel | puff jasonb good night its late out here :) 02h10 mrining |
| 01:11.14 | puff | G'night. |
| 01:11.20 | puff | Damn, that's an annoying UI. |
| 01:11.22 | jasonb | See ya. |
| 01:11.25 | *** part/#tomcat yel (n=yel@xdsl-84-44-176-237.netcologne.de) |
| 01:11.39 | puff | I've really gotta get around to writing that private tomcat-users search engine I keep talking about. |
| 01:12.10 | jasonb | The tomcat-users list is actually already searchable. |
| 01:12.26 | puff | Yeah, I know, but all the UIs suck :-) |
| 01:12.38 | puff | gmane's at least provides nntp, so I could in theory write my own search client. |
| 01:13.02 | jasonb | http://marc.theaimsgroup.com/?l=tomcat-user&r=1&w=2 |
| 01:13.07 | puff | This is a project I've been thinking about for years. |
| 01:13.37 | puff | Right, I just referenced a link to a specific message there :-). |
| 01:13.46 | puff | Okay, so the error is spreading to other apps. |
| 01:14.06 | jasonb | The too many open files error is a max file descriptors problem. |
| 01:17.32 | *** join/#tomcat FreemaniaX (n=Freemani@pcp0011117215pcs.elkrdg01.md.comcast.net) |
| 01:20.18 | puff | Yeah. |
| 01:20.25 | puff | Which we shouldn't be having, since we have ulimit set to 4096. |
| 01:20.47 | puff | Is there any good rule of thumb for calculating file descriptor usage by tomcat? |
| 01:21.24 | puff | We may need to tweak the kernel on this box and upgrade it, 4096 is the hard limit. |
| 01:21.38 | puff | And thinking about it, 400 per webapp is maybe a bit low, all things considered. |
| 01:21.55 | puff | Er, 100 per webapp. |
| 01:22.06 | puff | 10 for database connections leave 89. |
| 01:30.43 | jasonb | Yeah, to calculate the file descriptor usage by tomcat, use lsof and grep to find out the exact number. :) |
| 01:31.28 | jasonb | You really don't want to come within 100 file descriptors of your upper limit. That's asking for trouble. |
| 01:32.06 | jasonb | There eventually comes a limit to what you can do in a single process. :) |
| 01:33.28 | puff | Hm. |
| 01:33.41 | puff | So the question is, should we split tomcats or up the file system limit? |
| 01:33.43 | puff | Hm. |
| 01:33.50 | puff | What's the memory overhead per tomcat? |
| 01:34.18 | jasonb | If you already have your file descriptor max set to 4096, it would likely be hard to raise that higher. |
| 01:34.30 | puff | I've read it's tweakable at the OS level. |
| 01:34.40 | puff | I'm mainly thinking that might buy us time. |
| 01:34.43 | jasonb | The memory overhead for Tomcat itself depends to some degree on how many contexts you have deployed in it. |
| 01:34.55 | puff | Well, I meant the non-webapp part. |
| 01:35.08 | puff | I.e. if we set up 41 catalina-bases.... |
| 01:35.34 | jasonb | You could look into tweaking your file descriptor limit higher.. I think it may be raiseable to 16384. But, I don't know how to do that, and it could even require a new kernel binary. It is possible it wouldn't though. I don't know. |
| 01:35.37 | puff | I guess our options are to a) set up on tomcat process per webapp, b) break the webapps up in blocks of 10 or whatever number is appropriate. |
| 01:35.49 | puff | And put something in front of the tomcat process, or port 80, proxying to the tomcat processes. |
| 01:36.22 | jasonb | No, instantiate as few Tomcat runtimes as you can, because each one is a JVM also, and the JVM itself takes more RAM to start up than Tomcat itself does.. For example, each one must allocate heap RAM. |
| 01:37.17 | jasonb | Just use ethernet aliases, and bind multiple Tomcats to the same ethernet device on port 80. Give each ethernet alias its own IP address, and configure each Tomcat with one of those. |
| 01:37.32 | puff | Reallyu? |
| 01:37.33 | puff | Hm. |
| 01:37.36 | jasonb | Sure. |
| 01:37.48 | puff | Have to talk to our hosting center and see if they can arrange that. |
| 01:38.00 | jasonb | I'll bet they can. |
| 01:51.55 | puff | Hm, just checking, the ulimit on the existing tomcat process was putatively set to 8096, though I seem to remember reading that you can't _really_ set it above 4096. |
| 01:52.03 | puff | Without kernel tweaks. |
| 01:54.00 | jasonb | Probably true. |
| 01:54.12 | puff | I think I'm going to have to kick some asses and get us upgraded. |
| 01:54.18 | jasonb | It will just set it back down to the kernel's max. |
| 01:54.53 | jasonb | If it's really a file descriptor problem, upgrading Tomcat may not help. |
| 01:54.57 | puff | Yeah. |
| 01:55.04 | jasonb | Well, it may not help with that one problem, anyway. :) |
| 01:55.35 | puff | The #redhat folks are telling me that somewhere around 2.4, redhat upped the limit. |
| 01:56.16 | jasonb | To what? |
| 01:56.31 | jasonb | 4096 I believe. :) |
| 01:57.27 | puff | Hm: http://support.bea.com/application_content/product_portlets/support_patterns/wls/Too_Many_Open_Files_Pattern.html |
| 01:57.29 | jasonb | There's a kernel limit, and then there's a default shell configuration limit. I'm talking about the former. |
| 01:57.38 | puff | Yeah. |
| 01:57.46 | puff | they seem to be saying that the kernel limit was upped. |
| 01:59.08 | jasonb | To what? |
| 01:59.29 | jasonb | That URL is no good.. it just says how to set the max limit to 4096, which yours is already set to. :) |
| 01:59.34 | puff | Heh. |
| 01:59.40 | puff | Damn, hadn't gotten that far yet. |
| 01:59.54 | puff | #redhat folks seem to think 65536. |
| 02:00.14 | puff | Well, I'm reviewing all the other stuff about file descriptors, to see if it shakes something loose. |
| 02:00.31 | jasonb | So use lsof to count yours up.. doesn't take long. If it's near 4096, then up your limit to 65532 and see if that helps. |
| 02:00.57 | jasonb | Actually, 65535 may work. |
| 02:01.08 | jasonb | (if the kernel limit is indeed 65535. |
| 02:01.12 | jasonb | ) |
| 02:02.04 | puff | lsof | wc is taking forever. |
| 02:02.05 | jasonb | It may also be more than a kernel change.. it may also be a glibc change, or some other low level library like that. |
| 02:04.00 | puff | #qemu? |
| 02:04.25 | jasonb | If you don't know what qemu is, you should really start looking at it. |
| 02:05.14 | puff | Oh yeah, this stuff. |
| 02:05.18 | puff | cool technology. |
| 02:05.31 | jasonb | It's not Java, but it's amazing anyway. I use it just about every day for various things. |
| 02:05.49 | jasonb | And, the guys who work on it basically know everything. |
| 02:05.53 | puff | I have a buncha language/compiler geek friends who hang out at darksleep.com 7777, they talk about stuff like this. |
| 02:06.05 | puff | One of them found an old HP paper, years ago, about this topic. |
| 02:06.26 | puff | Basically the HP research people wanted something to emulate new chip architectures, so they could test out designs. |
| 02:06.30 | puff | So they wrote one. But it was slow. |
| 02:06.35 | puff | So they added a dynamic optimizer. |
| 02:06.51 | jasonb | This one does a whole architecture and CPU that you don't have, and it's still fast! |
| 02:06.56 | puff | Then one day somebody noticed that it ran an emulation of the same chip that was running the emulator, about 20% faster. |
| 02:07.19 | jasonb | hehe |
| 02:07.35 | puff | This is why I had more faith in the transmeta chip and in the potential of the hotspot compiler than most people did, a few years back :-). |
| 02:08.09 | puff | You use this thing a lot every day? |
| 02:08.15 | jasonb | Yes. |
| 02:08.18 | puff | Damn, that's too bad. |
| 02:08.21 | puff | The money, I mean. |
| 02:08.25 | jasonb | Yeah. |
| 02:08.30 | puff | I'm a big curious for what. |
| 02:08.33 | puff | What you use it for. |
| 02:08.54 | puff | But I guess right now I should concentrate more on this issue. |
| 02:09.26 | puff | Where would we do the bind-to-specific-ip thing? |
| 02:09.36 | jasonb | I run Linux on my desktop.. I'm most comfortable using Linux. But, then, all kinds of people I work with basically mandate that I use Microsoft software for work stuff. So, I run XP inside of qemu, and the software inside that (when I can't just use the software on Wine). |
| 02:09.42 | puff | Ah, cool. |
| 02:09.47 | puff | That would be handy as hell, yeah. |
| 02:09.54 | jasonb | For now I'm using a PowerBook G4 (another long story), and so I can't run wine at all. |
| 02:09.58 | puff | I'm going to have to ask you about that, later, since I'm running linux for my desktop. |
| 02:10.06 | jasonb | So, Qemu is all I have on this laptop. |
| 02:10.13 | puff | I have thus far managed to ignore windows. |
| 02:10.20 | puff | Since going fully linux :-). |
| 02:10.44 | jasonb | Also, in my last job, I had to emulate a whole different Linux OS that was a custom Linux distro, and I wasn't going to install that separately on my HD and keep rebooting all day.. so qemu came in nice and handy for that. |
| 02:11.20 | jasonb | And, I also needed the ability to roll back changes to the HD for debugging, and qemu also implements that nicely. So, I was able to debug things that nobody else could debug in the small amount of time I could. :) |
| 02:11.33 | puff | Jeesuz. lsof | wc 625781 5629940 82750965 |
| 02:11.40 | puff | Okay, something weird is going on here... |
| 02:11.47 | puff | There has *gotta* be some kinda file handle leak. |
| 02:11.53 | jasonb | lsof shows *all* of the file descriptors open for the whole system. |
| 02:11.59 | jasonb | You have to grep for the process you're interested in. |
| 02:12.09 | puff | All we have running on this system is tomcat, mysql. |
| 02:12.13 | jasonb | lsof | grep java | wc -l # (maybe) |
| 02:13.37 | puff | I'll check that next, yeah, but 625781 is a hell of al ot of open files. Isn't that near the file-max? |
| 02:13.39 | puff | lsof -i | wc 226289 2036598 27639668 |
| 02:13.55 | puff | lsof | fgrep java | wc -l --> 306512 |
| 02:14.33 | jasonb | Now have a look at that output to see if each line is really an open file descriptor. |
| 02:15.11 | puff | netstat | wc --> 775 4662 68405 |
| 02:15.28 | puff | How do you tell? |
| 02:15.53 | puff | netstat says 775 open connections, lsof -i says 226k connections. |
| 02:17.10 | jasonb | netstat also lists UNIX domain sockets. |
| 02:17.27 | jasonb | lsof -i will show all network connections for the entire host, for all processes at once. |
| 02:17.44 | puff | Yeah. But again, all this box is being used for is to run tomcat and mysql. |
| 02:18.10 | jasonb | Do "netstat -an | grep CLOSE_WAIT" |
| 02:18.15 | jasonb | Do you get lots and lots of those? |
| 02:18.32 | jasonb | Also, which kernel version are you running? |
| 02:18.45 | jasonb | Do "netstat -an | grep CLOSE_WAIT | wc -l" |
| 02:18.56 | puff | Well *that* was a mistake :-). |
| 02:19.03 | puff | I just tried to run head on every line in lsof :-). |
| 02:19.10 | puff | Let's try that again. |
| 02:20.10 | puff | netstat -an | fgrep CLOSE_WAIT == 37 lines. |
| 02:20.34 | puff | 2.4.21-27.0.2.ELsmp |
| 02:21.55 | jasonb | To some degree, you will likely just be chasing ghosts, since there are lots of bugs like these fixed in 2.6 kernels and not really fixed in 2.4. |
| 02:22.09 | puff | Yeah. |
| 02:22.17 | puff | We've got an upgrade to 2.6 kernel on the slate. |
| 02:22.19 | jasonb | You may be able to track down how many file descriptors Tomcat is using by using lsof.. And even mysql. |
| 02:22.43 | jasonb | But, if it's really malfunctioning, you won't have many options to fix it other than upgrading to kernel 2.6. |
| 02:22.47 | puff | I'm thinking at this point that I really, really want to upgrade. |
| 02:23.07 | puff | Well, if the 2.6 upgrade is likely to help, we'll just step on the gas for that. |
| 02:23.38 | jasonb | Try the lsof | grep java | wc -l both with and without Tomcat running and compare the output. |
| 02:23.41 | puff | Hm, just checked lsof | fgrep -i java | wc again and this time got only 90K. |
| 02:24.07 | jasonb | Well, how could it be 90k when the per process limit would be at most below 65k? |
| 02:25.11 | puff | http://pastebin.com/421327 |
| 02:26.35 | puff | Two tomcat installations. |
| 02:27.09 | puff | Hm. |
| 02:31.23 | puff | jasonb: rE: roll back changes to the HD, one of these language geek friends (who does not hang out at darksleep 7777), working on the ColdStore project, had that idea a few years back as a way to crash-proof windows. Run windows in a VM, snapshot memory to disk every 5 minutes, if windows crashes roll back to 5 minutes. |
| 02:31.37 | puff | Hm, now that I consider it, it's not quite the same idea. |
| 02:32.03 | puff | How would I tell from the output at http://pastebin.com/421327 if it's "really an open file descriptor"? |
| 02:34.54 | puff | USing "ps -ef | fgrep -i java" to find the lowest tomcat process id and then "lsof -p lowesttomcatpid | wc", I get 908 file descriptors. |
| 02:35.37 | puff | Of course, I've already restarted tomcat. Hm. |
| 02:37.25 | jasonb | If you are only running one Java VM on the system, you can just do lsof | grep java | wc -l |
| 02:37.57 | jasonb | I do indeed have the ability to roll back my Win XP hard drive with qemu, and I do indeed do it.. and get away with murder basically. :) |
| 02:38.00 | puff | Two tomcat processes, but yeah. |
| 02:38.04 | puff | Cute. |
| 02:38.08 | puff | Sounds like it'd be handy as hell. |
| 02:38.13 | jasonb | Oh yeah. |
| 02:38.49 | jasonb | Never mind what I said about looking at the lsof output. Each line should really be an open file descriptor. |
| 02:38.51 | puff | I'm intrigued that lsof -p tomcatpid | wc gives me 900 open files, but lsof | fgrep -i java gives me 90,000 |
| 02:39.03 | puff | Hm. |
| 02:39.30 | puff | then again, lsof | fgrep java gives me java-to-mysql connections too. I wonder if those count against java or against mysql. |
| 02:39.48 | jasonb | Well, in kernel 2.4, usually NPTL isn't used, so that means that each line in ps of a Java VM is really a *thread*, not a whole JVM. |
| 02:40.16 | jasonb | So when you look in lsof for the pid of one thread, you see the file descriptors associated just with that one thread of the JVM. :) In theory, anyway. |
| 02:42.33 | puff | Ah, so lsof is getting fooled by the process/thread thing as well? |
| 02:43.16 | puff | Any thoughts from the qemu folks? |
| 02:46.36 | puff | So the better way to get a "true" count is to lsof -p the tomcat process. |
| 02:46.37 | puff | Hm. |
| 02:49.02 | puff | Mysql appears to have the same issue, with respect to showing up in "ps aux" output numerous times. |
| 02:49.15 | puff | But "lsof -p mysqlpid" on each shows the same number of open files (821). |
| 02:49.18 | puff | Hm.... |
| 03:04.04 | jasonb | Nothing from the qemu folks. |
| 03:04.08 | jasonb | they must be out to dinner. |
| 03:04.25 | jasonb | mysql is also multithreaded. |
| 03:04.43 | jasonb | Anyway, I have to head home now. |
| 03:04.50 | jasonb | Have fun. :) |
| 03:07.08 | puff | jasonb: One last question: |
| 03:07.24 | puff | jasonb: The #mysql people are trying to convince me tha tI don't need pooling at all. |
| 03:07.31 | puff | I've actually heard this from a number of credible sources. |
| 03:07.41 | jasonb | What, db connection pooling? |
| 03:07.44 | puff | Is there any quick & dirty way I can disable it? |
| 03:07.44 | puff | Yeah. |
| 03:07.52 | puff | mysql's connection overhead is very, very low. |
| 03:07.58 | jasonb | Well, set the pool to 1 connection. :) |
| 03:08.01 | puff | Hm. |
| 03:08.17 | puff | Heh. |
| 03:08.20 | jasonb | I think it'll just slow down your Tomcat, since everything has to go through one connection. |
| 03:08.25 | puff | I don't think that'll do what I meant :-). |
| 03:08.37 | puff | Okay, I'll see about just ripping out the pool stuff. |
| 03:08.37 | puff | Hm. |
| 03:08.52 | jasonb | I think they're basically wrong about that, JSYK. |
| 03:08.56 | puff | I guess I could make a simple code change that makes it configurable to use JDBC pooling. |
| 03:08.59 | puff | Oh? |
| 03:09.09 | puff | Well, I like pooling for more reasons that saving overhead. |
| 03:09.19 | puff | I like the idea that it will spread the load aorund evenly, in essence. |
| 03:09.22 | jasonb | Your app is already written to access the db using the pool model, and it wouldn't be easy to change that, and even if you did, how would that improve concurrency? |
| 03:09.37 | puff | OR rather, that one app will slow down, rather than hogging it. |
| 03:09.47 | puff | Rather than hogging all the database resources. |
| 03:10.13 | puff | I have a swappable bit of code in my getConnection() method, so I can test stuff without a pool. |
| 03:10.32 | puff | So I'd just have to add some way to invoke that. |
| 03:11.10 | jasonb | (now that i have to go, everyone wants to talk. :) |
| 03:11.15 | puff | Ah? |
| 03:11.15 | puff | H |
| 03:11.16 | puff | Hee. |
| 03:11.25 | jasonb | Ugh. |
| 03:11.35 | lyken | yawn* |
| 03:11.36 | puff | Ugh? |
| 03:11.39 | puff | Oh, yeah. |
| 03:11.48 | jasonb | Good morning lyken! |
| 03:11.54 | puff | lyken: Morning. |
| 03:11.58 | lyken | heya jas, puff |
| 03:13.18 | puff | jasonb: I'll join #qemu so you can go :-). |
| 03:13.53 | jasonb | Okay. Yeah, I have to cruise. |
| 03:13.58 | jasonb | Cheers guys. |
| 03:14.06 | *** join/#tomcat pem3v78 (n=mike@dki63.neoplus.adsl.tpnet.pl) |
| 03:14.22 | pem3v78 | jsvc.exec error: Cannot find daemon loader org/apache/commons/daemon/support/DaemonLoader |
| 03:15.00 | pem3v78 | strange, jsvc runs on one machine and doesn't on the other |
| 03:15.33 | pem3v78 | but it's the same linux distribution with same jvm installed exacly in the same way |
| 03:15.55 | lyken | wow |
| 03:16.00 | pem3v78 | one thing is missing thou, compiler and development programs |
| 03:16.02 | lyken | from poland and you have good english :) |
| 03:16.22 | pem3v78 | :) |
| 03:16.35 | lyken | now uh, i have no idea about that problem |
| 03:16.46 | lyken | all i could do is try and google an answer |
| 03:17.00 | pem3v78 | there is none on google |
| 03:17.05 | pem3v78 | :( |
| 03:18.16 | lyken | ou'd be amazed at hat i can find :) |
| 03:25.44 | lyken | oks |
| 03:25.46 | lyken | found heaps of stuffs |
| 03:28.33 | lyken | is this error when you start up ? |
| 03:30.50 | lyken | you arent running on a fedora /redhat machine with the default jvm from redhat are you ? |
| 03:55.44 | strobedream | hey mang |
| 03:57.00 | lyken | yo |
| 03:57.39 | *** join/#tomcat SpiderX (n=faks@mctnnbsa24w-142167052151.nb.aliant.net) |
| 04:25.36 | *** join/#tomcat FreemaniaX (n=Freemani@pcp0011117215pcs.elkrdg01.md.comcast.net) |
| 04:35.02 | *** join/#tomcat vinse (n=alon@ppp-71-139-111-53.dsl.snfc21.pacbell.net) |
| 04:35.06 | *** join/#tomcat fallacy (i=[U2FsdGV@ool-45777c95.dyn.optonline.net) |
| 04:35.25 | fallacy | what kind of platforms is tomcat embedded designed for? |
| 04:36.26 | lyken | when you want to use tomcat as the engine for a stand alone application? |
| 04:36.40 | fallacy | oh I see |
| 04:37.00 | fallacy | I thought it was just an extra light version to run on some embedded systems |
| 04:38.06 | lyken | lemme check we are speaking the same terminology :) |
| 04:55.59 | lyken | appears my view is right. |
| 04:56.11 | lyken | i dont think tomcat could be made small enough to be embedded. |
| 04:58.10 | fallacy | oh ok |
| 04:58.13 | fallacy | just wondering |
| 05:01.34 | lyken | well thats from what i just read |
| 05:01.35 | lyken | i could be wrong |
| 05:01.36 | lyken | heh |
| 05:34.29 | *** join/#tomcat Zangdar`Away (n=jordimcl@ASt-Lambert-109-1-8-73.w193-252.abo.wanadoo.fr) |
| 05:43.24 | *** join/#tomcat jasonb (i=noneoyer@adsl-66-124-73-250.dsl.sntc01.pacbell.net) |
| 06:03.21 | lyken | wb |
| 06:12.08 | *** join/#tomcat Zangdar`Away (n=jordimcl@ASt-Lambert-109-1-8-73.w193-252.abo.wanadoo.fr) [NETSPLIT VICTIM] |
| 06:12.08 | *** join/#tomcat vinse (n=alon@ppp-71-139-111-53.dsl.snfc21.pacbell.net) [NETSPLIT VICTIM] |
| 06:12.08 | *** join/#tomcat Hugh (n=Hugh@base-aaisp.mxtelecom.com) [NETSPLIT VICTIM] |
| 06:12.08 | *** join/#tomcat Vanuatoo (n=vanob@227-35.adsl.caucasus.net) [NETSPLIT VICTIM] |
| 06:12.09 | *** join/#tomcat pucko (n=andreas@h232n2fls31o809.telia.com) [NETSPLIT VICTIM] |
| 06:12.09 | *** join/#tomcat lilalinux (i=e-trolle@deepthroat.deswahnsinns.de) [NETSPLIT VICTIM] |
| 06:12.09 | *** join/#tomcat strobedream (n=td@pcp05454736pcs.sandia01.nm.comcast.net) [NETSPLIT VICTIM] |
| 06:12.09 | *** join/#tomcat KingB (n=bking@208.246.215.5) [NETSPLIT VICTIM] |
| 06:12.09 | *** join/#tomcat vito__ (n=vito@www.perspectivepartners.com) [NETSPLIT VICTIM] |
| 06:12.09 | *** join/#tomcat roel- (n=rdvdijk@powderfinger.rustradio.org) [NETSPLIT VICTIM] |
| 06:12.09 | *** join/#tomcat Jara[zZ] (n=rimaraj@62.123.204.106) [NETSPLIT VICTIM] |
| 06:12.09 | *** join/#tomcat Yopyop (i=plop@83.173.67.7) [NETSPLIT VICTIM] |
| 06:12.09 | *** join/#tomcat HighTek (n=mukdog@Corp-NAT.connectria.com) [NETSPLIT VICTIM] |
| 06:12.09 | *** join/#tomcat memic (n=memic@217.172.179.89) [NETSPLIT VICTIM] |
| 06:24.02 | *** join/#tomcat Zangdar`Away (n=jordimcl@ASt-Lambert-109-1-8-73.w193-252.abo.wanadoo.fr) [NETSPLIT VICTIM] |
| 06:24.03 | *** join/#tomcat vinse (n=alon@ppp-71-139-111-53.dsl.snfc21.pacbell.net) [NETSPLIT VICTIM] |
| 06:24.03 | *** join/#tomcat Hugh (n=Hugh@base-aaisp.mxtelecom.com) [NETSPLIT VICTIM] |
| 06:24.03 | *** join/#tomcat Vanuatoo (n=vanob@227-35.adsl.caucasus.net) [NETSPLIT VICTIM] |
| 06:24.03 | *** join/#tomcat pucko (n=andreas@h232n2fls31o809.telia.com) [NETSPLIT VICTIM] |
| 06:24.03 | *** join/#tomcat lilalinux (i=e-trolle@deepthroat.deswahnsinns.de) [NETSPLIT VICTIM] |
| 06:24.03 | *** join/#tomcat strobedream (n=td@pcp05454736pcs.sandia01.nm.comcast.net) [NETSPLIT VICTIM] |
| 06:24.03 | *** join/#tomcat KingB (n=bking@208.246.215.5) [NETSPLIT VICTIM] |
| 06:24.03 | *** join/#tomcat vito__ (n=vito@www.perspectivepartners.com) [NETSPLIT VICTIM] |
| 06:24.03 | *** join/#tomcat memic (n=memic@217.172.179.89) [NETSPLIT VICTIM] |
| 06:24.03 | *** join/#tomcat HighTek (n=mukdog@Corp-NAT.connectria.com) [NETSPLIT VICTIM] |
| 06:24.03 | *** join/#tomcat Yopyop (i=plop@83.173.67.7) [NETSPLIT VICTIM] |
| 06:24.04 | *** join/#tomcat Jara[zZ] (n=rimaraj@62.123.204.106) [NETSPLIT VICTIM] |
| 06:24.04 | *** join/#tomcat roel- (n=rdvdijk@powderfinger.rustradio.org) [NETSPLIT VICTIM] |
| 06:31.18 | *** join/#tomcat vinse (n=alon@ppp-71-139-111-53.dsl.snfc21.pacbell.net) |
| 07:02.10 | puff | lyken: I think you're about right. |
| 07:05.23 | lyken | cool i thought so |
| 07:05.39 | lyken | based on the code ive found around and what i know about it, its like tomcat embedded into like netbeans and stuff |
| 07:05.46 | lyken | like embedded an sql engine inside an application |
| 07:47.18 | *** join/#tomcat Zarco (n=b@81.241.234.166) |
| 07:52.49 | Zarco | HI, I'm quite e newbie whenit comes to tomcat, but I have to following problem. We had a tomcat running well, but due to a power outage, the box wasn't properly shut down. When we rebooted the box, the tomcat does not startup. When I start tomcat and go to localhost:8080 nothing happens, not an error that no service is running on that port, not a tomcat-screen. Can it be that there is a stale file around preventing tomcat to startup? |
| 08:23.03 | *** join/#tomcat wfay (n=ask@adsl-068-209-227-025.sip.mem.bellsouth.net) |
| 08:29.36 | *** join/#tomcat nevs27 (i=lino@82.211.36.253) |
| 08:37.01 | *** join/#tomcat THOMAS (n=Thomas@p54B3B77C.dip0.t-ipconnect.de) |
| 08:37.10 | THOMAS | Hi! |
| 08:37.25 | THOMAS | Kann mir jmd. mit der installation von tomcat helfen? |
| 08:41.13 | THOMAS | ?? |
| 08:47.01 | *** part/#tomcat Zarco (n=b@81.241.234.166) |
| 09:06.28 | *** part/#tomcat THOMAS (n=Thomas@p54B3B77C.dip0.t-ipconnect.de) |
| 09:32.03 | *** join/#tomcat noodl (i=hidden-u@195-97-225-122.onyx.net) |
| 09:53.26 | *** join/#tomcat Conductor (n=thomas@62.8.240.185) |
| 09:53.39 | Conductor | how do i redirect all errors to a single page? |
| 09:54.24 | Conductor | <error-code>*</error-code> |
| 09:54.26 | Conductor | ? |
| 10:32.53 | *** join/#tomcat dreivier (n=dreivier@port-212-202-44-204.dynamic.qsc.de) |
| 11:45.05 | Conductor | how do i redirect all errors to a single page? |
| 11:45.07 | Conductor | <error-code>*</error-code> |
| 11:45.08 | Conductor | ? |
| 11:57.56 | *** join/#tomcat shuveb (n=shuveb@BTNL-TN-DSL-static-133.1.144.59.touchtelindia.net) |
| 11:58.14 | *** part/#tomcat shuveb (n=shuveb@BTNL-TN-DSL-static-133.1.144.59.touchtelindia.net) |
| 12:16.22 | *** join/#tomcat StarScream (n=am@217.169.25.114) |
| 12:23.12 | *** join/#tomcat silverwolf (n=silverwo@x1-6-00-d0-68-06-20-94.k592.webspeed.dk) |
| 12:33.08 | StarScream | is there a way to configure a differnent logfile for each <host> ? |
| 13:24.27 | *** join/#tomcat StarScream (n=am@217.169.25.114) [NETSPLIT VICTIM] |
| 13:24.27 | *** join/#tomcat Conductor (n=thomas@62.8.240.185) [NETSPLIT VICTIM] |
| 13:24.27 | *** join/#tomcat nevs27 (i=lino@82.211.36.253) |
| 13:24.27 | *** join/#tomcat Zangdar`Away (n=jordimcl@ASt-Lambert-109-1-8-73.w193-252.abo.wanadoo.fr) [NETSPLIT VICTIM] |
| 13:24.27 | *** join/#tomcat Hugh (n=Hugh@base-aaisp.mxtelecom.com) [NETSPLIT VICTIM] |
| 13:24.28 | *** join/#tomcat Vanuatoo (n=vanob@227-35.adsl.caucasus.net) [NETSPLIT VICTIM] |
| 13:24.28 | *** join/#tomcat pucko (n=andreas@h232n2fls31o809.telia.com) [NETSPLIT VICTIM] |
| 13:24.29 | *** join/#tomcat lilalinux (i=e-trolle@deepthroat.deswahnsinns.de) [NETSPLIT VICTIM] |
| 13:24.29 | *** join/#tomcat strobedream (n=td@pcp05454736pcs.sandia01.nm.comcast.net) [NETSPLIT VICTIM] |
| 13:24.29 | *** join/#tomcat KingB (n=bking@208.246.215.5) [NETSPLIT VICTIM] |
| 13:24.29 | *** join/#tomcat vito__ (n=vito@www.perspectivepartners.com) [NETSPLIT VICTIM] |
| 13:24.29 | *** join/#tomcat roel- (n=rdvdijk@powderfinger.rustradio.org) [NETSPLIT VICTIM] |
| 13:24.29 | *** join/#tomcat Jara[zZ] (n=rimaraj@62.123.204.106) [NETSPLIT VICTIM] |
| 13:24.29 | *** join/#tomcat Yopyop (i=plop@83.173.67.7) [NETSPLIT VICTIM] |
| 13:24.30 | *** join/#tomcat HighTek (n=mukdog@Corp-NAT.connectria.com) [NETSPLIT VICTIM] |
| 13:24.30 | *** join/#tomcat memic (n=memic@217.172.179.89) [NETSPLIT VICTIM] |
| 13:35.55 | *** join/#tomcat FreemaniaX (n=Freemani@pcp0011117215pcs.elkrdg01.md.comcast.net) |
| 13:44.19 | StarScream | helloo anyone here? |
| 13:44.37 | StarScream | has freenode recovered yet |
| 15:14.10 | *** join/#tomcat Zangdar (n=jordimcl@ASt-Lambert-109-1-6-237.w193-252.abo.wanadoo.fr) |
| 16:33.34 | *** join/#tomcat dd (i=deeeed@m108.net195-132-184.noos.fr) |
| 16:33.37 | dd | hi |
| 16:38.16 | dd | is it possible to set defaultHost="*" in server.xml ? (<Engine defaultHost="localhost" name="Catalina">) |
| 16:38.43 | kibab | What would that accomplish? |
| 16:39.10 | dd | i would like to do http://localhost and http://myip |
| 16:39.34 | dd | i would like that to work for the both. |
| 16:39.40 | dd | do you understand ? |
| 16:40.23 | kibab | I think anything unidentified by an explicit host entry will go to the default host, so you should be able to set the default to whatever you like. |
| 16:44.55 | dd | ah yes ;) |
| 16:44.59 | dd | sorry for the questions ;) |
| 16:45.10 | dd | thx |
| 16:47.12 | *** join/#tomcat FreemaniaPrime (n=Freemani@ip67-153-106-58.z106-153-67.customer.algx.net) |
| 16:59.18 | kibab | Hey all... mod_jk isn't reading the workers.properties file for some reason. Details at http://pastebin.com/421869. Any ideas? |
| 17:02.08 | kibab | To summarize, I have confirmed with strace that JkWorkersFile is stat'd but not open'd, which causes the mapping to fail as it can't find any details on the workers. |
| 17:02.15 | *** join/#tomcat valmont (n=chrishol@dsl092-043-004.lax1.dsl.speakeasy.net) |
| 17:04.53 | *** join/#tomcat johnmoore330 (n=johnm@mailgw.pdsi-software.com) |
| 17:06.02 | kibab | I added my workers.properties file contents to this paste: http://pastebin.com/421881 |
| 17:14.04 | *** join/#tomcat jasonb (i=noneoyer@adsl-66-124-73-250.dsl.sntc01.pacbell.net) |
| 17:19.02 | noodl | Hi guys. I'm having a rough time getting jdbc wired in to tomcat. This is my WEB-INF/context.xml file. http://rafb.net/paste/results/UkcIaK59.html Is there anything clearly wrong here? |
| 17:20.48 | kibab | noodl: what's the error you're getting when you try to access it? |
| 17:21.06 | noodl | javax.servlet.ServletException: Cannot create JDBC driver of class '' for connect URL 'null' |
| 17:21.47 | noodl | I'm very new to this, so please feel free to take the piss. |
| 17:22.27 | noodl | This is the code used to access the named resource: http://rafb.net/paste/results/ZmbIln35.html |
| 17:22.56 | noodl | And my WEB-INF/web.xml contains this: http://rafb.net/paste/results/6reToH35.html |
| 17:24.23 | noodl | kibab: What OS/Distro are you running mod_jk on? |
| 17:24.43 | kibab | Tomcat isn't picking it up right. I have only declared a db globally in server.xml... in which case I have the same code that you do in the context.xml file. But then in my Context definition I include a ResourceLink identifier which allows my app to see it. |
| 17:24.52 | kibab | noodl: Redhat Enterprise Linux |
| 17:25.27 | noodl | EL{3,4}? |
| 17:25.41 | kibab | noodl: 3... manually compiled the latest version. |
| 17:26.04 | noodl | Oh, I was thinking perhaps selinux, but 3 doesn't usually have that. |
| 17:26.23 | kibab | noodl: nope.. it's not running it. |
| 17:27.00 | noodl | Ok, so for the moment I don't care too much if I use server.xml or WEB-INF/context.xml to declare the database configuration parameters. I just want to get them wired together. |
| 17:27.11 | noodl | I've been trying to sort this out all day :( |
| 17:27.39 | noodl | kibab: Would you be prepared to show me your working config (password removed etc.)? |
| 17:32.21 | kibab | noodl: this is what I use: http://pastebin.com/421911 |
| 17:33.50 | noodl | kibab: Thanks, I'll try those out. Is your <resource-ref> (line 12) part of web.xml? |
| 17:34.18 | kibab | noodl: differences: I don't use res-sharing-scope, mine is global across apps, I grab the context using two lines instead of one, and that's about it. |
| 17:34.20 | kibab | noodl: yes. |
| 17:41.37 | kibab | noodl: not sure why it would make a difference... but does it matter if you change your code that reads in the entry to read in the env context before the ds context? (just noticing that I have never seen it read in one line before in any examples I have seen) |
| 17:42.03 | noodl | I've just tried that, and it did change the exception. |
| 17:42.07 | noodl | didn't |
| 17:42.28 | kibab | noodl: k. |
| 17:42.29 | noodl | I don't think it's the jndi lookup that has a problem. |
| 17:42.47 | kibab | It's that Tomcat isn't finding the Resource you have declared... the question is why. |
| 17:43.03 | kibab | I have that same problem if I mess up the resourcelink when using the global resource. |
| 17:43.06 | noodl | (If I falsify the jndi name, the lookup barfs as appropriate) |
| 17:43.20 | kibab | noodl: k. good simple test |
| 17:43.27 | FreemaniaPrime | does the manager app rely on the 'admin' app? |
| 17:43.36 | FreemaniaPrime | I want to get rid of my admin app |
| 17:43.38 | noodl | In the admin webapp, I can see the data source listed for my webapp. |
| 17:43.41 | kibab | FreemaniaPrime: no. |
| 17:44.05 | noodl | But, all the field (connection url, driver class) are empty. |
| 17:44.18 | kibab | noodl: hmm... let me see what mine shows |
| 17:47.08 | noodl | Oh, I'm testing this with netbeans' tomcat btw (perhaps I should have said that earlier). |
| 17:49.45 | *** join/#tomcat dreivier (n=dreivier@port-212-202-44-204.dynamic.qsc.de) |
| 17:50.01 | noodl | kibab: In your paste you have "# in my app context" (line 72). Do you mean WEB-INF/context.xml? |
| 17:51.03 | kibab | noodl: it's in a different location (not WEB-INF/context.xml) but yes. |
| 17:52.30 | noodl | Oh, you mean the <Context> element of server.xml for your app? |
| 17:52.55 | kibab | noodl: in the admin app, all the resource params that I specified are available |
| 17:53.00 | noodl | Root cause: org.apache.tomcat.dbcp.dbcp.SQLNestedException: Cannot create JDBC driver of class '' for connect URL 'null' |
| 17:53.13 | kibab | noodl: yeah... never used WEB-INF/context.xml |
| 17:53.44 | noodl | Hmm, perhaps that's my problem. I'll move this away from netbeans and try fiddling with standalone tomcat's server.xml. |
| 17:54.06 | kibab | k. |
| 18:01.46 | noodl | whaaaa! |
| 18:01.57 | noodl | no joy, time to go home :( |
| 18:02.09 | noodl | Thanks anyway kibab. Good luck fixing mod_jk. |
| 19:15.50 | *** join/#tomcat valmont (n=chrishol@germany.pspdev.pas.earthlink.net) |
| 19:38.14 | *** join/#tomcat linuxfreck (n=freck@fry.fmi.uni-passau.de) |
| 20:00.47 | *** join/#tomcat leonel (n=leonel@201.153.63.251) |
| 20:45.15 | *** join/#tomcat yel (n=yel@xdsl-84-44-176-237.netcologne.de) |
| 20:45.21 | yel | hi everyone! |
| 20:45.40 | linuxfreck | hi yel |
| 20:45.53 | yel | hi linuxfreck :) |
| 21:03.56 | FreeManiaX | Tomcat 5.5 hates my Realm |
| 21:05.05 | FreeManiaX | I recompiled it , repackaged it . But besides setting the attribute/properties , it will not authenticate |
| 21:05.34 | FreeManiaX | I'd love to see an example of the right way to make a custom realm |
| 21:07.50 | yel | which realm do you have FreeManiaX |
| 21:16.38 | *** join/#tomcat snibril (n=jh@p54A3E1E7.dip.t-dialin.net) |
| 21:17.21 | jasonb | FreeManiaX: Writing a custom realm really does mean you are agreeing to debug it when it doesn't work. :) |
| 21:51.20 | *** join/#tomcat peterb (n=peterb@c-67-162-138-236.hsd1.co.comcast.net) |
| 21:52.46 | *** join/#tomcat ucarpeterb (n=peterb@vpn13.ucar.edu) |
| 21:55.46 | *** join/#tomcat pem3v78 (n=pem3v78@dkr143.neoplus.adsl.tpnet.pl) |
| 21:56.04 | pem3v78 | I'm using tomcat 5.5.12 on debian. |
| 21:56.27 | pem3v78 | when run on one system jsvc works ok, but on the other I have 'Cannot find daemon loader org/apache/commons/daemon/support/DaemonLoader' |
| 21:56.41 | pem3v78 | java and tomcat installation is exaclly the same |
| 21:56.55 | pem3v78 | maybe enrionment is diffrent |
| 21:57.12 | pem3v78 | is this a known feature to occur ? |
| 21:58.30 | pem3v78 | and jdk1.5.0_05 |
| 22:03.11 | *** join/#tomcat ucarpeterb (n=peterb@vpn13.ucar.edu) |
| 22:03.35 | *** join/#tomcat pchapman (n=pchapman@misc-148-65-81-58.pool.starband.net) |
| 22:04.10 | ucarpeterb | I hope someone here can help me with an AJP1.3 question on tomcat 5.5.12 |
| 22:05.53 | ucarpeterb | One of the development teams here has redesigned an application, and it now runs on 5.5.12 and Java 1.5 (Sun). I"m deploying under RHEL 3. |
| 22:05.53 | ucarpeterb | The HTTP connector works great. But the AJP connector to the front-end Apache (1.3): not so good. |
| 22:08.01 | ucarpeterb | Running tethereal, when I do a GET on the Apache server, I see the 3-way handshake between the apache server and the tomcat server, then the Apache server sends a PSH, ACK packet with the request, and ACK comes back, then nothing. |
| 22:08.44 | pchapman | I'm having some problems with tomcat and apache 2.0. The config is at http://rafb.net/paste/ |
| 22:09.16 | pchapman | I can telnet to localhost 8081, but apache reports that it cannot connect to 127.0.0.1:8081 |
| 22:09.24 | yel | pchapman, there is nothing there :) |
| 22:09.37 | pchapman | oops |
| 22:09.43 | pchapman | http://rafb.net/paste/results/7sXcUv29.html |
| 22:13.52 | *** join/#tomcat ucarpeterb (n=peterb@vpn13.ucar.edu) |
| 22:16.16 | pchapman | yel, Do you see any mistakes in the config? |
| 22:17.26 | pchapman | There error I see in apache's logs is http://rafb.net/paste/results/mqYHXx75.html |
| 22:17.47 | yel | i will have a look at it now |
| 22:18.43 | yel | seems like if its refused the connection their are you sure its listinig to that port ? |
| 22:18.56 | pchapman | Is there some reason why the apache user wouldn't be able to make the connection |
| 22:19.14 | *** join/#tomcat UCARpeterb (n=chatzill@vpn13.ucar.edu) |
| 22:19.59 | pchapman | I can telnet to 127.0.0.1 8081 from that machine's console. |
| 22:20.16 | pchapman | It closes pretty fast though. |
| 22:21.38 | yel | if tomcat is not serving is the connection refused or you do connect and than the connection get closed ? |
| 22:22.01 | pchapman | I do connect, then after a short wait, the connection is closed. |
| 22:28.13 | UCARpeterb | Hello -- I hope someone can offer me some clues on a AJP1.3 issue (and if I'm repeating myself; sorry, my other client crapped on me) |
| 22:28.59 | UCARpeterb | I have a refactored application from one of my development teams, and I'm trying to do a test deployment. |
| 22:29.43 | UCARpeterb | Everything goes smoothly with the HTTP connector. But the AJP connector from Apache 1.3 to Tomcat 5.5, not so good. |
| 22:31.16 | UCARpeterb | When I do my GET on the Apache side, tethereal shows the 3-way handshake between Apache on Tomcat, then Apache sends a PSH,ACK packet with the request, the Tomcat sends an ACK, then nothing. |
| 22:31.30 | UCARpeterb | I don't know AJP well enough to know what a normal connection looks like. |
| 22:32.16 | snibril | well, consider using apache 2.0, that works |
| 22:32.21 | UCARpeterb | Details: tomcat server is RHEL 3.0 on x86; Tomcat 5.5.12, Sun Java 1.5.05; apache is 1.3.X, mod_jk version unknown, also on RHEL3 |
| 22:33.25 | UCARpeterb | This works just great with a bunch of other apps running tomcat 5.0.X and the Apache 1.3. apache 2.0 isn't an immediate option. |
| 22:34.21 | *** part/#tomcat vinse_work (n=Administ@208.253.223.146) |
| 22:34.35 | *** join/#tomcat vinse_work (n=Administ@208.253.223.146) |
| 22:34.54 | UCARpeterb | Hmm, but I have an Apache 2.0 server elsewhere that I can test w/ and see if problem goes away... |
| 22:36.50 | snibril | btw, the symptoms your describe would appear if you activated the shared secret in ajp |
| 22:36.57 | snibril | and forgot to config it on both sides |
| 22:37.47 | puff | Evening. |
| 22:39.46 | yel | pchapman, sorry i can not see what the your problem is ! |
| 22:39.51 | yel | hi puff :) |
| 22:40.35 | UCARpeterb | shared secret -- hmm -- very interesting -- I'll keep that in mind as I poke around. Thx. |
| 22:45.08 | puff | snibril: 2.0 or 2.1? I thought 2.1 was the version with the new-and-improved mod_proxy that incorporates AJP. |
| 22:45.48 | pchapman | yel, thanks for the effort. |
| 22:46.21 | pchapman | I'm rather at a loss myself. |
| 22:47.16 | UCARpeterb | Problem solved -- nothing to do with AJP -- but without someplace to bounce this question I don't know how long I"d be stuck. The answer |
| 22:48.50 | snibril | puff: we use ap2 + mod_jk + 5.5.9 |
| 22:48.56 | snibril | and that works |
| 22:49.32 | UCARpeterb | seems to be an application error. A query that works to the HTTP side simply doesn't work to the AJP side |
| 22:49.44 | pchapman | is mod_jk better than mod_proxy? Why 2 different way to proxy? |
| 22:50.16 | snibril | pchapman: because mod_proxy only lately learned to speak ajp |
| 22:51.09 | pchapman | I'm having trouble getting mod_proxy to work. Maybe mod_jk is my path.... |
| 22:52.33 | *** join/#tomcat pem3v78 (n=mike@dkr143.neoplus.adsl.tpnet.pl) |
| 22:57.09 | *** join/#tomcat jasonb (i=noneoyer@adsl-66-124-73-250.dsl.sntc01.pacbell.net) |
| 22:57.40 | *** join/#tomcat jasonb (i=noneoyer@adsl-66-124-73-250.dsl.sntc01.pacbell.net) |
| 23:04.01 | vinse_work | jasonb! |
| 23:04.02 | vinse_work | hola |
| 23:05.04 | pem3v78 | finally got jsvc to work, but jdk and tomcat need to be in user home dir, why is that ? |
| 23:05.11 | UCARpeterb | Now that I'm out of the woods, does anyone know why 5.5. has both /host-manager and /manager?? |
| 23:05.47 | pem3v78 | when I move tomcat outside user dir classes from commons-daemon.jar are not found |
| 23:06.54 | pem3v78 | where is the proper help for this server ? |
| 23:07.04 | pem3v78 | this channel seems to be pretty dead |
| 23:08.51 | puff | snibril: Is mod_proxy standard in apache? |
| 23:08.53 | snibril | uca: manager manages webapss |
| 23:09.01 | UCARpeterb | Dead? Actually, I've found it pretty helpful. but it can be SLOW-MOTION |
| 23:09.02 | snibril | host-manager manages vhosts |
| 23:09.16 | snibril | which the topic states clearly |
| 23:10.06 | FreeManiaX | jasonb, you are right |
| 23:10.20 | FreeManiaX | about debugging a custom realm |
| 23:12.51 | UCARpeterb | Which the topic states clearly but nothing in the documentation. I guess I should endeavour to help the community, read the code, and write the documentation. |
| 23:12.59 | pem3v78 | ah, lost whole day to figure out that jsvc needs commons-daemon inside user dir, shit this could be written in INSTALL or README at least |
| 23:13.05 | jasonb | FreeManiaX: Yeah. I've spent enough time with the default implementations that come with Tomcat to know this is the case. Custom realms are even harder since you can't just search the mailing lists to see if other people are having success with it or not. :) |
| 23:13.11 | jasonb | vinse_work: Hi. How's it going? |
| 23:13.46 | jasonb | puff: Yes, mod_proxy is standard in Apache. |
| 23:14.13 | *** join/#tomcat algebra (n=yel@xdsl-84-44-155-74.netcologne.de) |
| 23:19.36 | *** join/#tomcat andydan (i=andydan@159-134-176-54.as1.cra.dublin.eircom.net) |
| 23:23.38 | FreeManiaX | it goes jasonb |
| 23:24.18 | FreeManiaX | I'm serioussly thinking about using a security filter for the app instead of using a Realm |
| 23:24.33 | FreeManiaX | have to reboot bbl |
| 23:27.24 | *** join/#tomcat snibri1 (n=jh@p54A3E1FA.dip.t-dialin.net) |
| 23:30.01 | *** join/#tomcat FreemaniaX (n=Freemani@ip67-153-106-58.z106-153-67.customer.algx.net) |
| 23:31.09 | puff | jasonb: Ah, cool, so when 2.1 is mainstream, apache/tomcat integration should be a lot simpler to set up. |
| 23:31.26 | puff | FreemaniaX: I highly recommend securityfilter instead of a realm. |
| 23:32.05 | puff | FreemaniaX: We're using it quite nicely, though I'm thinking about cracking open the code to add in some hooks or something, some way for us to specify things that should happen right after the user logs in, not to mention some easier way to get at the username. |
| 23:33.03 | puff | Bah. |
| 23:33.11 | puff | Whoops, 6:30pm, time for dinner. See ya all in a bit. |
| 23:35.33 | jasonb | Yes, Filters are a newer construct than Valves, and Valves were the precurser that made people want to add that functionality to the servlet spec. So, yes, it's usually elegant enough for a security filter or even custom authentication, and it's servlet container implementation independent. |
| 23:36.17 | jasonb | But, Valves are also closer to where the request comes into the container (Tomcat), and some things are best done as valves, mainly for performance. |
| 23:36.32 | jasonb | So it looks as if valves aren't really going away. |
| 23:38.17 | jasonb | puff: Yes, when Apache 2.1 is mainstream, just about everyone will recommend using mod_proxy since it knows AJP. But, also, there are some load balancing features that look like they will only work when you use mod_jk.. Unless you use a "real" (hardware) load balancer, or LVS, that is. |
| 23:38.51 | jasonb | puff: Generally (hopefully), lots of people will end up switching from mod_jk to mod_proxy, or from mod_jk to plain old Tomcat. :) |
| 23:39.13 | FreemaniaX | is it true that the latest tomcat has some optional native parts ? |
| 23:39.45 | FreemaniaX | I'm a plain ole Tomcat guy |
| 23:40.08 | FreemaniaX | hooking to mod_X is just too much hassle |
| 23:40.24 | jasonb | FreemaniaX: There are two new connectors that are based on native code (APR code) for performance. |
| 23:40.30 | jasonb | FreemaniaX: And, yes, those are optional. |
| 23:41.28 | jasonb | FreemaniaX: But, from everything I could find about it in the mailing list, it turns out that the native APR web server connector is still about the same speed as the pure-Java coyote web server connector. |
| 23:41.55 | jasonb | FreemaniaX: So, I'm not sure why anyone would go that far out of their way to use the native one in that case. There aren't many reasons. |
| 23:46.42 | FreemaniaX | I agree jasonb |
| 23:49.57 | FreemaniaX | damn, SecurityFilter does not support single sign on between contexts |
| 23:50.55 | yel | im working on an idea to realize single sign on withs securityfilter FreemaniaX |
| 23:51.28 | FreemaniaX | cool |
| 23:51.31 | yel | well i have to deal with tomcat and apache2 |
| 23:51.38 | FreemaniaX | ah |
| 23:51.45 | yel | but the aprocach can be aplied for anyother apps |
| 23:52.32 | yel | my main idea is to have two tables the first one conains the real user data and theire passwords (md5) |
| 23:53.35 | yel | the second table is just a temporary one wich mean when a user autheticate on the first table he will gets his username+password in the cookie just tha this password he gets is a faked one |
| 23:53.57 | yel | and this password and his username will be stored in the second table |
| 23:54.33 | yel | so other application can check that as long as the HttpSessionListner can track the user and there is no timeout event |
| 23:54.48 | yel | the user cann access all the resouce you specified |
| 23:55.44 | yel | when a timeout occure the faked password and the user name will be deleted from the second table so no apps is able ti authenticate the user anymore |
| 23:56.01 | yel | FreemaniaX, can you seen what i mean : ) ? |