00:59.28 | *** join/#tomcat copland (n=rob@209.241.118.121) |
01:53.48 | *** part/#tomcat copland (n=rob@209.241.118.121) |
04:22.26 | *** join/#tomcat acidjnk (i=acid@p5B06AA67.dip0.t-ipconnect.de) |
04:54.32 | *** join/#tomcat codeshepherd (n=codeshep@61.8.208.15) |
05:16.23 | *** join/#tomcat LongBeach (n=mike@AFontenayssB-152-1-52-80.w82-121.abo.wanadoo.fr) |
05:45.05 | *** join/#tomcat pfn_cIc (n=pfnguyen@64.235.249.50) |
07:13.14 | *** join/#tomcat fix_ (n=fix@cable-87-244-191-178.upc.chello.be) |
07:13.18 | *** join/#tomcat seevash (n=jixjax@user-1121ng4.dsl.mindspring.com) |
07:40.37 | *** join/#tomcat crummygummy_ (n=CrummyGu@dsl-242-11-30.telkomadsl.co.za) |
07:42.07 | *** join/#tomcat crummygummy_ (n=CrummyGu@dsl-242-11-30.telkomadsl.co.za) |
07:45.46 | *** join/#tomcat blaa32123 (n=user1@2.80-203-34.nextgentel.com) |
07:47.09 | blaa32123 | I can only acces my pages from localhost:8080, how can I set up tomcat to make the pages available from ip:8080 ? |
07:53.16 | *** join/#tomcat crummygummy_ (n=CrummyGu@dsl-242-11-30.telkomadsl.co.za) |
07:54.52 | *** join/#tomcat matej_k (n=Matej@nat-88-212-20-60.antik.sk) |
08:21.38 | *** join/#tomcat Nicke (n=niclasa@ua-83-227-140-135.cust.bredbandsbolaget.se) |
09:24.51 | *** join/#tomcat CrummyGummy (n=CrummyGu@dsl-242-11-30.telkomadsl.co.za) |
09:28.52 | *** join/#tomcat matej_k (n=Matej@nat-88-212-20-60.antik.sk) |
09:31.48 | *** join/#tomcat CrummyGummy (n=CrummyGu@dsl-242-11-30.telkomadsl.co.za) |
09:44.46 | *** join/#tomcat tourix (n=tourix@LPuteaux-151-42-21-52.w193-252.abo.wanadoo.fr) |
09:44.52 | tourix | hello |
09:53.33 | tourix | [Tomcat 5.5.17 - JVM 1.5 Debian 4.0] Hello I'm running un webapp that uses a lot of cpu charge. It keeps increasing regardless the number of users. I guess that I have an source code issue (infinite loop or something like that) |
09:54.00 | tourix | Is there a way that I can locate the problem? (maybe by analysing the jvm?) |
10:01.12 | lyken | audit the source code |
10:04.33 | tourix | yes but how? |
10:08.00 | lyken | one line at a time |
10:08.12 | lyken | beat the developer |
10:09.19 | tourix | but the web app is quite huge ^^ and I'm not really sure that it's a infinite loop |
10:42.10 | *** part/#tomcat blaa32123 (n=user1@2.80-203-34.nextgentel.com) |
11:02.16 | *** join/#tomcat Hugh__ (n=Hugh@puma.mxtelecom.com) |
11:14.57 | *** join/#tomcat varek (n=sh2-136@storkey.plus.com) |
11:20.18 | *** join/#tomcat crummygummy_ (n=CrummyGu@dsl-242-11-30.telkomadsl.co.za) |
11:32.53 | *** join/#tomcat wam (n=wam@p57A57D3D.dip.t-dialin.net) |
11:53.46 | *** join/#tomcat dvayanu1 (n=another@ppp-82-135-3-210.dynamic.mnet-online.de) |
11:54.50 | dvayanu1 | re |
12:23.09 | *** join/#tomcat CrummyGummy (n=CrummyGu@dsl-242-11-30.telkomadsl.co.za) |
12:25.09 | *** join/#tomcat _seshf_ (i=kvirc@mx1.hesgdv.de) |
13:00.56 | *** join/#tomcat randrew (n=raj@dolmen.cc.columbia.edu) |
13:42.01 | dvayanu1 | someone familiar with tomcat debian package? |
13:48.45 | *** join/#tomcat aker (n=aker@87.18.109.211) |
13:48.50 | aker | !list |
13:49.24 | *** join/#tomcat matej_k (n=Matej@nat-88-212-20-60.antik.sk) |
13:50.17 | *** join/#tomcat compair1 (i=ariel@gateway/tor/x-d50ff1657146ed55) |
13:59.46 | *** join/#tomcat a4akb (n=a4akb@62.215.156.215) |
14:00.44 | *** join/#tomcat codeRat (n=codeRat@84-255-206-208.static.dsl.t-2.net) |
14:17.22 | *** join/#tomcat codeshepherd (n=codeshep@cm236.delta24.maxonline.com.sg) |
14:56.03 | *** join/#tomcat fzlogik (n=fuzelogi@213.254.185.115) |
14:56.45 | *** join/#tomcat codeshepherd (n=codeshep@cm236.delta24.maxonline.com.sg) |
15:21.52 | *** join/#tomcat Hali_303 (n=surfk@dsl5400CC90.pool.t-online.hu) |
15:38.50 | *** join/#tomcat markl__ (n=mark@c-24-10-214-91.hsd1.ut.comcast.net) |
15:39.25 | markl__ | newb question: i am trying my first tomcat servlet; is there a way to log errors like 404's from coyote? |
15:39.45 | markl__ | i'm using 5.5.23, just untar'd to /usr/local |
15:39.59 | markl__ | logs/catalina.out doesn't really give much of a log |
15:40.17 | *** join/#tomcat awelynant_ (n=cskinfil@208.65.172.147) |
15:40.22 | a4akb | Hi. |
16:00.24 | *** join/#tomcat [algo] (n=ilia@88.247.218.172) |
16:09.26 | *** join/#tomcat sylkwyrm (n=No@160.79.67.10) |
16:09.46 | sylkwyrm | Tomcat 6, Java 6, and WIndows XP . I can't seem to configure juli to log anything but info |
16:10.12 | sylkwyrm | Tomcat 6.0.13 |
16:10.33 | sylkwyrm | shouldn't just dropping a logging.properties file into WEB-INF/classes be enough? |
16:24.36 | *** join/#tomcat karstensrage (n=karstens@75.55.199.5) |
16:37.03 | a4akb | hi |
16:41.48 | *** join/#tomcat Hydra_ (n=Hydra@80-169-47-2.stanhope.mezzonet.net) |
17:09.34 | *** join/#tomcat slrhes (n=lrademey@64.31.174.28) |
17:14.27 | slrhes | TC: 6.0.13 |
17:14.43 | slrhes | Java: 1.6.0_01 |
17:14.54 | slrhes | OS: Windows 2003 Server |
17:15.28 | slrhes | I'm having trouble getting tribes to produce detailed logs |
17:15.42 | slrhes | All I get is some membership info by default |
17:15.58 | slrhes | Tried editing logging.properties but didn't get very far |
17:16.25 | slrhes | Docs say to enable logging on the key org.apache.tribes.MESSAGES - what does that translate to? |
17:16.26 | slrhes | thanks |
17:39.22 | *** join/#tomcat agentx0r (n=Zach@c-68-51-119-43.hsd1.in.comcast.net) |
17:39.40 | agentx0r | does tomcat support "silent" installs with the options needed by the installer being passed via the command line or some other method? |
17:43.21 | *** join/#tomcat compair1 (i=ariel@gateway/tor/x-a69efa6bd698069a) |
18:03.59 | *** join/#tomcat Nicke (n=niclasa@h161n2fls31o808.telia.com) |
18:04.29 | *** join/#tomcat fix_ (n=fix@cable-87-244-191-178.upc.chello.be) |
18:05.23 | *** join/#tomcat Kebap23 (i=Gebag@dslb-084-061-183-254.pools.arcor-ip.net) |
18:06.06 | *** part/#tomcat Kebap23 (i=Gebag@dslb-084-061-183-254.pools.arcor-ip.net) |
18:31.39 | dvayanu | silent installs? |
18:32.31 | dvayanu | agentx0r: i doubt tomcat "supports" installs at all... |
18:32.42 | agentx0r | i mean the installer, sorry |
18:33.04 | dvayanu | which installer? |
18:33.23 | dvayanu | usually you download it and unpack it :-) |
18:33.52 | agentx0r | ive got the windows installer, apache-tomcat-6.0.10.exe |
18:37.01 | *** join/#tomcat _ramo (n=ramo@p54A97646.dip.t-dialin.net) |
18:37.09 | _ramo | hi. |
18:42.29 | _ramo | I have this versions: 2.6.18-4-486 #1 Wed May 9 22:23:40 UTC 2007 i686 GNU/Linux |
18:42.29 | _ramo | tomcat Apache Tomcat/5.5.23 and JavaVersion:Java(TM) SE Runtime Environment (build 1.6.0-b105) |
18:42.29 | _ramo | My problem is, when i'm deploying my project (and it's copied into my apache.../webapps directory), it's not going to work on demand. i have to restart my tomcat-server every time. is there a trick how to get a workaround? |
18:52.37 | randrew | _ramo: you may want to install the manager app, secure it, and use it to restart your app |
18:57.27 | _ramo | okay, thank you randrew. i will google for it |
18:58.10 | *** join/#tomcat rellis (n=rellis@66.60.179.61) |
18:59.10 | rellis | If I add to tomcat-users.xml do I need to restart tomcat for that to take effect? I'm using Tomcat 5.5.20 on AIX 5.3 w/IBM Java 6 |
18:59.44 | _ramo | but isn't there a way to modify a config file, so that i have only to refresh my page and see if my changes take effect? |
19:01.57 | randrew | _ramo: see also the "deployer" docs for your version of tc |
19:02.20 | _ramo | thank you |
19:14.18 | _ramo | so i read the chapter about "Deploying on a running tomcat server", i looked at my con/server.xml and autodeploy is set true, but this isn't working. I'll try the Tomcat Manager |
19:29.43 | *** join/#tomcat codeshepherd (n=codeshep@bb219-74-162-141.singnet.com.sg) |
19:43.57 | *** part/#tomcat rellis (n=rellis@66.60.179.61) |
19:44.34 | *** join/#tomcat wam (n=wam@p57B57E9F.dip.t-dialin.net) |
19:49.20 | *** part/#tomcat slrhes (n=lrademey@64.31.174.28) |
20:02.09 | *** join/#tomcat wltjr (n=wltjr@gentoo/developer/wltjr) |
20:50.20 | *** join/#tomcat slrhes (n=lrademey@64.31.174.28) |
20:51.07 | slrhes | Java 1.6.0_01, Tomcat 6.0.13, Windows 2003 Server |
20:51.35 | slrhes | Anyone here know how to enable logging for Tribes? |
20:54.20 | *** part/#tomcat _ramo (n=ramo@p54A97646.dip.t-dialin.net) |
20:59.10 | *** join/#tomcat Nicke (n=niclasa@h161n2fls31o808.telia.com) |
21:08.19 | *** join/#tomcat fzlogik (n=fuzelogi@213.254.185.115) |
21:35.04 | *** part/#tomcat agentx0r (n=Zach@c-68-51-119-43.hsd1.in.comcast.net) |
21:56.43 | *** join/#tomcat asymptonic (n=scgmille@216-188-226-118.dyn.grandenetworks.net) |
21:58.12 | asymptonic | Can anyone help me figure out why Tomcat is not sending session data to a cluster peer when the session changes? I'm seeing no traffic other than membership traffic (which is working correctly). Tomcat 5.5.20 on Ubuntu Server Feisty, Java 1.6.0. |
21:58.38 | asymptonic | If you guys have a paste bot I'll paste my configuration |
21:59.11 | dvayanu | http://rafb.net/paste |
21:59.27 | dvayanu | but i doubt someone who is experienced with tomcat clustering is here |
21:59.47 | asymptonic | :( |
21:59.58 | asymptonic | It seems very few people use it |
22:00.08 | dvayanu | for a reason :-) |
22:00.17 | asymptonic | I wish there were an alternative |
22:00.24 | dvayanu | loadbalancer? |
22:00.45 | dvayanu | own implementation? |
22:00.49 | asymptonic | http://rafb.net/p/hWMVNK73.html |
22:01.16 | asymptonic | Load balancing is only half the problem. We can't have sessions expiring when a machine goes down, and they 'go down' often as we deploy new functionality |
22:01.26 | dvayanu | why |
22:01.28 | asymptonic | Apache with mod_proxy_balancer btw |
22:01.41 | dvayanu | is it a business requiremnet? |
22:02.01 | asymptonic | Why? Because our site is very dynamic. Session expirations look very bad to the users |
22:02.36 | asymptonic | pretty much everything a user does, even when not logged in, requires state to be maintained. |
22:02.48 | dvayanu | thats clear, but how often do you release? |
22:02.54 | asymptonic | a couple of times a day |
22:03.01 | dvayanu | thats a lot |
22:03.05 | asymptonic | yes, it is. |
22:03.10 | asymptonic | Hence the requirement. |
22:03.11 | dvayanu | it sounds like testing |
22:03.17 | asymptonic | It would be no big deal if it were once a week or month. |
22:03.22 | asymptonic | No, its feature delivery. |
22:03.31 | asymptonic | Like I said, its a pretty dynamic site. |
22:03.44 | dvayanu | how much do you store in session? |
22:03.45 | asymptonic | Or at least will be for the next couple of months. |
22:03.50 | asymptonic | About 180k of data |
22:03.56 | dvayanu | hell |
22:04.04 | dvayanu | how often does it change? |
22:04.13 | asymptonic | every request basically. |
22:04.22 | asymptonic | And there are frequent AJAX calls |
22:04.22 | dvayanu | read or write? |
22:04.28 | asymptonic | both |
22:04.50 | dvayanu | if it change on every request what is the sense of having it cached? :-) |
22:04.54 | dvayanu | anyway |
22:04.58 | dvayanu | how many webservers? |
22:05.00 | asymptonic | Statistical data. |
22:05.07 | asymptonic | 2 web servers, 2 tomcat servers |
22:05.16 | asymptonic | currently, but that will expand with time |
22:05.26 | dvayanu | and behind? |
22:05.42 | asymptonic | two database servers in warm standby (one active) |
22:06.09 | dvayanu | no business tier? |
22:06.19 | asymptonic | no |
22:07.18 | dvayanu | why? |
22:07.38 | asymptonic | because we're not that kind of site. Its fairly simple logic, but lots of state. |
22:07.54 | dvayanu | and, how many concurrent users and how many requests per second per server |
22:09.56 | dvayanu | approx. values |
22:09.57 | asymptonic | we anticipate around 20000 concurrent users, ~50 req/sec/server |
22:10.03 | asymptonic | Currently its about 1/10th that. |
22:10.45 | dvayanu | ~50 req/sec/server is rather low for 20K users |
22:10.46 | ibot | dvayanu: okay |
22:11.10 | dvayanu | but for 20K users i urgently recommend you to switch to 3 tier |
22:11.16 | dvayanu | without ejb |
22:11.23 | dvayanu | for your current problem |
22:11.54 | dvayanu | the easiest solution would be to have a sessionservice (pojo) remotely, which holds the session, and the local httpsession only contains a cached copy of it |
22:11.55 | asymptonic | whats the advantage to that? |
22:12.13 | dvayanu | the global session identifier is stored in a cookie, so you can always retrieve the global session |
22:13.00 | dvayanu | you need a fair amount of ram, but 50 req per second over net per server won't be a problem, especially since the requests to the global session service are perfectly distributed |
22:13.04 | dvayanu | distributable |
22:13.04 | *** join/#tomcat sanity (n=ian@216-188-231-97.dyn.grandenetworks.net) |
22:13.04 | asymptonic | Sure, but half the reason for using clustering right now is high availability. Wouldn't this necessitate some other sort of clustering then in the business layer? |
22:13.46 | dvayanu | asymptonic: sure, but the business layer is much easier to HA |
22:13.48 | asymptonic | er, replication that is |
22:14.18 | dvayanu | distributable in sense that you can have multiple copies of this service which are adressed by mod userid or something |
22:14.32 | dvayanu | i'm currently running a site with 20K concurrent users |
22:14.54 | dvayanu | well, running is the wrong word, i was the architect :-) |
22:15.10 | asymptonic | It feels like I'm just pushing that problem down a layer. Does the business layer container cluster more easily? |
22:15.48 | dvayanu | if you use a container - no. but why should you use it? |
22:15.51 | asymptonic | So you're suggesting a custom solution |
22:15.51 | dvayanu | lets start from the beginning |
22:16.20 | dvayanu | the advantage of the business layer is that you have one more scalability point |
22:16.32 | dvayanu | clustering db is extremly expensive |
22:16.38 | dvayanu | and it doesn't work usually |
22:16.58 | dvayanu | a clustered DB is not faster than an unclustered one |
22:17.04 | asymptonic | sure |
22:17.11 | dvayanu | business layer is cheap |
22:17.22 | dvayanu | all you need are machines with a lot of memory |
22:17.41 | dvayanu | you get a dual core hp servers with 16GB ram for lets say <10K |
22:17.55 | dvayanu | and this is a lot of ram |
22:18.03 | asymptonic | But I still have the problem of getting session replication working between at least two of these machines |
22:18.33 | dvayanu | not necessary |
22:18.39 | asymptonic | OH? |
22:18.43 | asymptonic | s/H/h |
22:18.54 | dvayanu | if you always can determine which session is on which machine |
22:19.06 | asymptonic | No, the problem there is if a machine fails, the session is lost. |
22:19.32 | asymptonic | Again, its the HA thats important to me. |
22:19.43 | dvayanu | you can have crossed backup service instances |
22:19.47 | dvayanu | with asynchronous updates |
22:19.55 | dvayanu | jms or corba event service |
22:19.58 | asymptonic | Yes, but thats what tomcat clustering gives me as well. |
22:20.06 | asymptonic | So its the same solution in a different place. |
22:20.21 | asymptonic | With the added complexity of another layer. |
22:20.22 | dvayanu | but its more scalable |
22:20.31 | dvayanu | you can put 10 webservers in front |
22:20.34 | asymptonic | Actually, I don't think it is in this case |
22:20.35 | dvayanu | of those 2 machines |
22:21.04 | dvayanu | or 20 |
22:21.10 | asymptonic | I can put 10 web servers in front of my database in this case, as the session has to be reconstructed fron changing global information anyway. |
22:21.42 | asymptonic | All the business layer would be buying me is longer term in RAM storage of temporary state. |
22:21.46 | dvayanu | 10 webservers writing into the db at the same time? |
22:21.56 | dvayanu | thats not fast |
22:21.57 | asymptonic | That has to happen anyway |
22:22.12 | asymptonic | As each of our requests modifies state that must be persisted beyond the session anyway |
22:22.20 | dvayanu | no, since the business layer can manage the connection better |
22:22.33 | asymptonic | Thats irrelevant, the write still has to happen. |
22:22.36 | dvayanu | but thats speculative :-) |
22:22.49 | dvayanu | can you afford to loose some data? |
22:22.52 | dvayanu | like 5 minutes? |
22:23.09 | asymptonic | Usually not, no. In our site, scrolling the browser generates events we must record. |
22:23.31 | dvayanu | 100% availability is not reachable anyway |
22:23.59 | dvayanu | achieveable |
22:24.01 | asymptonic | Of course not. But so far I don't think adding another layer solves any of my problems. |
22:24.25 | dvayanu | it does, you just dont yet see it, or we haven't found the case |
22:24.47 | dvayanu | you could for example persist everything to RAM in the business layer, and make copies once a minute |
22:24.49 | asymptonic | Thanks for the help, but I'm going to bang on clustering some more. We are a very atypical site. |
22:25.16 | dvayanu | or even continuosly but asynchron |
22:25.19 | asymptonic | dvayanu: but the problem is the writes I'm talking about aren't to session data, they're to global data |
22:25.40 | dvayanu | and they have to be visible to other users at once? |
22:25.42 | asymptonic | So either way, they must be written. They're already written asynchronously, so passing them off to a second third party isn't an improvement. |
22:25.48 | asymptonic | no, and they aren't. |
22:25.58 | dvayanu | than you can cache them |
22:26.02 | asymptonic | They are already cached |
22:26.09 | dvayanu | cached by ? |
22:26.11 | dvayanu | tomcat? |
22:26.27 | asymptonic | User specific datastructures |
22:26.45 | dvayanu | you might want to have a look at moskito btw |
22:26.52 | asymptonic | The problem is, each user sees a separate view of the data, so its not enough to maintain state globally. |
22:27.29 | dvayanu | since performance seems to be an issue for you |
22:28.09 | asymptonic | thanks |
22:28.22 | asymptonic | I have to go actually, I may be back later tonight. |
22:28.39 | dvayanu | its 00:30 here, so i doubt i'll be here: -) |
22:28.40 | dvayanu | cu |
22:28.43 | asymptonic | later |
22:28.45 | *** part/#tomcat asymptonic (n=scgmille@216-188-226-118.dyn.grandenetworks.net) |
22:53.37 | *** join/#tomcat sanity (n=ian@cpe-66-25-165-188.austin.res.rr.com) |