| 00:38.34 | puff | yel: I have no idea. |
| 00:38.46 | yel | danke |
| 00:38.51 | yel | oups thanks puff :) |
| 02:23.34 | *** join/#tomcat FreemaniaX (n=Freemani@pcp0011117215pcs.elkrdg01.md.comcast.net) |
| 02:55.27 | *** join/#tomcat CAz (n=me@wnpgmb02dc1-62-52.dynamic.mts.net) |
| 02:56.20 | CAz | hi all working on apache/tomcat on solaris 10 nothing fancy right off the box install but don't know how to start tomcat |
| 03:16.30 | *** join/#tomcat Folgore (n=noone@203.125.84.22) |
| 03:16.59 | Folgore | hi tomcat still requires the original apache installed right? it is not a standalone? |
| 03:17.00 | *** join/#tomcat b00t (n=ss@222.178.152.81) |
| 03:17.15 | CAz | hi Folgore |
| 03:17.42 | Folgore | hi |
| 03:17.47 | CAz | you can run it stand-alone to my knowledge just is better to go with apache (to my knowledge not an expert) |
| 03:17.58 | CAz | especially if for dev purposes |
| 03:18.21 | Folgore | oh i see thanks |
| 03:18.31 | *** part/#tomcat yel (n=nnnnnnye@xdsl-87-78-33-160.netcologne.de) |
| 03:18.46 | CAz | just might have to go to 8080 to see results instead of 80 |
| 03:19.47 | *** join/#tomcat bblack (n=bblack@h216.70.13.4.superiorbroadband.com) |
| 03:19.51 | Folgore | thanks alot |
| 03:20.10 | CAz | no problem |
| 03:20.15 | CAz | what os you installing this on ? |
| 03:22.02 | Folgore | i have a solaris and windows box |
| 03:22.06 | Folgore | prob going to try on both |
| 03:22.14 | CAz | we should stay in contact |
| 03:22.22 | CAz | have gotten it working on windows |
| 03:22.30 | CAz | and am now trying to do it on solaris 10 |
| 03:22.38 | CAz | but ran into a wall |
| 03:22.48 | Folgore | woah I'm still using 8 |
| 03:22.50 | Folgore | lol |
| 03:23.08 | CAz | this is at home for dev/admin learning purposes so i don't have any limitations |
| 03:23.28 | CAz | would be nice if i could find some docs on tomcat seeing as how the bundled it with solaris 10... |
| 03:25.16 | Folgore | well the documents on the tomcat homepage seem straight forward , will try it out later |
| 03:25.29 | CAz | that's where i am looking right now |
| 03:25.48 | CAz | just annoying that it's in solaris 10 and they don't even say hey look at the tomcat homepage.. |
| 03:25.56 | bblack | Anyone know how to allow cgi style PATH_INFO data via a .jsp page? Had no luck setting the .jsp servlet's <url-pattern> to *.jsp* (returned the jsp page as plain/text) and *.jsp/* (Invalid pattern) |
| 03:26.23 | CAz | woot there are other people alive in here |
| 03:31.18 | lyken | is not ! |
| 03:37.59 | bblack | anyone know if the invoker servlet can be used allow passing of PATH_INFO style data to jsp pages? |
| 04:02.57 | *** join/#tomcat omay (n=Oleg@dslb-088-072-041-091.pools.arcor-ip.net) |
| 04:09.04 | CAz | hey omay |
| 04:09.09 | lyken | *yawn* |
| 04:11.06 | CAz | hey lyken |
| 04:11.35 | CAz | you one of the knowers of tomcat around here or one of the people seeking knowledge ? |
| 04:27.03 | FreemaniaX | question about classloading issues: I have a class and an xml file both in the same jar. the class needs to load this xml file. but I keep getting an IOException |
| 04:27.49 | FreemaniaX | it cant find the file |
| 04:28.23 | FreemaniaX | Reader reader = Resources.getResourceAsReader("jaassqlmapconfig.xml"); |
| 04:36.30 | *** join/#tomcat blumber (n=klumber@adsl-69-232-203-172.dsl.pltn13.pacbell.net) |
| 04:38.21 | blumber | anyone around? where do you put the Listener to listen in on users public_html? |
| 04:38.57 | blumber | I enabled this on conf/server.xml below <Host >... |
| 04:39.04 | blumber | but it was throwing an error |
| 04:40.01 | blumber | nm my fault..somehow i missed the " |
| 05:33.32 | *** join/#tomcat jasonb (i=noneoyer@adsl-66-124-73-250.dsl.sntc01.pacbell.net) |
| 05:45.09 | FreemaniaX | hey jasonb |
| 05:58.59 | *** join/#tomcat NamNT (n=namnt@203.117.170.20) |
| 05:59.03 | NamNT | hi |
| 05:59.30 | NamNT | is anyone running tomcat 5.5.14? |
| 05:59.56 | NamNT | or latest development version |
| 06:09.39 | jasonb | Hi FreemaniaX |
| 07:13.06 | *** join/#tomcat FreemaniaX (n=Freemani@pcp0011117215pcs.elkrdg01.md.comcast.net) |
| 07:17.00 | *** join/#tomcat dorileo (n=dorileo@200.175.149.184.adsl.gvt.net.br) |
| 07:27.17 | *** part/#tomcat dorileo (n=dorileo@200.175.149.184.adsl.gvt.net.br) |
| 08:46.12 | *** join/#tomcat GedasRx (i=usr492@c1.edrana.lt) |
| 09:14.28 | *** join/#tomcat Twilight\ (n=ask@17.84-48-138.nextgentel.com) |
| 09:36.48 | *** join/#tomcat silverwolf (i=silverwo@83.93.43.159) |
| 10:11.46 | *** join/#tomcat AvengerMoJo (n=alex@avenger.brainfood.com) |
| 10:11.49 | AvengerMoJo | hi |
| 10:12.08 | AvengerMoJo | is Tomcat possible after start up , force it to update the DNS server ip? |
| 10:12.44 | AvengerMoJo | e.g. resolve used to be 192.168.0.1 ... not it is down ... and DNS switch to 192.168.0.2 .... |
| 10:17.28 | *** part/#tomcat vinse_ (n=vinse_@ppp-71-139-106-234.dsl.snfc21.pacbell.net) |
| 10:18.04 | *** join/#tomcat aicra (n=aicra@ip70-190-89-221.ph.ph.cox.net) |
| 10:18.31 | *** join/#tomcat vinse_ (n=vinse_@ppp-71-139-106-234.dsl.snfc21.pacbell.net) |
| 10:22.08 | *** join/#tomcat SubNova81 (n=SubNova8@ny-lackawna-cadent1-grp2e-b-241.bflony.adelphia.net) |
| 10:24.22 | *** join/#tomcat SubNova81 (n=SubNova8@ny-lackawna-cadent1-grp2e-b-241.bflony.adelphia.net) |
| 10:25.10 | SubNova81 | hi, i have problem with java Web services DP 1.6 finding tomcat 5.5.X |
| 11:49.25 | *** join/#tomcat randrew (n=user1@128.sub-70-194-52.myvzw.com) |
| 12:01.20 | *** join/#tomcat mnemonic (i=none@80.91.145.226) |
| 12:33.50 | *** part/#tomcat AvengerMoJo (n=alex@avenger.brainfood.com) |
| 12:38.23 | *** join/#tomcat jco (n=jconti@dsl-du-83-173-253-224.cybernet.ch) |
| 12:54.43 | *** part/#tomcat jco (n=jconti@dsl-du-83-173-253-224.cybernet.ch) |
| 13:24.24 | *** join/#tomcat FreemaniaX (n=Freemani@pcp0011117215pcs.elkrdg01.md.comcast.net) |
| 14:09.20 | *** join/#tomcat adamc (n=adamc@batmanuel.its.unc.edu) |
| 14:14.22 | FreemaniaX | ugh, I wrote a login module and an xml file that is included in the jar can not be found by my class |
| 14:20.09 | adamc | how are you trying to load it? |
| 14:47.48 | FreemaniaX | like this Reader reader = Resources.getResourceAsReader("/com/aaxistechnologies/core/jaassqlmapconfig.xml"); |
| 14:48.26 | FreemaniaX | the calling class is in the same package/directory as the xml file |
| 14:52.46 | adamc | how about getClass().getResourceAsStream("file.xml") |
| 14:53.37 | adamc | http://www.javaworld.com/javaworld/javaqa/2002-11/02-qa-1122-resources.html |
| 15:05.08 | FreemaniaX | that will grab a file in the same package as that class? |
| 15:06.00 | adamc | that's the idea, anyhow =) |
| 15:07.16 | FreemaniaX | that looks good adamc , I'll give it a try |
| 15:22.40 | FreemaniaX | ah, the only problem with getClass is that you cant use it in a static context |
| 15:23.13 | adamc | ohsix, then ClassName.class.getResource() ... |
| 15:24.02 | adamc | that's the static form; it should get you the same Class object. |
| 15:32.48 | FreemaniaX | Thanks adamc , I think that might be the answer |
| 15:33.04 | FreemaniaX | will test it, got drive on icy roads to work this morn |
| 15:33.09 | FreemaniaX | bbl |
| 15:33.12 | FreemaniaX | thanks again |
| 16:06.13 | *** join/#tomcat blumber (n=klumber@adsl-69-232-198-185.dsl.pltn13.pacbell.net) |
| 16:26.17 | *** join/#tomcat Ward (i=ward@hope.webware.be) |
| 16:27.06 | blumber | where is the context.xml file for the jsp-examples ? |
| 16:27.25 | Ward | has tomcat features to use an own UID/GID for each web application? Or has it features to lock/chroot user's webapps ? |
| 16:28.58 | randrew | blumber: it's hiding |
| 16:31.21 | blumber | Ward: I dont know if this is what you look for org.apache.catalina.startup.UserConfig |
| 16:31.36 | blumber | or org.apache.catalina.startup.PasswdUserDatabase |
| 16:31.53 | blumber | those are Listeners |
| 16:32.23 | Ward | I think not. I 'm searching a way to use tomcat in a shared user environment where every user has 1 webapplication |
| 16:32.25 | randrew | blumber: sorry, jsp-examples uses defaults |
| 16:33.02 | swente | Ward: i guess it's impossible [uid|gid] as long as they run in a single process [==vm] |
| 16:33.32 | swente | Ward: i'm not sure, but afaik you might set up appropiate java-security-managers |
| 16:33.36 | blumber | Ward: would it satisfy the /home/username/public_html for each user having its own home directory? |
| 16:34.21 | Ward | swente: ic, and where can I read about the security managers? somewhere in the tomcat documentation 'forest' |
| 16:34.47 | swente | Ward: uuh .. basic information e.g. from sun. see tomcat for how to enable specific securitymanagers for specific webapps |
| 16:34.53 | *** join/#tomcat aaronc (n=acline@12.104.6.129) |
| 16:35.05 | swente | Ward: point is - i just "think" it's possible that way:) |
| 16:35.29 | Ward | ic |
| 16:36.00 | swente | Ward: http://tomcat.apache.org/tomcat-5.5-doc/security-manager-howto.html |
| 16:36.46 | Ward | yeah, I already found that one, but it is very vague |
| 16:37.03 | adamc | how many apps/users do you need? |
| 16:37.35 | Ward | 100-200 or more |
| 16:38.03 | adamc | ohsix, then one vm per uid ain't gonna cut it ... unless you have an E15K in your back pocket ... =) |
| 16:38.39 | Ward | indeed |
| 16:39.07 | Ward | it's on a student development machine, so it hasn't many RAM :p |
| 16:39.52 | adamc | you want each webapp to run as a specified user, and use the OS' access controls? Like suExec for Apache CGI? |
| 16:40.26 | Ward | that would be nice |
| 16:40.35 | Ward | indeed, like suExec |
| 16:41.29 | Ward | there are many requests after a system like that... other people asked some weeks ago after advise how they could set up a system like that and I had to told them I didnt know :) |
| 16:41.42 | Ward | I thought it wasnt possible |
| 16:41.57 | Ward | but it would be nice if there is a solution somehow |
| 16:42.13 | adamc | the thing is that a servlet container is a single process (more or less) while suExec uses the fact that a CGI is a separate process. |
| 16:42.19 | Ward | indeed |
| 16:43.12 | adamc | i'unno, maybe there's some kinda magic you can pull with runAs() and some deep OS voodoo, but it *seems* to me you'd have to basically reimplement the OS' access controls within the JVM. |
| 16:44.19 | Ward | isnt it possible to fork a process from the jvm ? |
| 16:44.27 | Ward | and do a setuid() setgid() in that child process |
| 16:44.53 | adamc | you can spawn a thread, or spawn a process; but if it were a java process, you'd be starting a separate VM. |
| 16:44.55 | Ward | I wrote a php 'wrapper' some time ago, so I could implement such system if it is technically possible |
| 16:47.42 | Ward | if you want to do setuid(), you need to spawn a new process, because it's impossible to get the root right after it |
| 16:48.45 | swente | Ward: if you'd spawn java-processes with different UID|GIDs, you would end up with several hundred processes. probably using the security-manager is more efficient. |
| 16:49.15 | aaronc | anyone know of any good utilities that help you metric the memory usage of a tomcat instance? |
| 16:49.39 | Ward | yeah probably |
| 16:49.52 | adamc | aaronc: run it under a 5.0 JRE and enable the management console? |
| 16:50.04 | swente | Ward: and, security-managers in general allow very fine-grained control about what's accessible and what is not |
| 16:50.11 | aaronc | jre 1.5? |
| 16:50.13 | adamc | aaronc: or a profiler, like the one built into Netbeans |
| 16:50.20 | adamc | aaronc: yeah |
| 16:50.34 | aaronc | adamc: just so you know... I'm not a developer I'm a sysadmin |
| 16:50.57 | adamc | swente: what I'm worried about with that approach is that the securityManager would have to replicate the OS' access controls. |
| 16:51.10 | Ward | thanks for the help, I will take a look at the security manager |
| 16:51.14 | aaronc | management console in Tomcat? |
| 16:51.19 | adamc | aaronc: oh, gotcha ... yeah from a system perspective a JVM is largely a black box. |
| 16:51.20 | aaronc | or is there a management console for java? |
| 16:51.31 | aaronc | adamc: hmmmmmm |
| 16:51.47 | aaronc | that sucks |
| 16:52.05 | adamc | aaronc: with the 1.5 versions of the JRE, you can enable management and connect to it using jconsole, which ships with the JDK. |
| 16:52.26 | adamc | there's also stuff like mc4j (I think that's what it's called) |
| 16:53.50 | Ward | swente: the security manager is not tomcat dependant ? it's something from the java vm? |
| 16:53.57 | swente | Ward: yes |
| 17:06.43 | *** join/#tomcat linuxfreck (n=freck@fry.fmi.uni-passau.de) |
| 17:16.08 | *** join/#tomcat FreemaniaX (n=Freemani@ip67-153-106-58.z106-153-67.customer.algx.net) |
| 17:25.59 | FreemaniaX | adamc, it worked |
| 17:26.24 | FreemaniaX | now I can look at other things like why I can't find this datasource :) |
| 17:30.23 | adamc | FreemaniaX: cool, glad I could help. |
| 18:53.54 | *** join/#tomcat jasonb (i=noneoyer@adsl-66-124-73-250.dsl.sntc01.pacbell.net) |
| 18:55.27 | FreemaniaX | yo jasonb |
| 18:55.34 | FreemaniaX | how goes it? |
| 18:55.36 | jasonb | Hi FreemaniaX |
| 18:55.58 | jasonb | Going okay. A zillion things going on at once as usual. |
| 18:56.03 | jasonb | How bout you? |
| 19:02.14 | FreemaniaX | not quite a zillion |
| 19:02.35 | jasonb | zillion - 3 ? :) |
| 19:02.48 | FreemaniaX | yeah, that's about right |
| 19:03.51 | jasonb | Did you see that netcraft survey that recently was released on web server linux distro market share? I thought that was very cool. |
| 19:04.29 | jasonb | This one: http://news.netcraft.com/archives/2005/12/05/strong_growth_for_debian.html |
| 19:06.21 | jasonb | Well, I'm hearing that ubuntu is growing fast.. and yeah, probably netcraft counts it as Debian -- it is! :) |
| 19:07.05 | linuxfreck | I'm a bit surprised that the market share by cobalt is still that big |
| 19:07.10 | linuxfreck | do they even exist anymore? |
| 19:07.34 | adamc | it's probably server appliances of some sort. |
| 19:07.46 | jasonb | adamc: Check this out: http://www.desktoplinux.com/files/article019/osdl-dtl-survey-12.jpg |
| 19:08.10 | linuxfreck | yes, but afaik sun bought cobalt a few years ago and then decided to obsolete the whole stuff |
| 19:08.16 | jasonb | They did. |
| 19:09.03 | linuxfreck | I'd prefer debian over fedora anytime |
| 19:09.12 | adamc | ubuntu is debian with a more polished desktop. |
| 19:09.17 | linuxfreck | you just have to avoid the debian java stuff |
| 19:09.25 | adamc | (in terms of where it lives in the ecosystem) |
| 19:10.53 | jasonb | We need a Java distro. |
| 19:11.16 | jasonb | No matter which distro you pick right now, Java on it is screwed. |
| 19:11.51 | adamc | heh ... JDS |
| 19:12.12 | jasonb | Hmm.. but that one sucks in too many other ways. |
| 19:13.00 | adamc | it's the gnu userland stuff running on top of a solaris kernel. |
| 19:13.03 | linuxfreck | jasonb: just convince sun or ibm to give their jdk a proper licence |
| 19:13.42 | jasonb | linuxfreck: IBM can't because they're under contract with Sun. Sun won't because they fear Microsoft will end up taking it and polluting it and redistributing it. |
| 19:13.55 | jasonb | It's all Microsoft's fault in the end. |
| 19:14.06 | adamc | best thing evar: javaws on amd64 linux => busted. |
| 19:14.38 | adamc | yet Sun is marketing Opteron "java workstations" |
| 19:14.39 | jasonb | javaws?? What's that? |
| 19:14.43 | adamc | web start. |
| 19:14.46 | jasonb | ahh. |
| 19:14.56 | linuxfreck | jasonb: It doesn't have to be opensource, there are lots of commercial stuff in debian non-free |
| 19:15.10 | adamc | suse ships jdk anyhow. |
| 19:16.36 | jasonb | linuxfreck: Yeah. You're right. But, 1) why would Sun change the license for all of Java only to please broken Debian distro folks? :) 2) if it's not really open source, lots of other distros still won't bundle it at all, so it doesn't really get us out of the problem. |
| 19:17.18 | linuxfreck | fedora is broken in the same way, no? |
| 19:17.22 | jasonb | Java implementations will remain commercial & restricted until Harmony or some other from-scratch OSS implementation matures. |
| 19:18.21 | jasonb | linuxfreck: Fedora ships broken non-compliant Java packages, in violation of Sun's Java license & "IP rights" (why Sun lets them do this I have NO idea). But, Fedora isn't broken in all the other ways that debian is broken. |
| 19:18.47 | adamc | fedora ships gcj/GNU classpath, afaik, not the sun JDK. |
| 19:19.03 | linuxfreck | debian does that, too. where is the difference? |
| 19:19.41 | jasonb | In the case of Java functionality, there isn't much difference between Debian and Fedora. |
| 19:20.05 | jasonb | In the case of lots and lots of other things, there's a huge difference between Debian and Fedora. |
| 19:21.24 | linuxfreck | well I don't quite agree with that. debian just is a bit slow when adopting new things. but that isn't that bad for a server os |
| 19:21.44 | jasonb | Yes it is. |
| 19:22.23 | linuxfreck | well I prefer stability over the latest bleeding edge version |
| 19:22.39 | jasonb | Stability is important, but not at the cost of security, and not at the cost of a huge amount of lost productivity when it becomes time prohibitive to update things in the OS. |
| 19:24.17 | linuxfreck | I'm not quite convinced that e.g. fedora is more secure than debian |
| 19:26.13 | linuxfreck | especially if your fedora install is so old that there are no security updates anymore, and it's too much time to upgrade |
| 19:26.49 | jasonb | It's difficult to prove security. But, with Fedora, adoption of new packages & new package versions is _very_ fast, so there's at least a convenient way to try out new things in a convenient manner (and even roll back in most cases), and there's a way to update to security-patched versions of packages, so if the fix is out there, then you can easily apply it quickly. |
| 19:28.35 | jasonb | Within a major version of Fedora, you can update to new package versions within that major version really easily. But, updating to the next major version isn't as easy, and the main reason is usually that either the kernel version has had a major rev change, or that the libraries have changed in some big way. That'd be hard to upgrade smoothly on any distro. |
| 19:29.04 | linuxfreck | the debian security team is quite fast, too. it also tries to avoid new upstream versions in a stable distro, so you don't have to worry about incompatibilities between versions. and if you want a newer version, you can try the one from testing |
| 19:29.37 | jasonb | Fedora does support upgrading from major version to another major version, but it often fails.. this is *very bad*. Normally, Fedora folks have to avoid doing that. If I could fix one thing about Fedora it'd be the Java packages, but if I could fix two things, it would be the major upgrade migration path. |
| 19:31.09 | adamc | debian will often backport fixes from more recent versions of key packages (e.g. apache) so even if you're running 1.3.19 or whatever, it's got all the security fixes up to 1.3.31 |
| 19:31.20 | adamc | (just as a "pulled out of you know where" example) |
| 19:32.17 | jasonb | I consider that wrong & confusing. |
| 19:32.22 | jasonb | I realize others don't. |
| 19:33.09 | jasonb | But, operationally, it makes things too confusing for when things go wrong and you have to debug and find out where everything happened.. You've got all your versions mixed up in various trees. |
| 19:33.16 | adamc | maybe a little confusing, if you don't know that's going on, but insofar as you're worried about security, it doesn't bother me at all. |
| 19:33.40 | jasonb | Until you have some serious problem and you have to dig through source to find out what went wrong. |
| 19:34.36 | adamc | if you have that sort of capability, to actually dig through the code and fix stuff, then you probably don't need the distribution to package stuff for you. |
| 19:36.26 | linuxfreck | if you use a distro package, you have to use the distro supplied source to debug anyway, since most distros patch their sources in some way or the other |
| 19:39.26 | jasonb | Well, Fedora / RedHat packages take pristine sources from the OSS project and include those untouched in the distro's package, then also include patches that are applied at package build time. So the pristine sources are often patched in some ways, but you can get the patches right out of the distro package source. You know exactly what got patched and it's easy to reproduce it all. |
| 19:39.47 | jasonb | Plus, Fedora / RedHat doesn't try to make one version of the OSS project into another version of the OSS project. :) |
| 19:41.40 | linuxfreck | debian actually does the same |
| 19:41.40 | jasonb | For example, when you do "gcc --version" and it says the version, you know you have that version. You can go into the original gcc source tree and see the source for it (minus patches). It's true if that if you want the full story, you also have to look to see what patches were applied at package build time, but at least you know the versions aren't mixed. |
| 19:41.58 | adamc | I'll have to demur on that point, since I don't know how source .debs are structured, but I do know it's true that srpms unpack the 'official' version and past when you call rpmbuild. |
| 19:42.09 | adamc | s/past/patch/ |
| 19:42.19 | linuxfreck | if you run apt-get source <yourpackage> it will download the original source plus the patches |
| 19:42.55 | jasonb | linuxfreck: What I've seen of debian is: there are suggestions floating around on various web pages on how one should build a debian package, and there may even be relatively strong conventions for how to do it, but that many, many package maintainers don't follow them, and what you end up with is quite a bit more random. |
| 19:43.49 | aaronc | for a linux sysadmin, does anyone have any books they'd recommend on getting going with Tomcat and optimizing the OS for tomcat? |
| 19:44.39 | jasonb | I've unfortunatly seen too many debian packages that are glorified wrappers around a tar.gz file containing source that used to be from the original OSS project but was hacked and slashed all to hell. |
| 19:44.46 | jasonb | (and no patch files to be found) |
| 19:45.18 | jasonb | aaronc: Yes. :) Tomcat: The Definitive Guide, Second Edition.. due out late next year. :) |
| 19:45.36 | aaronc | late NEXT year? *gasp* |
| 19:45.40 | jasonb | :) |
| 19:45.53 | aaronc | Is there a "Tomcat: Less Than Definitive Guide" for in the meantime? |
| 19:46.01 | jasonb | aaronc: You could get the first edition.. it does have some good stuff for that still. |
| 19:46.14 | jasonb | aaronc: That's what this channel is, really. |
| 19:46.33 | aaronc | hrmmmmmm |
| 19:46.58 | adamc | most of the tweaking you can do isn't terribly OS specific tho' |
| 19:47.22 | jasonb | Yeah, it's more JVM specific. |
| 19:47.31 | aaronc | makes sense |
| 19:47.47 | adamc | like "do I have dual opterons or an old PII-400?" |
| 19:47.50 | jasonb | aaronc: Huge hint: Use the IBM JDK 1.5.0. |
| 19:47.52 | aaronc | so no decent books out there? |
| 19:47.58 | jasonb | aaronc: Nope! |
| 19:48.07 | aaronc | IBM JDK is not an option at this point |
| 19:48.21 | jasonb | aaronc: And, use Linux 2.6 running on AMD64. |
| 19:48.36 | adamc | "How Tomcat Works" is supposed to be pretty nitty-gritty. |
| 19:48.48 | adamc | but I have no firsthand experience. |
| 19:49.47 | adamc | I just slather arctic silver all over the inside of the case and plug the server into a 220V output. |
| 19:50.06 | jasonb | heh |
| 19:50.10 | aaronc | I have been handed existing production systems that run either Sun Java 1.4.2 and tomcat 4.1.2 or Sun Java 1.5_05 and Tomcat 5.0.28 |
| 19:50.32 | aaronc | I don't have the option at this point of changing the java, its more like dealing with the systems to make them run as good as possible |
| 19:50.36 | aaronc | and also for my own understanding |
| 19:51.07 | aaronc | althought I appreciate your opinion that IBM Java is the way to go |
| 19:51.29 | adamc | aaronc: well, tweaking the number of threads on your connectors, hooking up Apache / AJP in front if you're not doing that already, and running a profiler to see where code bottlenecks are. |
| 19:51.31 | jasonb | They have put more resources on their implementation than Sun can.. |
| 19:51.59 | adamc | it's usually the developers' fault when the server runs slowly. |
| 19:52.10 | aaronc | adamc: its always the devs fault ;) |
| 19:52.17 | aaronc | heheh |
| 19:52.56 | adamc | Of course I can blame the higher ups for making me admin the service, write the code, and do all the research without training. |
| 19:53.08 | aaronc | good point |
| 19:53.27 | aaronc | and then force the deploy without proper testing |
| 19:53.31 | aaronc | its the same everywhere |
| 19:54.27 | adamc | anyhow ... there are plenty of decent guides out there on the web to tweaking your tomcat installation, but it all boils down to "this may work for you, but then again it might not, so test and benchmark and let us know how it works out" |
| 19:54.48 | aaronc | alright |
| 21:03.27 | *** join/#tomcat ibot (i=ibot@pdpc/supporter/active/TimRiker/bot/apt) |
| 21:03.27 | *** topic/#tomcat is 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 (we all have jobs): Ask your question, then wait; check back often to see if anybody answered. |
| 21:08.30 | aaronc | I have a tomcat instance with like 12,000+ sessions... I know this is most likely caused by a problem in the program |
| 21:08.50 | aaronc | however, this problem doesn't replicate on another server with similar setup (program, java, tomcat are all the same) |
| 21:08.59 | aaronc | any thoughts about why this might occur? |
| 21:09.00 | puff | aaronc: Might be a difference in usage patterns. |
| 21:09.10 | aaronc | hmmmm |
| 21:09.15 | puff | aaronc: Typically why this would occur is because something is causing the browser to lose track of the sessions. |
| 21:09.15 | aaronc | good point... that's likely |
| 21:09.33 | puff | aaronc: More to the point, something is preventing the browser from sending the JSESSIONID parameter. |
| 21:09.44 | puff | aaronc: So tomcat is assigning a new session. |
| 21:09.49 | aaronc | is there a common cause of this? |
| 21:09.52 | puff | aaronc: Do you have any idea what's in those sessions? |
| 21:10.01 | aaronc | I do not |
| 21:10.06 | aaronc | is there a way to find out? |
| 21:10.13 | puff | aaronc: Well, a proxy of some sort that's screwing with the browser reqeust headers, maybe. |
| 21:10.23 | puff | aaronc: I ended up writing a simple class to iterate through the sessions. |
| 21:10.23 | aaronc | don't think its a proxy |
| 21:10.27 | aaronc | user is on the intranet |
| 21:10.54 | aaronc | does that simple class have to be a part of the main program or can I use such a class as a troubleshooting tool? |
| 21:12.13 | aaronc | how can I completely clear out the sessions if I have tomcat shut off? |
| 21:12.45 | puff | aaronc: Hm... |
| 21:13.06 | puff | aaronc: They should clear out when you restart tomcat... UNLESS you have something in web.xml to persist sessions, maybe. |
| 21:14.00 | aaronc | the only thing I see about sessions defined in web.xml is a timeout of 30 |
| 21:14.14 | puff | aaronc: As far as the simple class, basically it needs to be a class that implements HttpSessionListener, which then gets notified of each session creation or destruction. |
| 21:14.27 | DarkStorm | Ok this is annoying: |
| 21:14.27 | DarkStorm | http://issues.apache.org/bugzilla/show_bug.cgi?id=23929 |
| 21:14.36 | puff | aaronc: Hm, by default I believe sessions get cleared out when you restart tomca,t unless you have it configured to persist them across retarts. |
| 21:14.37 | DarkStorm | It says here the standard for URL encoding in GET requests is UTF-8. |
| 21:14.39 | puff | restarts. |
| 21:14.45 | DarkStorm | Yeah heh I'm using UTF-8 but it still doesn't decode correctly. |
| 21:15.28 | puff | aaronc: The class has to be installed as part of the webapp in order to hook into the session events. However, you can just code up the class, include a config entry for the listener in the app's web.xml, and code up a simple JSP to display the sessions. |
| 21:16.06 | puff | aaronc: An alternative (though the session tracking is probably a good idea for the long run) is to get a packet sniffer or a logging proxy, and watch what's going on with the problem app. |
| 21:16.36 | puff | aaronc: Specifically, watch evedrything that goes back and forth between one browser and the server, for a representtive session. |
| 21:16.50 | puff | aaronc: If that doesn't turn up anything obvious, then it may be something that the users are actuallying doing. |
| 21:28.35 | puff | aaronc: Hello? |
| 21:28.47 | aaronc | puff: sorry... one moment |
| 21:29.53 | *** part/#tomcat DarkStorm (i=DarkStor@p54B04F40.dip.t-dialin.net) |
| 21:36.36 | aaronc | puff: I'm a sysadmin, so doing the packet sniffer isn't hard at all... writing the class and hooking that into the session might be more than I want to get into ;) |
| 21:38.16 | aaronc | one other thing to note.... many of the connection come from outside through an apache connector |
| 21:38.23 | aaronc | could apache have something to do with it? |
| 21:44.07 | puff | Hm, could. |
| 21:44.43 | puff | aaronc: I wouldn't see it as especially likely, but is the second (non-problematic) tomcat install using the connector also? |
| 21:44.55 | puff | aaronc: Do you mean mod_jk? |
| 21:45.08 | *** join/#tomcat bdudney (n=bdudney@c-24-9-189-43.hsd1.co.comcast.net) |
| 21:45.16 | aaronc | don't think its mod_jk... its an ajp13 connector |
| 21:45.34 | puff | aaronc: mod_jk just proxies requests to tomcat, AFAIK, shouldn't be interfering with the headers. Of course anything could be going on at the apache level, depending on rewrites, etc. |
| 21:46.15 | puff | aaronc: mod_jk and ajp13 in general are unfamiliar territory for me; I can muddle through setting up a new jkMount enttry, that's about it. |
| 21:46.25 | yel | whats the deiffernce between mod_jk and jk used on Jboss ?? |
| 21:46.26 | puff | aaronc: Oh, what version of tomcat and jdk? |
| 21:46.28 | aaronc | ok |
| 21:46.31 | aaronc | old |
| 21:46.34 | puff | yel: No idea, haven't used Jboss. |
| 21:46.39 | aaronc | 4.1.2 tomcat, 1.4.1 java |
| 21:46.42 | puff | aaronc: Ah, well, of course that could be an issue. |
| 21:46.44 | yel | :) |
| 21:46.46 | yel | thanks |
| 21:46.51 | puff | aaronc: yeeikes, you should at least upgrade toimcat to 4.1.31. |
| 21:46.55 | jasonb | yel: Ask in #jboss, or just call the company and pay for support. |
| 21:47.04 | puff | aaronc: And java to 1.4.2_10. |
| 21:47.09 | aaronc | upgrading around here is more difficult than pulling teeth |
| 21:47.16 | puff | aaronc: Not that I'm going to get all holier-than-thou on you... I know exactly how you feel. |
| 21:47.16 | yel | im already there jasonb but thanks you anyway :) |
| 21:47.32 | aaronc | puff: I do appreciate the sympathy ... heheh |
| 21:47.34 | puff | aaronc: My own company is still using 4.1.31 and 1.4.2_10. |
| 21:47.39 | aaronc | 1lbb&1cl |
| 21:47.57 | puff | aaronc: But upgrades to 4.1.31 and 1.4.2_wahtever should be about as safe as you can get. |
| 21:48.12 | linuxfreck | 4.1.31 will probably be around for quite some time |
| 21:48.28 | puff | aaronc: jasonb has berated me repeatedly about not upgrading :-). |
| 21:48.41 | jasonb | :) |
| 21:48.48 | puff | jasonb: How's life? |
| 21:48.59 | jasonb | "But it's only because I want the best for you.." |
| 21:49.00 | puff | jasonb: I'm still negotiating with my ex-company about what happens next, severance, etc. |
| 21:49.18 | aaronc | upgrading is nice... but the funny thing about this place is that simple upgrades of subversions like 4.1.x to 4.1.y seem to break apps |
| 21:49.21 | aaronc | its nuts |
| 21:49.23 | jasonb | puff: Wow. Are they being nice to you yet? :) |
| 21:49.37 | puff | jasonb: They're stalling right now. |
| 21:50.02 | jasonb | puff: Did you get a new job, or are you hanging out? :) |
| 21:50.02 | puff | jasonb: They gave me an "offer" on monday morning, said to think it over and consult an attorney. I did so, sent them a marked up counter-offer tuesday evening. |
| 21:50.10 | puff | jasonb: Right now I'm hanging out, but funny thing... |
| 21:50.40 | puff | Tuesday afternoon, as I finished up the call with my lawyer, got an email from a local recruiter asking me if I'm interested in a contract gig at a local bank. |
| 21:50.51 | adamc | jboss packages tomcat and wraps it in a jmx blanket, but I'd be willing to bet the JK stuff is exactly the same. |
| 21:50.59 | puff | This is a recruiter and a large-ish bank that I've done business with numerous times in the past. |
| 21:51.03 | adamc | anyhow, happy weekend, for those of you having a weekend. |
| 21:51.17 | puff | So I'm not especially interested in contracting again, but I'm going to go in for the interview, see what they have to say. |
| 21:51.47 | puff | Several of my friends are recommending that I do a short contract (6 months) as a sort of rebound relationship, so I'm not depressed over the ex-company thing. |
| 21:51.58 | puff | Rather than go straight into the next startup project. |
| 21:53.52 | *** join/#tomcat putzel (n=putzel@183-237-221-213-pool.cable.lan.ch) |
| 21:54.16 | puff | aaronc: Yeah, that can be a bitch... |
| 21:54.32 | puff | aaronc: But the alternative is that you get stuck with old, unsupported tools. |
| 21:54.49 | puff | aaronc: automated testing is a really good idea, for that reason... unit tests as well as automated functional tests. |
| 21:55.13 | aaronc | :q |
| 21:55.58 | jasonb | puff: Sometimes it'll make you feel much better to just jump on the next good startup project you can find. Not all of them will be the right one, of course. But, if you did a 6-month contract thing that would give you 6 months to find a good one to join. :) |
| 21:57.02 | puff | jasonb: I'd rather start my own, at this point :-). |
| 21:57.36 | jasonb | puff: Or that. |
| 22:05.33 | puff | jasonb: Also thinking about rekindling an old project, the Bandwidth Collective. |
| 22:05.52 | jasonb | puff: Anything like bittorrent? :) |
| 22:11.14 | puff | jasonb: Kind of odd, actually... started it back in 1996 as basically a club T1. |
| 22:11.26 | puff | This was back when real web hosting didn't really exist :-). |
| 22:12.14 | puff | Pair networks came along and really kinda removed that problem, but we kept it going until mid-1998. |
| 22:12.41 | puff | It was basically a mutant geek club/incubator/researchlab |
| 22:13.23 | puff | Essentially, an alternative to the more traditional options of workplace, university, or research institution. |
| 22:15.44 | jasonb | So is it software, or a geek club? |
| 22:16.59 | puff | A club. |
| 22:17.03 | puff | jasonb: I'm not sure what you're asking. |
| 22:17.25 | puff | jasonb: Basic intent was to create a context that would result in increased opportunities for interesting projects. |
| 22:17.51 | puff | The guy who hosts darksleep for me was somehat involved when we did it back in 1996/97/98, has been suggesting we should revisit the idea. |
| 22:17.59 | puff | I'm thinking that grant funding might be a real possibility. |
| 22:18.20 | puff | We don't have precisely the same vision, though, so I'm not sure. |
| 22:18.41 | puff | Well, I'm sure it's a good idea, and I'm sure he's a good guy, but we have to be on the same page before we start. |
| 22:18.57 | puff | Yeah, I've wanted something like that for years, too.... |
| 22:19.01 | puff | jasonb: Hm... |
| 22:19.15 | jasonb | It could be like an "underweb". |
| 22:19.57 | jasonb | All I know is: hosting a web site should be free .. you shouldn't have to pay slimy companies like Network Solutions for name "protection". |
| 22:21.10 | jasonb | I have a file full of these kinds of ideas, and no time to implement them. :( |
| 22:23.24 | vinse | do you have an idea for how i can make my pc know that my phone is ringing? |
| 22:23.32 | vinse | preferably one that doesnt involve soldering |
| 22:23.41 | vinse | i want to be able to turn my headphones up all the way |
| 22:23.52 | vinse | but people are complaining that i never answer my calls :x |
| 22:24.05 | puff | vinse: bluetooth? |
| 22:24.17 | vinse | my regular old analog deskphone |
| 22:24.19 | vinse | not a cell |
| 22:24.24 | puff | vinse: There are also bluetooth headphones that are designed to listen to music and use with a phone as well. |
| 22:24.42 | vinse | puff: eeenteresting |
| 22:25.03 | jasonb | vinse: Yes, it's called a modem! |
| 22:25.23 | jasonb | vinse: Modems can tell your computer when the phone is ringing, it's one of the things a modem is designed to do. |
| 22:25.33 | jasonb | vinse: Also, they're quite cheap these days. |
| 22:26.30 | vinse | jasonb: it's so crazy it just might work! |
| 22:26.43 | vinse | so just get some super crappy modem and use it just for that |
| 22:26.46 | jasonb | vinse: I've done all kinds of fun things with modems. |
| 22:27.36 | jasonb | vinse: Yes! Just plug it in.. and then when the phone rings, the modem will send "RING." via the serial port. |
| 22:27.58 | vinse | this is an external modem? |
| 22:28.16 | jasonb | Internal or external, it will still be surfaced as a serial device in Linux. |
| 22:28.27 | jasonb | you can run minicom or some other software to interface with it.. |
| 22:28.53 | vinse | god |
| 22:29.04 | vinse | i have bad memories of interfacing with serial devices |
| 22:29.07 | jasonb | And, it would be very easy to write an expect script that does the rest. |
| 22:29.14 | vinse | i used to program "control systems" |
| 22:29.50 | vinse | so we did a lot of things like high end stereo systems that had serial control capabilities |
| 22:29.59 | vinse | but hardly anyone ever used them |
| 22:30.08 | vinse | and the protocols were pretty much an afterthought |
| 22:30.27 | vinse | and the documentation was written by people who claimed to knwo english on their resume |
| 22:30.32 | vinse | but were apparently lying |
| 22:30.36 | vinse | not good times |
| 22:30.39 | vinse | bad times |
| 22:30.40 | jasonb | heh! |
| 22:30.43 | aaronc | Hey puff |
| 22:30.45 | aaronc | you still here? |
| 22:31.32 | vinse | like, an english speaker might have written: BYTE 3: 0 -- unused field, leave as zero |
| 22:31.42 | vinse | but what whoever wrote them did would be |
| 22:31.52 | vinse | BYTE 3: 0 -- i am stationary in zero |
| 22:32.08 | vinse | ^ not a joke, that's in teh protocol docs for a sony switcher |
| 22:37.19 | jasonb | haha |
| 22:37.37 | aaronc | thanks for the help folks... I'm sure I'll be in here again on monday ;) |
| 22:37.39 | *** part/#tomcat aaronc (n=acline@12.104.6.129) |
| 22:38.23 | jasonb | Aparently the olympics will be in China sometime soon? I heard a news report that they're asking english speakers to help them with public signage translation.. one said "Beware of Waterfall" when it really should have said "Slippery When Wet". :) |
| 22:44.34 | linuxfreck | heh |
| 22:44.56 | linuxfreck | olympics are in beijing in 2008 |
| 22:46.01 | puff | Back. |
| 22:49.24 | *** join/#tomcat BelfordCarioca (n=belfordc@200165168081.user.veloxzone.com.br) |
| 22:57.07 | puff | jasonb: Aren't you a mac user? |
| 22:57.16 | puff | jasonb: A friend is asking me about setting linux onna mac. |
| 22:58.29 | jasonb | puff: I really don't want people to think I'm a mac user! :) I'm currently "borrowing" a PowerBook G4 (I hate it, basically). First thing I did was nuke OSuX and put Fedora Core 4 on. :) |
| 22:58.49 | jasonb | puff: So, yes, I run Linux on a powerbook right now. |
| 22:59.00 | jasonb | puff: And, I have just about everything working. |
| 23:08.56 | puff | hee. |
| 23:08.57 | puff | Okay. |
| 23:09.03 | puff | Cool, I'll mention that to him. |
| 23:09.10 | puff | jasonb: I do like the idea of the undernet or whatever. |
| 23:09.32 | jasonb | under*web* |
| 23:09.33 | puff | jasonb: I had a similar idea, back in 2000 when that domain flap went on... |
| 23:09.36 | puff | Underweb, right. |
| 23:09.39 | puff | UnterWeb |
| 23:09.43 | jasonb | heh |
| 23:10.11 | puff | jasonb: Basically the idea of truly competitive DNS, so you could choose which DNS backbone(s) you want to use. |
| 23:10.29 | puff | I think the flap at that point was about AlterNet. |
| 23:10.34 | jasonb | I even saw people doing this very thing back in like 1996.. having their own top-level DNS that wasn't hooked into the root DNS servers, and you could set it as one of your hosts's nameservers and bam! you got all the new web sites. :) |
| 23:11.03 | jasonb | But, people don't always have control over their own DNS settings.. but they almost always have control over whether to install a plugin in their browser or not. :) |
| 23:11.27 | jasonb | Yeah, maybe it was AlterNet. |
| 23:11.51 | jasonb | But IIRC they tried to charge for domains as well or something. |
| 23:12.47 | *** part/#tomcat BelfordCarioca (n=belfordc@200165168081.user.veloxzone.com.br) |
| 23:15.04 | puff | jasonb: Yeah, that was the altenret thing... my thought in 2000 when it came up again was that it oughta be doable to have a plug-in so people could choose their dns bacbone. |
| 23:15.28 | puff | jasonb: Hell, that'd be fun to do as a geek project, a special "secret" DNS backbone so you could avoid all of the bullshit :-). |
| 23:16.07 | puff | jasonb: I have this project idea that I'm trying to flesh out, not sure whether/how it works as a commercial project, but I think it has some truly profound potential. |
| 23:17.07 | jasonb | puff: I'd love to hear, and to give you feedback. |
| 23:17.14 | puff | jasonb: The underlying concept is to take the vocabulary of interaction that people have, by now, by and large become accustomed to, and to reimplement them all as a cooperating suite of protocols/tools, basing it as much as possible on existing internet protocols. And, in the process, turn many of the "rules" upside down. |
| 23:17.39 | puff | jasonb: A simple example would be DNS - everybody "gets" domain names now (regular users, I mean). |
| 23:17.54 | puff | But domains are somethign that are slow to set up, and most people don't have one. |
| 23:18.05 | puff | There's a high amount of intertia to overcome. |
| 23:18.05 | jasonb | Right. |
| 23:18.10 | puff | Also, nobody uses subdomains :-). |
| 23:18.18 | puff | What if you used a zillion subdomains. |
| 23:18.44 | puff | What if you installed a GUI chat client and one of the things it did was automatically go out and register a subdomain for you on one of a number of subdomain serverS? |
| 23:18.51 | puff | And one fo the domain fields would be a public key. |
| 23:18.54 | jasonb | Well, it's human nature to not use subdomains. Everyone is so vein they want to be top-level. They also want their web site to be easy to type in. |
| 23:19.25 | puff | Yeah, the point is to transform the concept of a domain from a static, high-intertia, rare thing, to a fluid, dynamic,c ommon thing. |
| 23:19.32 | jasonb | People want 100% control over what their web site URL is. |
| 23:19.36 | puff | Subdomains are just an easy way to fit it into the existing backbone, etc. |
| 23:19.49 | puff | jasonb: Who says it has anything to do with the web? :-) |
| 23:19.55 | puff | jasonb: Though that's an option too. |
| 23:20.08 | jasonb | I personally take issue with the top-level domain organization. |
| 23:20.21 | puff | jasonb: Me too, but that's not what I'm talking about here. |
| 23:20.42 | jasonb | I think it would be great if top level domains were essentially free, and you could add them whenever you want. |
| 23:21.06 | puff | Gotta phone call about dinner, back later maybe/probably. |
| 23:21.12 | puff | Well, probably not,. It's friday, after all :-). |
| 23:21.13 | jasonb | Oh, okay. |
| 23:21.14 | jasonb | heh |
| 23:21.22 | puff | I'll try to write up a clearer description of this idea, to bounce off you. |
| 23:21.24 | jasonb | Well, I'd like to talk more about this when you get a chance. |
| 23:21.34 | jasonb | Okay. Have fun. |
| 23:21.41 | jasonb | I'll be around all weekend I think. |
| 23:21.43 | puff | jasonb: You wanna be one of the "remote" members of the bandwidth collective? :-) |
| 23:21.54 | jasonb | puff: Sure. |
| 23:22.07 | puff | I'll put you on my little list (they never will be missed). |
| 23:22.09 | jasonb | puff: Is there a mailing list or something? :) |
| 23:22.11 | puff | Ciao. |
| 23:22.13 | puff | And for me, chow. |
| 23:22.17 | puff | jasonb: Right now it's all in our heads. |
| 23:22.23 | jasonb | okay.. see ya... |
| 23:22.27 | puff | jasonb: There probably will be a mailing list and some flavor of chat server. |
| 23:22.41 | jasonb | IRC IRC IRC |
| 23:22.50 | puff | jasonb: Why something so crude? |
| 23:22.55 | jasonb | heh |
| 23:23.03 | puff | jasonb: You forget I started out in the mud world, back in 1988. |
| 23:23.10 | jasonb | Cool. |
| 23:23.17 | jasonb | I started in the BBS world in 1986. |
| 23:23.19 | puff | Though I guess that technically I started out on BITnet. |
| 23:23.29 | puff | jasonb: Yeah, BBSes too, fidonet a lot. |
| 23:23.33 | jasonb | hehe |
| 23:23.37 | puff | jasonb: My first modem was 300 baud :-0 |
| 23:23.42 | jasonb | ditto. |
| 23:23.51 | jasonb | I narrowly avoided 110 baud. |
| 23:24.00 | puff | Something I always meant to explore when I was working out of my house for the first couple of years was using voice synth/recog and similar technology to blur the line between physical and virtual space. |
| 23:24.19 | puff | So if we reignite the collective, I'll probably explore that angle a little. |
| 23:24.22 | jasonb | That would rule. Have you tried festival? |
| 23:24.34 | puff | skitch, the guy I was talking about, is big into telephony tech. |
| 23:24.49 | puff | We've been talking about using festival and sphinx. |
| 23:24.55 | puff | Because I really don't like voicexml... |
| 23:24.58 | puff | Okay, gotta idle. |
| 23:25.02 | jasonb | sphinx will be useful someday. |
| 23:25.06 | jasonb | Okay, see ya! |
| 23:29.25 | vinse | jasonb: i dont know if your modem idea will work for me |
| 23:29.49 | vinse | jasonb: most seem to key off call waiting, i.e. you need to be on the phone |
| 23:30.13 | vinse | i basically just want to rig something up to get a pop up on my desktop when the phone rings, whether i'm on it or not |
| 23:30.55 | vinse | i bet there's something for hearing impaired |
| 23:36.07 | jasonb | Nah.. All you have to do is hook up a hayes compatible modem and when the phone rings (when you're not on the phone already) it sends "RING." to the serial port. |
| 23:36.44 | jasonb | Try it. |
| 23:46.02 | vinse | cool thx |
