00:00.04 | *** join/#tomcat Cas (n=me@wnpgmb02dc1-180-8.dynamic.mts.net) |
00:00.14 | Cas | hi all |
00:02.50 | Cas | trying to connect apache with tomcat on solaris 10 and i can't seem to get the connector to work - mod_jk is the one i should be using right ? |
00:03.05 | jasonb | Zero: My guess is that your Tomcat code has just enough bugs to cause you this trouble. Time to upgrade it. :) |
00:03.35 | jasonb | Cas: Either use mod_proxy, or just use Tomcat stand-alone. |
00:03.50 | Cas | easy to get it to work jasonb ? |
00:03.59 | Cas | or is that more the standard that people use ? |
00:04.26 | jasonb | Cas: It's the easiest way. |
00:04.34 | Cas | alright thanks |
00:04.41 | jasonb | Cas: The very easist way is to use Tomcat stand-alone. |
00:04.51 | jasonb | (which is also the most performant way) |
00:04.59 | Cas | yeah just trying to learn things all the way through |
00:05.05 | jasonb | Cas: If you really must use httpd, use mod_proxy. |
00:05.08 | jasonb | ibot: mod_proxy |
00:05.10 | ibot | rumour has it, mod_proxy is at http://jakarta.apache.org/tomcat/tomcat-5.5-doc/proxy-howto.html |
00:05.15 | Cas | ok thanks |
00:05.28 | jasonb | Tomcat stand-alone is "all the way through". |
00:05.34 | Zero | jasonb: :( Not something we have access to sadly. |
00:06.07 | Cas | yeah but i was reading that tomcat standalone isn't as capable as if it goes through apache |
00:06.39 | jasonb | Zero: Then you're stuck unless you can use a debugger to narrow it down to the exact class file that is causing it to not work properly. If you could do that, you could compile a new .class file for just that one class and put it in server/classes/ and the problem would go away after the next Tomcat restart. |
00:06.47 | Cas | mind if I ask a couple of jsp concept questions ? |
00:06.47 | jasonb | Cas: You read wrong. |
00:07.04 | Cas | thanks for the clarification jasonb |
00:07.17 | jasonb | Go ahead. Anyone in the channel can answer.. I'm working in the background here, so it may not be me. |
00:07.29 | Cas | just wanted to make sure it wasn't taboo |
00:07.48 | Cas | i've done lots of java development but almost no jsp work |
00:08.04 | Cas | i want to write a tiny little web application using authentication for a couple of pages |
00:08.22 | Cas | would using JAAS and JSF sound like they would work ? |
00:09.00 | jasonb | Yes, as long as you use a recent Tomcat v5.5. |
00:09.02 | Zero | jasonb: I'm not really sure how that helps find the problem. The current filter class has public void doFilter(ServletRequest request, ServletResponse response,...), but I can see clearly through both logging and debugging that it's not getting executed in these instances. It would be nice if there were some way of seeing at the application level what it's actually executing in terms of filters :/ |
00:09.07 | Zero | or what it thinks it should be, etc. |
00:09.32 | Cas | going to be using the lastest stable build of tomcat |
00:10.29 | jasonb | Zero: You can try turning up Tomcat's logging all the way and seeing if some of the output shows you what the request gets mapped to, but I have a feeling it won't answer your question. Now, I could be wrong about that, so you should try turning the logging up anyway. Do debug="9" on everything in server.xml and restart. |
00:11.09 | jasonb | Cas: Okay. JAAS would have been a problem if you didn't run a recent Tomcat version. With newer versions, it should work. |
00:11.25 | jasonb | Cas: I don't use JSF, so I can't tell you how well/poorly that works. |
00:12.12 | gverig | Is there any way to hook JSP servlet to external location (to pull in JSP from a source external to WebApplication)? |
00:13.22 | Cas | thanks jasonb appreciate the answers |
00:17.25 | jasonb | Cas: You're welcome. |
00:17.34 | jasonb | Cas: Why would you want to do that? |
00:19.14 | *** join/#tomcat L-----D (n=LongkerD@211.160.165.186) |
00:21.08 | Cas | why would i want to do what ? |
00:22.03 | jasonb | err. gverig I meant.. sorry.. :) |
00:22.55 | Cas | oh alright |
00:23.04 | Cas | thought you were asking why i said thanks ;-) |
00:24.00 | jasonb | hehe |
00:27.36 | gverig | jasonb: because I want to have JSP (web) stuff in one system (web publishing and what not) and application in another. And I want application deployment not to whack my JSPs |
00:32.13 | jasonb | gverig: So why shouldn't they be in separate webapps then? |
00:32.46 | gverig | jasonb: err... They should be. The question is how to make web app reference an external location for JSP files? |
00:32.52 | jasonb | gverig: Are you worried that deploying your application code would end up changing the modification time on each JSP? |
00:33.04 | gverig | Oh, *in* a separate web app |
00:33.53 | jasonb | Yes.. in. |
00:33.55 | gverig | no. My application has basic configurations and JARs. My JSPs have a lot of HTMLs and are updated through separate channels. |
00:34.26 | gverig | jasonb: JSPs use facilities in my web app. |
00:35.41 | jasonb | Okay, so then why not have it as one webapp, and you do your own scripted deployment of files in the webapp dir.. (do not use any .war files.. use an exploded dir instead). That way, you can touch exactly what you want to touch.. then use the manager webapp to tell it when to reload the webapp after you've deployed something new. |
00:35.47 | gverig | jasonb: JSPs carry basic HTML rendering with some references to java beans and other JSPs. Java beans are dropped into context by filters and servlets. But I still want to update JSPs separately |
00:37.30 | gverig | jasonb: That is an option but not one I like for several reasons. 1) we have a content deployment system that works (with workflows and all that good stuff). And files end up in staging location 2) I like to have web apps and logic as far as HTML and other content as possble |
00:38.44 | jasonb | gverig: Well, so then you know lots of details about the organization and operation of your webapp that I don't know, so I can't make further suggestions really. |
00:39.01 | gverig | jasonb: basically, I understand that there could be other ways to do this (like one you suggested), however we have a very explicit separation of responsibilities between people who work with logic and people who do HTML and I would like to keep my stuff on my side of the fence and pull JSPs in |
00:40.11 | gverig | jasonb: Well, as much as I appreciate recommendations (and I do), the original question was pretty simple one, is there a way to pull JSPs from a location outside web app |
00:40.15 | jasonb | gverig: So you have a fence construct, and I don't know what it's composed of. |
00:40.51 | gverig | jasonb: Just different directories in file system (NFS mapped). Nothing too fancy |
00:41.22 | jasonb | gverig: Yes, the way you do it is: you write some code to copy a JSP file from a different absolute path on the file system into a path inside the webapp dir, and then you can invoke that JSP via either a forward or an include. |
00:41.57 | jasonb | gverig: Nothing I suggested would mix files in those dirs in any way other than how you currently have them organized. |
00:43.16 | gverig | jasonb: :-\ ok, thanks. BTW, how does Jasper pull JSP files in? Through ServletContext.getResrouce or through some other way? |
00:44.55 | jasonb | gverig: I don't remember the answer to that question, and it might have even changed since the last time I looked. |
00:45.17 | gverig | jasonb: Again, I understand that there are ways to get JSPs into web app directory and this is an option. The question was whether it was possible to set this up without copying files and I guess the answer is "no", at least to the best of your knowledge... |
00:45.41 | gverig | (which probably still means "no" since in my experience you know Tomcat pretty well...) |
00:46.51 | jasonb | gverig: Well, what's wrong with copying the file? You don't want the server to perform the work of the copy? Okay, if it's on a Linux or Solaris or BSD you could use symlinks, or even hard links if the files are on the same partition. Short of that, you really do need to access them as you would access any file outside the webapp.. by opening it at the absolute path of the file. |
00:47.10 | Cas | bbl |
00:47.58 | jasonb | I doubt that you could point the JSPServlet at a JSP file in some other absolute dir easily because it wasn't really designed for that. But, I must admit that it is possible, and one would need to look at the source to find out. |
00:48.11 | gverig | jasonb: I don't mind setting absolute path (another way is to build a filesystem resource based on absolute path and then have Servlet use that resource) |
00:48.49 | gverig | *nod*, thank you. I will probably do just that. |
00:48.54 | jasonb | JSPServlet, being just a built-in servlet, would allow you to map it in your own web.xml in addition to the one in the CATALINA_HOME/conf/web.xml (the global web.xml), so if you start up your own instance of it, maybe you can control it in a way that would let you point it to JSP files outside the webapp.. heh. |
00:49.52 | jasonb | gverig: The resource thing is overkill if all you're trying to do is make a JSP file get served, and the only reason why it can't is because its path is outside the scope of the webapp dir. This is indeed what symlinks and hard links are for. |
00:50.11 | gverig | Symlinks are an option too, but there are still issues (I don't really want to go there since that's not the primary goal) |
00:50.17 | jasonb | gverig: symlinks and hard links are for bridging gaps like this, while not moving the content around by copying. |
00:51.02 | gverig | jasonb: It adds another step to operations and it adds an extra leg in path if JSP is hit directly |
00:52.19 | jasonb | gverig: Hard links add no such extra leg. |
00:52.50 | jasonb | gverig: As for an extra step.. you're going to take some extra step to make the file serve, whatever you do. |
00:54.58 | gverig | jasonb: 1) they do (they are still named) , 2) I mentioned NFS. Hard links do not work cross- FS. *If* I can set up location in context.xml for the web application I can redeploy application as much as I want and I get no such extra element in path (I would reassign root for JSPServlet) |
00:55.27 | gverig | jasonb: Again, this is what I am trying to do, which does not mean it is the best or smartest. I am researching my options |
00:56.51 | jasonb | What does 1) above mean? |
00:57.09 | jasonb | Okay, with NFS yeah hard links won't really work. |
00:57.28 | gverig | hard links do add another element in path IIRC, since even hard links are named |
00:58.37 | jasonb | A hard link is just another file path to the same data. You have to reference a JSP with a path no matter what. But, with a hard link that is within the webapp dir, at least you have a useable path inside the weabpp dir. |
00:58.49 | jasonb | But, if you're using NFS, you can't do it I don't think. |
00:59.31 | gverig | right, the point is that if I have an app "myapp" and under it I have a link (hard or soft) "jsp", my http path will still be /myapp/jsp |
00:59.42 | gverig | I can do symlinks, I cannot do hard links |
01:00.18 | jasonb | Well, so then you're not talking about referencing individual files, you're talking about mapping a directory with files inside it. |
01:00.23 | jasonb | You didn't mention that earlier. :) |
01:00.53 | gverig | O_O GOD FORBID!!!! |
01:01.08 | gverig | I want to assign a new root for JSPServlet |
01:01.23 | jasonb | Those are two different problems, and my answers were for the former, not the latter. |
01:01.33 | *** part/#tomcat Spiky-{vaio}- (n=Len@85-210-50-89.dsl.pipex.com) |
01:01.37 | gverig | %( I thought I explained the problem clear, sorry |
01:01.51 | *** join/#tomcat larryBoy (n=me@72-160-4-12.dyn.centurytel.net) |
01:01.55 | gverig | I guess I have been inside for a while and it seemed obvious |
01:01.57 | jasonb | re larryBoy |
01:02.06 | jasonb | gverig: It happens. :) |
01:02.17 | larryBoy | hey jasonb |
01:02.33 | gverig | No, most definitely new root directory for JSP servlet, so the whole filesystem for JSP is located someplace else |
01:02.59 | jasonb | gverig: I also don't know the deployment use cases.. but if you have a dir on NFS where new content comes in, you could use rsync to copy in only what changed to a dir in the weabpp, like the root of the webapp for example. |
01:03.59 | jasonb | gverig: Okay, welp, whether JSPServlet will let you do that depends on whether that's an init parameter or not.. I'd guess it isn't, but I'd have to look. |
01:04.00 | gverig | I understand that but I would really like to eliminate any copying if I can, to often that had gone wrong, far too often and it's another step. |
01:04.13 | jasonb | Okay. |
01:04.14 | gverig | I am looking at the source right now ;) |
01:04.26 | gverig | Thanks for your help. |
01:06.51 | jasonb | gverig: Another option would be: put the JSPs in a separate webapp, turn on crossContext="true" on both contexts, and hope for the best. :) |
01:07.31 | gverig | heh... that is an option. Thanks |
01:09.43 | jasonb | You're welcome. |
01:11.45 | *** join/#tomcat Cas (n=me@wnpgmb02dc1-180-8.dynamic.mts.net) |
01:11.59 | Cas | it's annoying that you have to reboot windows for a path change. |
01:12.48 | Cas | how is that ? |
01:13.01 | Cas | what eveil thing does it involve ? |
01:13.27 | jasonb | No evil at all.. |
01:13.39 | jasonb | Just install FC5. |
01:13.42 | Cas | ah |
01:14.07 | Cas | trying to start working more and more on my solaris machine but i don't have a solaris laptop yet |
01:14.32 | jasonb | Run qemu on your laptop and you can run Solaris x86 inside that on your laptop. |
01:14.51 | Cas | can't run solaris on this thing |
01:14.57 | jasonb | Too slow? |
01:14.59 | Cas | the nick isn't hcl compatible |
01:15.12 | jasonb | Doesn't matter. Qemu's nic is. |
01:15.20 | Cas | really .. |
01:15.23 | Cas | interesting |
01:15.25 | jasonb | Yup. |
01:15.41 | Cas | you my friend are a wealth of knowledge |
01:15.46 | gverig | jasonb: I think I have the answer and it's not one I was hoping for. |
01:15.58 | jasonb | gverig: Hehe! |
01:16.04 | gverig | jasonb: it uses ServletContext.getResource(...) and getResourceAsStream(...) |
01:16.28 | gverig | jasonb: And those are defined as "The path must begin with a "/" and is interpreted as relative to the current context root." |
01:16.28 | jasonb | gverig: That's a classloader kind of thing.. probably going through the webapp class loader or something. |
01:16.30 | gverig | :.( |
01:16.36 | gverig | Nope, that's not |
01:16.49 | jasonb | oh.. servletcontext. |
01:16.55 | gverig | Well, I mean it is but it's not the same as Class.getResource |
01:17.20 | jasonb | well, but if you look at the implementation, underneath it's probably implemented as a classloader. |
01:17.27 | gverig | Oh well, I have JSPs anyways :) |
01:17.36 | gverig | s/have/hate |
01:17.59 | gverig | jasonb: Yes it is. I just meant that it's not the same as ClassLoader.getResource or Class.getResource |
01:19.22 | jasonb | gverig: What you're trying to do is reach outside the confines of your webapp's dir, and that's not exactly against the servlet spec, but it breaks the standard pattern. So, IMHO it's perfectly okay to write code that opens the file via an absolute path, and either copies it to a location inside your webapp, or reads it into RAM and parses it as a JSP from in there. |
01:20.18 | jasonb | gverig: Also, you could write a custom classloader and set that in your context instead of the standard loader. |
01:20.21 | gverig | I definitely agree that it's OK, I am just not about to write my own JSP parser ;) |
01:20.33 | jasonb | In your classloader, you could include the path to the other JSPs. |
01:21.06 | gverig | jasonb: I don't think I can do that. Again, it uses a classloader undernieth ServletContext, which is defined by the container |
01:21.38 | gverig | I could wrap JSP servlet with my own and spoof ServletContext but boy would that be ugly ;) |
01:21.42 | jasonb | gverig: That classloader is the webappclassloader I believe. |
01:22.08 | gverig | right |
01:22.37 | jasonb | gverig: http://tomcat.apache.org/tomcat-5.5-doc/config/loader.html |
01:22.48 | gverig | I could really do many things, like rewrite JspCompillationContext |
01:22.58 | gverig | hmm... what's that? (reading) |
01:23.40 | jasonb | Also, more detailed info about how it works here: http://tomcat.apache.org/tomcat-5.5-doc/class-loader-howto.html |
01:23.48 | gverig | Yeah, I could do that. |
01:23.57 | jasonb | This is a pluggable component of the Tomcat container system. |
01:24.25 | jasonb | Writing one of those wouldn't be bad, and it's not so much of a hack. |
01:25.09 | gverig | I don't know about "not so much of a hack" part but yeah, it should be relatively easy |
01:25.11 | jasonb | And, it would at least give you a standard way to load the content. I'm not sure that the JSPServlet uses the webapploader to load the JSP files.. it may not. So, I'm not sure it directly solves the problem you're trying to solve, but it may. |
01:25.36 | gverig | (basically just chain a URLClassLoader with standard web class loader and call it a day) |
01:25.42 | jasonb | Yup. |
01:26.33 | jasonb | Anyway, by this point I probably gave you enough hints. :) |
01:26.45 | gverig | jasonb: Well, from what I can see at the moment JspCompillationContext class uses ServletContext to resolve resources and I am pretty sure (probably ~80%) that ServletContext relies on this class loader |
01:26.59 | gverig | Yup, more then. Thank you very much |
01:27.10 | jasonb | You're welcome. |
01:27.17 | jasonb | Let me know later how you solved this. :) |
01:27.55 | gverig | Sure. |
02:08.13 | *** join/#tomcat valmont (n=chrishol@germany.pspdev.pas.earthlink.net) |
02:12.47 | map7 | to use JSTL do I need to add anything to my web.xml file? |
02:39.05 | *** part/#tomcat DaDave (i=DaDave@dslb-084-057-232-098.pools.arcor-ip.net) |
02:41.34 | *** join/#tomcat Brian_L (n=brian@c-69-137-239-43.hsd1.md.comcast.net) |
03:01.41 | *** join/#tomcat intrpngn (n=intrpngn@pool-71-98-0-6.mdsnwi.dsl-w.verizon.net) |
03:10.05 | russt | map7 not necessarily, afaik |
03:10.37 | map7 | I have this in my JSP, <%@ taglib uri="http://java.sun.com/jsp/jstl/fmt" prefix="fmt" %> |
03:11.13 | russt | that's correct |
03:11.17 | map7 | is this all I need to start using <fmt:message key="welcome.title"/> to call my MessageResource.properties file? |
03:11.43 | map7 | because that's not working, I get ???welcome.heading??? |
03:11.50 | *** join/#tomcat GhettoJava (n=Freemani@c-68-55-101-112.hsd1.md.comcast.net) |
03:11.55 | russt | map7 I think you have to specify that in the configuration file |
03:12.31 | map7 | Yes I have MessageResource specified in there |
03:12.43 | map7 | and it works with struts tags, but not JSTL. |
03:12.58 | *** join/#tomcat intrpngn (n=chatzill@pool-71-98-0-6.mdsnwi.dsl-w.verizon.net) |
03:13.01 | russt | map7 sorry gotta go afk |
03:13.07 | map7 | do I have to tell JSTL where MessageResource.properties is |
03:13.17 | map7 | ok cya |
03:14.02 | map7 | ah, I found it |
03:14.19 | map7 | sometimes it helps just to talk through a problem with someone, thanks russt |
03:35.31 | *** join/#tomcat vinse (n=vinse@c-67-169-25-31.hsd1.ca.comcast.net) |
03:49.34 | russt | map7, you're welcom |
04:11.35 | *** join/#tomcat kRabbit (n=chatzill@24-207-197-231.dhcp.stls.mo.charter.com) |
04:12.02 | kRabbit | Can anyone help with a tomcat 5.* JSP 2.0 issue? |
04:14.57 | larryBoy | oh, probably |
04:15.09 | larryBoy | we might not, but maybe we will |
04:15.19 | larryBoy | you'll have to ask first tho |
04:15.34 | kRabbit | I've got a 5.5 instance of tomcat running |
04:15.54 | kRabbit | with multiple hosts set up |
04:16.05 | kRabbit | one host works with jsp 2.0 ${} |
04:16.07 | kRabbit | and the other doesn't |
04:17.02 | larryBoy | check the web.xml files |
04:17.04 | *** join/#tomcat fafa (n=travis@pool-216-211-72-148.tbaytel.net) |
04:17.12 | kRabbit | for? |
04:17.12 | larryBoy | i'l bet $1 they are not the same |
04:17.16 | larryBoy | the dtd |
04:17.22 | larryBoy | first 3 lines |
04:17.26 | fafa | when i start tomcat, and type links http://localhost a website dont pop up... is this normal.. because i thought tomcat was an http server in itself |
04:17.40 | larryBoy | try http://localhost:8080 |
04:17.43 | kRabbit | try port 8080 |
04:17.45 | kRabbit | wow |
04:17.47 | kRabbit | fast larry |
04:17.48 | larryBoy | haha |
04:17.56 | larryBoy | :P |
04:18.35 | fafa | no go |
04:18.42 | larryBoy | same? |
04:18.49 | larryBoy | oh |
04:18.54 | kRabbit | <web-app version="2.4" |
04:18.57 | larryBoy | try 8081 |
04:19.01 | kRabbit | ooh it didn't copy it all |
04:19.06 | larryBoy | fafa, some linux distros change the port |
04:19.10 | fafa | http://aur.archlinux.org/packages/tomcat/tomcat/tomcat |
04:19.15 | fafa | thats what is used to start the daemon |
04:19.25 | larryBoy | look in server.xml |
04:19.29 | larryBoy | in $CATALINA_HOME/bin |
04:19.53 | larryBoy | in $CATALINA_HOME/conf |
04:19.56 | larryBoy | sorry |
04:20.05 | fafa | ok i will thanks brb |
04:20.55 | larryBoy | look for <Connector port="XXXX" /> |
04:24.49 | fafa | no go |
04:25.40 | larryBoy | what port is the connector on? |
04:26.03 | fafa | it said connector port for a few of them, 8082, 8009, and 8442, none worked |
04:27.02 | larryBoy | 8082 looks like the one you want |
04:27.11 | larryBoy | what OS? |
04:27.16 | larryBoy | firewall? |
04:27.23 | fafa | archlinux |
04:27.36 | fafa | no firewall running, you think i have to pass it through my router? |
04:27.51 | larryBoy | dunno |
04:29.36 | fafa | i do |
04:30.15 | larryBoy | do we have joy? |
04:30.49 | fafa | we do thanks |
05:11.39 | Cas | i'm just starting with jsp's what's a good starting point ? have tomcat 5.5.16 running and ant working but now need a starting point of where to put code/how to compile |
05:12.29 | larryBoy | http://www.learntechnology.net/ |
05:12.36 | Cas | thanks larryBoy |
05:12.39 | larryBoy | good examples there Cas |
05:13.10 | larryBoy | if you want a full-on firehose example, check this one: http://www.learntechnology.net/struts-spring-ibatis.do |
05:13.26 | Cas | firehose ? |
05:16.31 | larryBoy | hah |
05:16.42 | larryBoy | as in "drinking from a firehose" |
05:17.03 | Cas | larryBoy you mind if I fire you a message ? |
05:17.21 | larryBoy | nope |
05:17.52 | larryBoy | how much java work have you done? |
05:18.57 | Cas | java lots jsp not enough |
05:19.24 | Cas | it's more the admin part of jsp's that i don't understand i get the coding part |
05:19.33 | larryBoy | gui stuff, or web with servlets? |
05:20.28 | Cas | it's more that i need to understand how/where all the files go in tomcat |
05:21.02 | Cas | any suggestions for sites for that info ? |
05:22.13 | larryBoy | it hink rick's examples are good for that |
05:22.25 | larryBoy | it hink? |
05:22.26 | Cas | ? |
05:22.29 | larryBoy | i think rick's examples are good for that |
05:22.33 | larryBoy | that url |
05:22.36 | Cas | ah alright |
05:22.41 | larryBoy | they are structured real well |
05:22.56 | larryBoy | basically, he is using struts to do model2 web apps |
05:23.00 | Cas | have to understand why things are put where they are. |
05:23.13 | larryBoy | with a front controller, you never want a user to go directly to a jsp |
05:23.21 | Cas | alright |
05:23.25 | larryBoy | always to the controller that forwards them to the jsp |
05:23.30 | Cas | ok |
05:23.42 | larryBoy | so you can do all the setup code in the controller |
05:23.46 | Cas | going to start looking at this war file deploy it and play with that |
05:24.07 | larryBoy | i helped with the struts-spring-ibatis one |
05:24.38 | Cas | that's the one i'm going to start looking at |
05:24.42 | Cas | thanks again for the help |
05:24.54 | larryBoy | ok, then i am going to bed - tomorrow we ski! |
05:24.56 | larryBoy | :) |
05:25.21 | Cas | ski ? |
05:25.23 | Cas | awesome |
05:25.27 | Cas | i went on saturday |
05:25.43 | Cas | got some pictures of me ski'ing in t-shirt and shorts |
05:26.25 | larryBoy | hah |
05:26.27 | larryBoy | where at? |
05:26.33 | Cas | frostfire |
05:26.44 | Cas | Yay my very first jsp error on my own |
05:26.46 | Cas | hehe |
05:26.55 | larryBoy | hah |
05:27.28 | Cas | no db on it yet.. |
05:27.37 | larryBoy | north dakota? |
05:27.41 | Cas | yeah |
05:27.56 | larryBoy | about 1 hour from my house -> http://blacktailmountain.com/ |
05:28.03 | larryBoy | i am in montana |
05:28.13 | larryBoy | very nw corner |
05:28.40 | Cas | that's a decent sized one |
05:28.51 | Cas | frostfire is a tiny little thing but it's alright for me |
05:29.02 | Cas | anything bigger and i'd likely end up killing myself |
05:29.09 | larryBoy | big mountain is bigger, but i have kids, and this one is friendlier :) |
05:29.20 | larryBoy | hah, yeah, me too |
05:29.27 | larryBoy | my kids make fun of me |
05:29.29 | larryBoy | hahah |
05:29.31 | Cas | lol |
05:29.33 | larryBoy | "hurry up dad" |
05:29.37 | Cas | haven't gotten that far yet |
05:29.57 | Cas | in a while i'll have kids but not for a while yet |
05:30.24 | Cas | have to learn all the things there is to be about being a geek get rich and then i can start dating |
05:30.25 | Cas | hehe |
05:30.41 | larryBoy | heh, they are fun, but if i do not get to bed soon, they won't be fun tomorrow because i'll be grouchy |
05:30.43 | larryBoy | later |
05:30.51 | Cas | thanks again for the help |
05:30.52 | Cas | later |
05:31.12 | *** join/#tomcat bugfixer (n=bugfixer@42gis175.gulftel.com) |
06:02.19 | *** join/#tomcat jasonb (i=noneoyer@adsl-66-124-73-250.dsl.sntc01.pacbell.net) |
06:02.24 | *** part/#tomcat NorthDragon (n=IzhikovN@213.251.192.43) |
06:08.05 | *** join/#tomcat L-----D (n=LongkerD@211.160.165.186) |
06:11.42 | *** join/#tomcat akowal (n=Miranda@62.33.232.135) |
06:15.30 | *** join/#tomcat reefshark (n=reefshar@fw.sogeti.nl) |
06:54.53 | *** join/#tomcat valmont (n=chrishol@dsl092-043-004.lax1.dsl.speakeasy.net) |
07:25.08 | *** join/#tomcat Evellia (n=Evellia@LAubervilliers-151-11-60-55.w193-251.abo.wanadoo.fr) |
07:55.01 | *** join/#tomcat Evellia (n=Evellia@LAubervilliers-151-11-60-55.w193-251.abo.wanadoo.fr) |
07:56.44 | Evellia | hello |
08:01.08 | *** join/#tomcat bugfixer (n=bugfixer@42gis175.gulftel.com) |
08:26.21 | *** join/#tomcat noodl (n=noodl@pdpc/supporter/base/noodl) |
08:38.30 | *** join/#tomcat harobed (n=harobed@LNeuilly-152-23-1-114.w193-252.abo.wanadoo.fr) |
08:56.35 | harobed | hi, are you url about http://mydomain:8180/admin/login.jsp don't work, I've always error 404 |
09:31.35 | *** join/#tomcat Hugh (n=Hugh@puma.mxtelecom.com) |
11:05.30 | *** join/#tomcat omay (n=omay@dslb-088-072-035-126.pools.arcor-ip.net) |
11:05.36 | *** join/#tomcat akowal (n=Miranda@62.33.232.135) |
11:17.22 | *** join/#tomcat cypo (n=opera@82.83.48.185) |
11:22.06 | *** join/#tomcat Evellia (n=Evellia@LAubervilliers-151-11-60-55.w193-251.abo.wanadoo.fr) |
12:05.22 | eeeyore | How do you set the caching headers such as "expires or max-age" on static resources using tomcat? |
13:16.48 | *** join/#tomcat zurg (i=a4akb@ask10-182.qualitynet.net) |
13:21.19 | *** join/#tomcat katerX (n=marco@wotan.wpp-service.de) |
13:21.45 | katerX | Hi, any hints for the following error message? org.apache.tomcat.jni.Error: Too many open files |
13:21.53 | katerX | This is Tomcat 5.5.15. |
13:25.03 | a4akb | katerX: www.google.com tat error |
14:02.43 | *** join/#tomcat joered (n=chatter@81-208-83-246.fastres.net) |
14:04.40 | a4akb | joered am boered |
14:08.17 | larryBoy | haah |
14:08.20 | larryBoy | asshole |
14:24.24 | tkoskine | katerX: Your system doesn't allow Tomcat to open any more files because it (Tomcat) has opened the maximum amount already. |
14:25.41 | tkoskine | You need to either increase the limits (file descriptor count) or teach Tomcat to open less files at the same time. |
14:38.44 | katerX | tkoskine: Yes, I would try to increate the file-max, but I would have to now if I have a wrong tomcat configuration |
14:43.30 | *** join/#tomcat intrpngn (n=intrpngn@69-223-214-2.ded.ameritech.net) |
14:44.00 | *** join/#tomcat GhettoJava (n=Freemani@68.55.101.112) |
14:59.22 | *** join/#tomcat nihilis7 (n=Nihilist@69.93.60.82) |
15:12.09 | *** join/#tomcat bdudney (n=bdudney@71-211-141-201.hlrn.qwest.net) |
15:12.34 | *** join/#tomcat randrew (n=raj@dolmen.cc.columbia.edu) |
15:25.42 | *** join/#tomcat phactor (n=luis@195-23-250-58.net.novis.pt) |
15:25.44 | phactor | hello |
15:25.57 | phactor | anyway available to help? |
15:30.55 | swente | it depends on your issue |
15:31.40 | phactor | i'm trying to deploy a .war file |
15:31.42 | phactor | and i receive |
15:31.52 | phactor | Deploy Upload Failed : Exception null |
15:32.03 | phactor | dunno why |
15:32.12 | swente | you're using the manager-tool to upload the archive? |
15:32.37 | phactor | yes |
15:32.51 | phactor | i configured the access in tomcat-users.xml |
15:33.23 | swente | hmm .. i've *never ever* used that upload-mechanism .. hmm .. the logfiles don't give any more clue? how large is the the .war-file? |
15:33.44 | phactor | 1.5mb is the file |
15:34.07 | swente | that shouldn't exceed any upload-size-restriction. |
15:34.13 | swente | phactor: do you have full access to the box? |
15:34.28 | phactor | yes i'm root |
15:34.42 | phactor | how do i define that access pointing to the customer account? |
15:34.45 | phactor | do i need to do that? |
15:35.06 | phactor | can i pm? |
15:35.15 | phactor | INFO: HTMLManager: list: Listing contexts for virtual host 'alojamentos13.com' |
15:35.15 | phactor | Mar 30, 2006 4:34:49 PM org.apache.catalina.core.ApplicationContext log |
15:35.15 | phactor | SEVERE: HTMLManager: FAIL - Deploy Upload Failed, Exception: null |
15:35.15 | phactor | java.lang.NullPointerException |
15:35.35 | phactor | FAIL - Deploy Upload Failed, Exception: null |
15:35.55 | swente | what do you mean with the customer-account-question? [sorry] |
15:36.15 | phactor | i have about 150 domains in this server, but just one customer using jsp |
15:36.33 | phactor | do i need to put the manager access i created poiting to that customer account? |
15:37.37 | swente | phactor: sorry, on the one hand i still don't know what you exactly mean, on the other hand i never have no clue about hosting multiple customers on one tomcat. |
15:37.55 | swente | phactor: the deployment issue .. |
15:38.00 | swente | phactor: did you try to deploy it manually? |
15:38.35 | phactor | just via the manager |
15:39.18 | swente | phactor: i'd try dropping in it /webapps, to see if tomcat still complains [perhaps that way one gets a more helpful error-message. or it might simply work -- who knows] |
15:40.35 | phactor | FAIL - Unknown command /webapps/ |
15:40.52 | swente | $TOMCAT/webapps ? |
15:42.08 | *** join/#tomcat andres_ (n=andres@200.75.48.240) |
15:42.16 | andres_ | hi |
15:42.39 | phactor | if i try to upload as root |
15:42.40 | phactor | it works |
15:42.42 | andres_ | i'm using tomcat with jaas realm..i want to know how i can get the subject and principals from a servlet or filter |
15:57.10 | phactor | :( |
15:58.12 | swente | phactor: as root? |
15:59.09 | swente | phactor: might it be that the tomcat-processes owner [the user it is running as] has no privileges to write to the directory the manager-app wants to write to? |
16:04.40 | phactor | and how should i specify what directory the files should be writen? |
16:09.44 | phactor | ? |
16:19.42 | *** join/#tomcat GhettoJava (n=Freemani@ip67-153-106-58.z106-153-67.customer.algx.net) |
17:30.11 | *** join/#tomcat putzel (n=burps@148-202-221-213-pool.cable.lan.ch) |
17:54.00 | *** join/#tomcat Ontolog (n=Ontolog@68-118-230-41.dhcp.oxfr.ma.charter.com) |
18:28.44 | *** join/#tomcat valmont (n=chrishol@germany.pspdev.pas.earthlink.net) |
19:12.36 | *** join/#tomcat jasonb (i=noneoyer@adsl-66-124-73-250.dsl.sntc01.pacbell.net) |
19:54.09 | *** join/#tomcat dreivier (n=dreivier@port-212-202-44-200.dynamic.qsc.de) |
19:55.46 | *** join/#tomcat philosophia (n=possum@tangerine.uchicago.edu) |
19:59.25 | *** join/#tomcat Shrdlu (n=etaoin@h-66-167-233-19.mclnva23.dynamic.covad.net) |
20:01.18 | Shrdlu | Hey everybody. I'm having issues setting up mod_jk 1.2.14 with Apache 2.0.55 and Tomcat 5.5.16 -- I've followed the howto in the connector docs, but when I run 'apachectl -t' with the new config, I get an error saying it can't find mod_jk.so. |
20:01.39 | Shrdlu | (Despite the fact that it exists at that precise path - I can open it by copy-and-pasting from the error message.) |
20:08.27 | *** join/#tomcat zurg (i=a4akb@ask13-30.qualitynet.net) |
20:11.51 | a4akb | ibot hmmm |
20:12.39 | a4akb | ibot hmm is huh? |
20:12.41 | ibot | ...but hmm is already something else... |
20:12.47 | a4akb | ibot hmm |
20:12.48 | ibot | [hmm] hidden markov model |
20:12.58 | a4akb | ibot hmmm is huh? |
20:13.01 | ibot | a4akb: okay |
20:28.18 | a4akb | ibot boo |
20:28.20 | ibot | a4akb, for heven's sake, don't do that! |
20:31.10 | a4akb | so |
20:31.15 | a4akb | whats new today |
20:38.09 | *** join/#tomcat D351-B345T (i=none@cpc2-asht1-3-0-cust20.manc.cable.ntl.com) |
20:38.12 | D351-B345T | hey |
20:38.26 | a4akb | Yo. |
20:38.39 | D351-B345T | can someone help me with installing tomcat |
20:38.44 | D351-B345T | ive installed it but i cant run it |
20:39.43 | a4akb | Great! |
20:39.51 | a4akb | And? |
20:40.07 | D351-B345T | and how do i run it |
20:40.25 | D351-B345T | when i click on start service it tries to start but it actually doesnot start |
20:40.35 | a4akb | cant u find the ignition butoon? |
20:40.38 | a4akb | *button |
20:40.55 | a4akb | which OS |
20:41.10 | D351-B345T | XP[ |
20:41.23 | a4akb | YaY |
20:41.32 | a4akb | what error do u get? |
20:42.04 | D351-B345T | no error |
20:42.07 | D351-B345T | it just dont start |
20:42.15 | D351-B345T | when i right click and click on stasrt service |
20:42.25 | D351-B345T | it shows me a start window |
20:42.28 | D351-B345T | then it stops again |
20:42.33 | a4akb | wait |
20:42.37 | D351-B345T | ok |
20:42.46 | a4akb | goto command prompt |
20:42.56 | D351-B345T | ok |
20:43.03 | a4akb | but first explain your alpha numeric nick |
20:43.16 | D351-B345T | DESI-BEAST |
20:43.18 | a4akb | in not less than 1000 words |
20:44.03 | a4akb | beasts are not allowed here |
20:45.25 | a4akb | hmm |
20:45.59 | Shrdlu | Don't you hate it when they believe you? |
20:46.33 | a4akb | hmmm |
20:46.56 | a4akb | he didnt believe me |
20:47.25 | a4akb | is his one of the remaining 90% windows users pissed with the whole world...as the OS wont run the way you want to run it |
20:47.40 | a4akb | *he |
20:49.04 | a4akb | maybe he does not even know how to change his nick |
20:49.27 | a4akb | or even worst, he went to write a composition on his nick (not less than 1000 words) |
20:50.17 | a4akb | blame it on windows and userfriendly_making them lazy attitude |
20:50.35 | a4akb | *its |
20:50.56 | a4akb | Shrdlu mod_jk is a big jk why use a connector? |
20:51.43 | Shrdlu | Is there another way to use both Apache and Tomcat on port 80? |
20:52.09 | a4akb | apache and tomcat are different entities |
20:52.25 | a4akb | what happens when you mix water with petrol/gas/benzene |
20:52.29 | *** join/#tomcat jsisson (n=sissonj@59.167.31.104) |
20:52.44 | Shrdlu | You get something that's got water in it, but is also flammable! Awesome! |
20:52.51 | a4akb | the machine backfires |
20:52.56 | a4akb | its not smooth |
20:53.07 | a4akb | it gets bumpy lesser efficient and unreliable |
20:53.14 | Shrdlu | Metaphors aside, I eagerly await an alternate methodology. |
20:53.16 | a4akb | thats what apache and tomcat do together |
20:53.25 | a4akb | Shrdlu: create your own |
20:53.38 | Shrdlu | Create my own what? |
20:53.49 | a4akb | there is not much work going on mod_jk/2 last knew of |
20:53.58 | a4akb | create your own technology. dont wait for it |
20:54.39 | Shrdlu | I am leet, but not that leet. |
20:54.44 | Shrdlu | Besides, I don't like reinventing wheels. |
20:54.52 | a4akb | k |
20:55.00 | a4akb | then shut up n live with what u got |
20:55.06 | Shrdlu | Anyway, I'd be just fine if Apache weren't erroring out when it tries to load. |
20:55.24 | a4akb | abcdefghijklmnopqrstuvwxyz |
20:55.49 | a4akb | zyxwvutsrqponmlkjihgfedcba |
20:56.01 | Shrdlu | It keeps saying mod_jk.so doesn't exist, but the path is right - I can open it by copy and pasting from the error message. |
20:56.03 | a4akb | 1234567890987654321 |
20:56.28 | *** part/#tomcat jsisson (n=sissonj@59.167.31.104) |
20:56.52 | Shrdlu | And I know the config syntax is right, because it works for libphp5.so on the line above. |
20:56.55 | a4akb | Shrdlu: google up |
20:57.00 | Shrdlu | Tried it, got nothing. |
20:57.03 | a4akb | google up the error |
20:57.17 | Shrdlu | Did. Worked through most of the results, couldn't find anything relevant. |
20:57.25 | a4akb | which os |
20:57.28 | Shrdlu | So I probably botched something fundamental, but I can't think what. |
20:57.30 | Shrdlu | Ubuntu. |
20:57.43 | a4akb | i did not ask ur lastname Shrdlu |
20:57.49 | a4akb | i asked what operating system |
20:57.54 | Shrdlu | 0.o |
20:58.05 | a4akb | Shrdlu Ubuntu |
20:58.15 | Shrdlu | Uh, okay. |
20:58.25 | Shrdlu | Meanwhile, I am still running Ubuntu Linux. |
20:59.36 | a4akb | try #ubuntu |
21:00.02 | a4akb | maybe some secret configuration is required |
21:00.13 | a4akb | i cant evens pell ubuntu |
21:09.44 | *** join/#tomcat cas_ (n=cas@83.98.233.2) |
21:10.29 | *** join/#tomcat Cas (n=me@wnpgmb02dc1-180-8.dynamic.mts.net) |
21:11.01 | *** join/#tomcat Caspian (n=me@wnpgmb02dc1-180-8.dynamic.mts.net) |
21:11.32 | cas | sorry for that Caspian |
21:11.37 | *** part/#tomcat cas (n=cas@83.98.233.2) |
21:35.03 | *** join/#tomcat NsOmNiAc (i=nsomniac@freeshell.ORG) |
21:35.20 | NsOmNiAc | is there a way to find out the version from the command line for tomcat ? |
21:40.19 | *** join/#tomcat Zero (i=zero@c211-28-58-136.fernt1.vic.optusnet.com.au) |
22:16.37 | *** part/#tomcat randrew (n=raj@dolmen.cc.columbia.edu) |
22:18.01 | *** join/#tomcat jasonb (i=noneoyer@adsl-66-124-73-250.dsl.sntc01.pacbell.net) |
22:18.50 | vinse_ | NsOmNiAc: ther'es a version.bat or something that will tell you |
22:19.04 | vinse_ | oh jasonb is here now, he'll know for sure! |
22:19.17 | vinse_ | jasonb: 04 |
22:19.17 | vinse_ | 10[13:35:53] NsOmNiAc: 01is there a way to find out the version from the command line for tomcat ? |
22:20.02 | vinse_ | i believe there's a script of some kind included with tomcat for this purpose? |
22:20.59 | jasonb | Hmm, they added something recently.. let me look.. |
22:21.04 | vinse_ | http://tomcat.apache.org/faq/misc.html#version |
22:21.05 | jasonb | Well, actually, are you talking about 5.5? |
22:21.17 | vinse_ | 5.0 and later apparently |
22:21.56 | jasonb | Well, I don't think there was anything like that in 5.0.x -- a command line script I mean. |
22:22.39 | vinse_ | faq says 5.0.28 |
22:22.45 | jasonb | version.sh or version.bat now does it.. in the bin/ dir. |
22:22.59 | jasonb | The faq isn't talking about a command line script. |
22:23.15 | vinse_ | um |
22:23.24 | jasonb | oops |
22:23.26 | jasonb | yes it is. |
22:23.36 | NsOmNiAc | ok thanks |
22:23.37 | vinse_ | ;) |
22:23.56 | *** join/#tomcat jasonb (i=noneoyer@adsl-66-124-73-250.dsl.sntc01.pacbell.net) |
22:24.09 | jasonb | Sorry.. xchat goof. |
22:24.13 | vinse_ | heh |
22:24.25 | jasonb | Hmm.. I think that faq is misleading because I don't think that was introduced until 5.5.x, but let me check.. |
22:25.54 | jasonb | Nope, I'm wrong. |
22:25.58 | jasonb | It's in 5.0.28. |
22:26.13 | jasonb | seems like just yesterday that they put that script in there. :) |
22:26.32 | jasonb | At this point, I only believe it wasn't there in 4.1.x.. |
22:27.48 | jasonb | Confirmed. It's not in 4.1.31. |
22:28.45 | jasonb | Strangely, the faq was correct on this item.. |
22:37.14 | *** join/#tomcat Rienzilla (i=rien@sinas.rename-it.nl) |
22:37.23 | Rienzilla | Good evening |
22:39.43 | jasonb | Hi Rienzilla. Cool nick. |
22:41.05 | Rienzilla | :) |
22:42.04 | Rienzilla | Hey I have a question. It's not only tomcat related I think.... If I have a tomcat service with an axis webservice environment, and I execute some webservice from it, will I be able to catch the stdin and stdout of the executing java classes somewhere? |
22:44.03 | jasonb | If Tomcat is on your local machine, and you execute code that runs in the same JVM as Tomcat, then stdout does go to one of Tomcat's log files. I'm not quite sure what you mean about stdin.. |
22:44.46 | *** join/#tomcat [a]tomi (n=[a]tomi@pool-71-104-33-213.lsanca.dsl-w.verizon.net) |
22:45.54 | Rienzilla | Well |
22:46.04 | Rienzilla | I meant stdout only as a matter of fact :) |
22:46.23 | Rienzilla | So system.out.println("something") will show up in tomcat logfiles? |
22:46.31 | Rienzilla | s/sys/Sys |
22:48.36 | *** join/#tomcat obsidian-studios (n=obsidian@c-66-177-190-192.hsd1.fl.comcast.net) |
22:48.48 | Rienzilla | hmkay |
22:49.25 | Rienzilla | And did you by chance ever do something with the oracle BPEL process manager? |
22:50.03 | vinse_ | please kill me if i ever have to do anything associated with bpel again |
22:50.22 | Rienzilla | hehe |
22:51.23 | Rienzilla | I implemented a bpel process which should orchestrate the flow of some webservices, but I can't seem to figure out how I can tell an axis webservice where to call back if the bpel process does a request which expects an asynch callback |
22:51.30 | Rienzilla | any clues? |
22:52.29 | obsidian-studios | just stopping by to see if any gentoo people have stopped by lately having any issues with tomcat or etc on gentoo? if they do please send them to #gentoo-java, and I will see about helping them |
22:56.41 | obsidian-studios | jasonb: you know, I might have been one of the people over the past months here bitching about tomcat on gentoo :) unfortunately I lost those logs and etc when my lappy hd died |
23:00.59 | jasonb | obsidian-studios: Hi there. |
23:01.08 | obsidian-studios | jasonb: hello |
23:01.08 | jasonb | obsidian-studios: logs of what? Our conversations? |
23:01.22 | obsidian-studios | jasonb: my irc logs from when I was in here before, months ago |
23:01.39 | obsidian-studios | jasonb: not with you specifically, but I might have bitched here in Oct, or Nov, before getting involved in Dec |
23:01.41 | jasonb | obsidian-studios: http://ibot.rikers.org/%23tomcat/ |
23:01.51 | obsidian-studios | even better |
23:02.03 | jasonb | Yup. |
23:02.25 | jasonb | So, you could look through those and see who asked about or mentioned what. :) |
23:02.53 | obsidian-studios | cool, hopefully google can come to my rescue or I might have to write something myself to parse those pages and etc:) |
23:03.20 | jasonb | Use wget to pull the whole tree to your box and gunzip them all and grep for whatever. :) |
23:03.22 | obsidian-studios | also FYI, more than likely going to lose profiling support on Gentoo, I personally was not a fan of it, I am not sure anyone uses it etc |
23:03.30 | obsidian-studios | ah, doh |
23:03.51 | jasonb | oh, good.. no more silly profiling thingy. |
23:04.06 | obsidian-studios | I am more than likely going to leave the stuff in the fhs locations, but will make symlinks all over the place, so people can find what ever easily |
23:04.22 | obsidian-studios | like /opt/tomcat-5.5 will be liked to /usr/share/tomcat-5.5 |
23:04.36 | obsidian-studios | and inside the /usr/share/tomcat-5.5 will link to anything that is not in that dir |
23:04.50 | jasonb | obsidian-studios: I'm not sure that more symlinking is a better solution. :/ |
23:05.05 | jasonb | Maybe to the CATALINA_HOME dir.. |
23:05.13 | obsidian-studios | well other than the ones I create, once I remove profiling most others go away |
23:05.19 | jasonb | (the root of the CATALINA_HOME dir I mean) |
23:05.28 | jasonb | good! |
23:05.36 | obsidian-studios | oh, most stuff is already linked I believe |
23:05.54 | obsidian-studios | just want to adhere to fhs, while having stuff in the "standard" tomcat locations |
23:06.51 | jasonb | Something I don't like about the version number being in the dir name is: when the user upgrades to Tomcat 5.7 someday, then the CATALINA_HOME dir path changes, and all their scripts where they hard coded that break. Yeah, I know, they shouldn't hard code it.. but people do this. I've seen it a lot. |
23:06.57 | obsidian-studios | ok, there are no symlinks currently in CATALINA_HOME |
23:07.41 | obsidian-studios | jasonb: noted, I will see about losing it in the future |
23:07.57 | obsidian-studios | however it's possible someone might have a couple versions, like I still have 5.x on my dev box |
23:08.07 | jasonb | obsidian-studios: I knew of some additional reasons why putting the version in the dir name was bad, and I've already forgotten them. |
23:08.11 | obsidian-studios | we slotted 5.5 (slotting being a portage thing) |
23:08.30 | obsidian-studios | jasonb: yeah, I get your point, I don't have much of a problem with that |
23:08.46 | jasonb | Well, if the package is relocateable, then even if the CATALINA_HOME dir name is always the same that's not a problem becuase they can always relocate the second TC version package someplace else. |
23:08.57 | obsidian-studios | jasonb: just the multiple versions? have to have a way for that? maybe not? |
23:09.11 | jasonb | huh? |
23:09.19 | obsidian-studios | jasonb: yeah, FYI, we also have /etc/conf.d/tomcat* files |
23:09.30 | obsidian-studios | jasonb: someone might want tomcat 3.x 4.x, 5.x etc on the same machine |
23:10.10 | jasonb | obsidian-studios: At least with RPM it is possible to write relocateable RPMs that would allow relocating all of those dirs so their dir names & paths don't collide. |
23:10.19 | jasonb | obsidian-studios: My relocateable RPM set is like that. |
23:10.30 | obsidian-studios | jasonb: hmm, let me think about that |
23:10.40 | obsidian-studios | jasonb: your saying to relocate the existing stuff? |
23:10.59 | obsidian-studios | like moving /opt/tomcat to /opt/tomcat-5, when overwriting it with 5.5? |
23:11.30 | jasonb | So, when you have config files in /etc like /etc/conf.d/tomcat* <-- a file, then when you relocate it you could make the ending dirname of the relocation path be the name of the thing you're configuring, so they can name the path, say, /opt/tomcat-5.5, and then the config file becomes named /etc/conf.d/tomcat-5.5. |
23:11.54 | obsidian-studios | ok |
23:11.58 | obsidian-studios | that's how they are now |
23:12.08 | obsidian-studios | but your suggesting the primary one just be called tomcat |
23:12.18 | jasonb | You can move it after it was already installed, but then anything else already installed that depends on it may not find it anymore. I'm mainly referring to specifying on the command line where to install a package at install time. |
23:12.31 | obsidian-studios | only versioning when relocating |
23:13.00 | obsidian-studios | for the most part people should not be hard coding any paths no? |
23:13.11 | obsidian-studios | since they should be using the CATALINA_HOME, and BASE vars |
23:13.14 | jasonb | Yes, by default it should just be called "tomcat" because the usual case is that people have exactly one installation of Tomcat, the latest stable. For the pro users who want more than one, installed on the same machine at the same time, they will need to relocate at least one of the tomcat installations. |
23:13.34 | obsidian-studios | ok, not sure if that will be possible |
23:13.45 | obsidian-studios | since like in my case, I wanted to test 5.5 without effecting my 5.x install |
23:14.13 | obsidian-studios | I will see what I can do, but we might always have version #'s on gentoo, on the flip side, we never version jars |
23:14.27 | jasonb | Yes, people should use variables for paths where they can. But, you should try not to break the case where they want to hard code a path.. if you can't get around it in some spots, fine. But, where you can help it to work, it helps people (possibly people who aren't real bright, but they're still your users!) |
23:14.32 | obsidian-studios | just for the reason why you are suggesting we should remain tomcat to just tomcat |
23:14.53 | obsidian-studios | s/remain/rename |
23:15.08 | jasonb | Yeah. |
23:15.29 | obsidian-studios | it's noted, if I can do it I will, I think it might go against how gentoo does that stuff |
23:15.44 | obsidian-studios | jasonb: it's always worked great for me, I have yet to have any gentoo specific problems |
23:15.53 | jasonb | It might.. because the gentoo way isn't quite flexible enough yet. |
23:16.18 | jasonb | obsidian-studios: No, I mean relocating the packages so you can have two versions of the same package on a box and use both concurrently. |
23:16.20 | obsidian-studios | no it's the opposite |
23:16.31 | obsidian-studios | just like apache, you can have /etc/apache, and /etc/apache2 |
23:16.44 | jasonb | but /etc/apache is a dir, right? |
23:16.46 | obsidian-studios | we have that now, just no relocation necessary |
23:16.51 | obsidian-studios | both are the same |
23:17.18 | obsidian-studios | both are dirs, one would be for 1.x and the other for 2.x, not sure what's going to happen with 2.2 I would assme apache-2.2 or etc |
23:17.37 | obsidian-studios | jasonb: I can easily do what you are asking with a symlink though |
23:17.44 | jasonb | With the jpackage.org package, it creates an /etc/tomcat5 dir, and that's not relocateable.. so you can't have, say 5.0.28 and 5.5.15 both installed since the /etc/tomcat5 dir would collide. |
23:17.54 | obsidian-studios | and have the unversioned symlink point to what ever is the lastest installed tomcat |
23:17.58 | jasonb | Yeah.. see? |
23:18.07 | jasonb | ugh. |
23:18.17 | obsidian-studios | yes, hat's why we have /etc/tomcat-5.5 and /etc/tomcat-5 |
23:18.43 | jasonb | Okay, so with your system of doing things, how can I install 5.5.9 and 5.5.16 and run them both concurrently? |
23:18.58 | jasonb | See what I'm saying? :) |
23:19.00 | obsidian-studios | that you can't do since they are same generation, same "slot" |
23:19.18 | jasonb | Yeah, well, with my RPM set you can. |
23:19.21 | obsidian-studios | ok, so you want us to go another version deep, or etc, exact versions |
23:19.21 | jasonb | It's fully relocateable. |
23:19.32 | obsidian-studios | interesting |
23:19.36 | jasonb | No, I'm saying the version *is* the problem. |
23:19.46 | obsidian-studios | yes, but when relocated the full version |
23:20.06 | obsidian-studios | if we did full versions now, we could do just what you suggested |
23:20.18 | obsidian-studios | I get what you are saying |
23:21.01 | jasonb | IMHO, it's just as useful to have 5.0.x and 5.5.x installed and both working concurrently as it is to have 5.5.x and 5.5.y. |
23:21.09 | obsidian-studios | I agree |
23:21.17 | jasonb | They're for somewhat different purposes, but they're both very useful. |
23:21.40 | obsidian-studios | usually not much of an issue to go up in revisions though, usually bug fixes and etc, less likely for one to want to do that, than between versions |
23:21.47 | obsidian-studios | but for testing sake, I could see it happening |
23:21.48 | jasonb | Anyway, it's best to avoid any potential collision points. |
23:22.12 | jasonb | It does happen.. unexpected bugs when you go up a very minor rev. |
23:22.38 | jasonb | People often just want to try the next version, and decide not to because they'd have to rip out their otherwise well working Tomcat in order to do it. |
23:22.40 | obsidian-studios | yeah, I am not sure calling it just tomcat will be doing things the "gentoo" way, at the same time, I do not think how things are, to include full version, might make things worse |
23:22.40 | jasonb | Why should they have to? |
23:23.30 | jasonb | I think the distro ought to be kept separate from user configuration. |
23:23.33 | obsidian-studios | true, in our case so long as they did not nuke their older configs, could just re-emerge the older version, but I get your point and for testing alone would be a good thing |
23:23.42 | jasonb | And, the distro should be switchable. |
23:23.43 | obsidian-studios | but I think the issue is much larger than just tomcat on gentoo, in that regard |
23:23.52 | jasonb | distro being Tomcat distribution.. sorry. |
23:24.05 | jasonb | Yup. |
23:24.32 | jasonb | welp, I have to get back to work here. |
23:25.03 | obsidian-studios | like my current issue with modular X, I can't install that over my existing X, I have to uninstall X first, then emerge the modular one, and if something goes wrong, it will suck |
23:25.20 | obsidian-studios | np, I was just stopping but, appreciate the info, noted your suggestions |
23:25.27 | obsidian-studios | got to get back to working myself |
23:25.35 | jasonb | Cool. |
23:26.03 | jasonb | obsidian-studios: You might want to have a look at my RPM package source sometime. |
23:26.14 | jasonb | obsidian-studios: It's pretty clean, and well commented. |
23:26.52 | obsidian-studios | cool, not to familiar with rpm these days, but I will look into it when I get a chance |
23:26.54 | jasonb | http://www.webdroid.org:8080/repo/viewcvs.cgi/ |
23:27.01 | obsidian-studios | most likley going to start addressing this on Sat or Sunday |
23:27.13 | jasonb | Well, it's not all RPM-specific.. for instance I have an init script in there. |
23:27.38 | obsidian-studios | cool, did you ever look at that init one we made? you like? dislike? etc |
23:28.05 | jasonb | I have it saved, and looking at it is at the top of my todo list.. probably will look within the next day or so. |
23:28.17 | jasonb | For those, I like to take my time and really understand it. |
23:28.36 | obsidian-studios | np |
23:28.52 | obsidian-studios | it's compacted so not a clear easy read |
23:29.00 | obsidian-studios | not to bad, but not much duplication either ;) |
23:29.05 | jasonb | I'm really good with bash / sh script, so I'm sure it'll be understandable. |
23:29.10 | jasonb | I just like to soak in all the detail. :) |
23:29.17 | obsidian-studios | yeah, it's just the guts from catalina.sh |
23:29.21 | obsidian-studios | np |
23:29.55 | obsidian-studios | I helped out on it, but others who are better at bash scripting than me, provided allot of feedback, suggestions etc |
23:30.09 | obsidian-studios | in my experience it seems to at least stop tomcat faster ;) |
23:33.14 | jasonb | There's fast, and then there's reliable.. so that should be a user-tuneable setting. |
23:33.31 | *** join/#tomcat stjames (n=stjames@65.19.56.22) |
23:34.09 | obsidian-studios | it's reliable |
23:34.28 | obsidian-studios | I just seem to recall catalina.sh invoking both start stop when stopping? |
23:35.19 | jasonb | by reliable I'm referring to the threads inside of Tomcat. Do they have sufficient notice to shut down gracefully all the way before the JVM is forceably shut down? If you force the JVM process down too fast, the answer is no, and the user's webapp threads do not get enough time to properly save out everything, so the user's webapp actually loses data. |
23:36.46 | obsidian-studios | well we are still calling bootstrap and etc just the same |
23:36.48 | obsidian-studios | except we pass it a single value, not two |
23:36.53 | jasonb | My init script has an environment variable setting for how long to wait before shutting down the JVM. |
23:36.55 | obsidian-studios | only stop, not start stop |
23:37.01 | obsidian-studios | interesting |
23:37.11 | obsidian-studios | catalina.sh does not have that correct? |
23:37.27 | jasonb | Nope. Catalina.sh by itself implements unreliable shutdowns. |
23:37.46 | jasonb | In fact, it's not even reliable that the TC JVM comes down at all when using just catalina.sh. |
23:39.25 | obsidian-studios | exactly the problems I was seeing and trying to resolve with a direct init script |
23:42.47 | jasonb | Well, unless you're using kill, you're not going to prevent JVM hangs. |
23:51.06 | obsidian-studios | wow |