00:23.07 | *** part/#tomcat alx- (~alx-@69.165.240.218) |
00:53.36 | *** join/#tomcat revattar (~revattar@c-98-248-187-174.hsd1.ca.comcast.net) |
01:35.04 | *** join/#tomcat acidjnk22 (havenone@p4FDFDA25.dip.t-dialin.net) |
02:10.13 | *** join/#tomcat robottinosino_ (~robottino@255.Red-2-138-158.dynamicIP.rima-tde.net) |
02:38.02 | *** join/#tomcat revattar1 (~revattar@c-98-248-187-174.hsd1.ca.comcast.net) |
03:28.59 | *** join/#tomcat revattar (~revattar@c-98-248-187-174.hsd1.ca.comcast.net) |
04:03.23 | *** join/#tomcat nevada-smith (~nevada-sm@S0106c8fb2641d848.vs.shawcable.net) |
04:20.14 | *** join/#tomcat revattar (~revattar@c-98-248-187-174.hsd1.ca.comcast.net) |
05:19.24 | *** join/#tomcat vigneshwaran (~vigneshwa@182.71.239.158) |
05:40.05 | *** join/#tomcat rhonabwy (~rhonabwy@95.157.57.154.dynamic.cablesurf.de) |
06:08.15 | *** join/#tomcat chkal (~chkal@217.194.235.13) |
06:12.27 | *** join/#tomcat papegaaij (~papegaaij@topicus-leeuwenbrug-33-deventer.xs4all.nl) |
06:15.28 | *** join/#tomcat andre2g (~andre2g@dslb-088-074-218-068.pools.arcor-ip.net) |
06:28.36 | *** join/#tomcat mturk (~mturk@18-216.dsl.iskon.hr) |
06:28.37 | *** join/#tomcat mturk (~mturk@redhat/jboss/mturk) |
06:30.00 | *** join/#tomcat vigneshwaran (~vigneshwa@182.71.239.158) |
07:08.55 | *** join/#tomcat alcir (~alcir@fw.ftgm.it) |
07:34.14 | *** join/#tomcat _Qman (~Qman@213.177.160.238) |
07:45.01 | *** join/#tomcat jnhghy (~jalexandr@86.123.191.6) |
09:41.49 | *** join/#tomcat mturk (~mturk@18-216.dsl.iskon.hr) |
09:41.50 | *** join/#tomcat mturk (~mturk@redhat/jboss/mturk) |
09:44.09 | *** join/#tomcat Reddy (~laxman@118.102.128.225) |
10:22.40 | _Qman | Hello, can we configure Tomcat 6.0.36 to authenticate against multiple different OU with LDPA? |
10:22.54 | *** join/#tomcat hzilla (~hzilla@41.178.136.196) |
10:23.49 | hzilla | hello. i'm wasting my time here trying to get something as basic as app authentication working in tomcat7. I've got a tomcat installation at /etc/tomcat7 and tomcat-users.xml in that dir along with server.xml |
10:23.52 | _Qman | meaning a user may be located in ou=Users,dc=my,dc=domain,dc=com when another is located in ou=italy,dc=my,dc=domain,dc=com |
10:24.33 | hzilla | tomcat-users.xml has a role and user specified |
10:25.42 | hzilla | i'll pastie it |
10:31.49 | hzilla | http://pastie.org/7598242 - first problem i don't understand is taht it doesn;t look like the users file is loaded. i've garbled the xml in the file to test if the server restart fails but it starts fine |
10:32.04 | hzilla | please help |
12:15.34 | *** join/#tomcat Karel| (~Karel@94-226-237-238.access.telenet.be) |
12:28.24 | *** join/#tomcat xortim (~tim@167.246.61.8) |
12:30.15 | Karel| | what is the best way to do a proper analysis of the performance of a tomcat servlet? |
12:30.32 | Karel| | I have a servlet that quickly fills up the available RAM, and then drops back to practically 0 afterwards, every few minutes |
12:30.38 | Karel| | I assume that's because of garbage collection |
12:31.51 | xortim | i'm not a tomcat dev, but that's very likely to be garbage collection |
12:32.14 | xortim | try using JMeter or installing NewRelic (they have a trial) to get a better look at what's going on in there |
12:33.27 | Karel| | I use JMeter to stress test the servlet to get that behaviour |
12:33.43 | xortim | try tossing newrelic into the mix |
12:33.44 | Karel| | the only thing it tells me is raw memory, load, thread numbers |
12:33.51 | Karel| | alright, I'll check it out |
12:34.07 | xortim | another option is dynatrace |
12:34.15 | xortim | not sure if there's a trial available for that on |
12:34.17 | xortim | *one |
12:36.02 | xortim | also, depending on the JVM you're using you can turn on verbose garbage collection which should vomit out a lot more data to native_stderr.log |
12:48.37 | *** join/#tomcat pradhand (~pradhand@142.177.89.39) |
12:55.13 | *** join/#tomcat pradhand_ (~pradhand@142.177.122.31) |
12:57.18 | Karel| | wow xortim |
12:57.22 | Karel| | that newrelic program is like magic |
12:57.26 | Karel| | it tells me everything |
12:57.51 | *** join/#tomcat demetan (~chatzilla@gw1.ciep.fr) |
12:57.52 | *** join/#tomcat demetan (~chatzilla@unaffiliated/demetan) |
13:01.45 | *** join/#tomcat pradhand_ (~pradhand@142.177.121.208) |
13:02.51 | xortim | glad it could help |
13:03.07 | xortim | looks like the lite version helps mostly with triage |
13:03.25 | xortim | unfortunately there's no instrumentation profiling that's FOSS |
13:06.45 | xortim | and try out some of the oracle tools for java analysis http://docs.oracle.com/javase/6/docs/technotes/tools/share/jhat.html |
13:07.04 | xortim | http://docs.oracle.com/javase/7/docs/technotes/samples/hprof.html |
13:07.29 | *** join/#tomcat pradhand_ (~pradhand@142.177.41.11) |
13:10.51 | *** join/#tomcat demetan (~chatzilla@gw1.ciep.fr) |
13:10.52 | *** join/#tomcat demetan (~chatzilla@unaffiliated/demetan) |
13:11.15 | xortim | http://visualvm.java.net/ might also be useful if you're worried about giving newrelic your data |
13:11.30 | Karel| | so far I haven't encountered any limitations to what I can do with the free version |
13:11.48 | xortim | data retention and where the data is stored has been the biggest point of contention with my clients |
13:25.20 | Karel| | ah |
13:32.03 | *** join/#tomcat pradhand (~pradhand@142.177.89.99) |
13:50.01 | *** join/#tomcat _Qman (~Qman@213.177.160.238) |
13:57.53 | *** join/#tomcat pradhand (~pradhand@142.177.90.114) |
14:16.24 | LzrdKing | ok, i have two idle identical tomcats on two identical servers, so why is one almost continuously GCing? |
14:17.06 | LzrdKing | Viaual GC just tells me "Last Cause: unknown GCCause" |
14:17.21 | LzrdKing | Visual |
14:17.44 | xortim | under the same load? (check the load balancer) settings. Also check that the JVM settings are the same on both |
14:18.34 | LzrdKing | load is 1 on one and 1.5 on the other, identical JVM settings |
14:19.01 | LzrdKing | the one with the 1.5 load is the one thats GCing so much |
14:19.46 | xortim | might be a silly question but is the code the same on each? |
14:20.01 | LzrdKing | there is no code |
14:20.08 | LzrdKing | oh wait, maybe one has the silly default apps |
14:20.30 | xortim | code meaning the webapps, deployable artifacts, etc. |
14:20.42 | xortim | mdf5sum ${tomcat_home}/webapps/*.war |
14:20.46 | LzrdKing | no, that wasn;t it, there is only the root app |
14:20.50 | LzrdKing | no wars |
14:21.05 | xortim | you have tomcat idling on both servers? |
14:21.06 | LzrdKing | thats what i meant by idle, empty |
14:21.11 | LzrdKing | yup |
14:21.34 | LzrdKing | the one is gcing almost exactly once an hour |
14:22.08 | xortim | best bet is to look at the garbage collection logs, heap and core dumps |
14:22.24 | LzrdKing | the other is like every 30 seconds |
14:22.26 | xortim | http://wolfe.id.au/2011/10/16/monitoring-the-openjdk-from-the-cli/ |
14:22.39 | xortim | that has great resources |
14:22.58 | xortim | i believe one of the tools allows you to attach to a given pid and see what's going on |
14:23.34 | LzrdKing | i have the tools |
14:25.06 | Karel| | alright, so I pinpointed the major source of CPU usage in my tomcat to DriverManagerDataSource.getConnection(), which initializes the database connection |
14:25.10 | xortim | not sure. if it's just a basic installation of tomcat on two identical servers (same OS level, Java Version, architecture, physical resource such a RAM, etc.) then check to make sure that there's nothing else running that's preventing java from allocating the same amount of ram |
14:26.02 | xortim | Karel| - that's seems weird. is it not reusing connections? |
14:26.15 | Karel| | hmm it probably is not |
14:26.41 | Karel| | you got experience with spring? |
14:26.51 | xortim | not really, no |
14:26.59 | Karel| | Simple implementation of the standard JDBC DataSource interface, configuring a plain old JDBC Driver via bean properties, and returning a new Connection for every getConnection call. |
14:26.59 | Karel| | NOTE: This class is not an actual connection pool; it does not actually pool Connections. It just serves as simple replacement for a full-blown connection pool, implementing the same standard interface, but creating new Connections on every call. |
14:27.04 | xortim | mostly a middleware admin |
14:27.04 | Karel| | that might explain it |
14:27.18 | LzrdKing | ok, the one that is GCing more frequently has committed 8MB to eden, the other has committed 62GB |
14:27.31 | xortim | eden? |
14:27.39 | LzrdKing | new gen |
14:28.35 | LzrdKing | max is 1.3GB on each, i'm letting it autotune the heaps so it seems it has done so very differently on each |
14:28.47 | *** join/#tomcat jieryn (~jieryn@pool-173-77-215-221.nycmny.fios.verizon.net) |
14:28.48 | *** join/#tomcat jieryn (~jieryn@jenkins/developer/jieryn) |
15:04.42 | *** join/#tomcat revattar (~revattar@c-98-248-187-174.hsd1.ca.comcast.net) |
15:14.13 | *** join/#tomcat gcristian (~havok@mail.bixti.com.ar) |
15:14.50 | *** join/#tomcat AJC_Z0 (~AJ_Z0@pdpc/supporter/professional/ajz0) |
15:23.55 | LzrdKing | i restarted both tomcats with -Xmn64m and they both seem a lot better |
15:33.03 | _Qman | Hello, anyone can assist with setting LDAP authentication in Tomcat where the users are spread around multiple OU? |
15:39.37 | xortim | multiple OUs are usually tricky no matter what. set the search base to the lowest common denominator and come up with a reasonable LDAP query for the OUs (|(ou=one)(ou=two)) |
15:39.40 | xortim | so on and so forth |
15:40.40 | xortim | test your query against LDAP with JXExplorer or ApacheDS |
15:41.06 | xortim | get the results you want and make them work with the tomcat config |
15:48.25 | *** join/#tomcat pradhand (~pradhand@142.177.121.38) |
15:48.34 | LzrdKing | no, i take that back, the 01 server is STILL GCing more frequentrly than 2, but a lot less often than before |
15:57.32 | *** join/#tomcat pradhand_ (~pradhand@142.177.17.199) |
16:05.44 | LzrdKing | 7 collections in the time it took the other to have 3 |
16:09.59 | LzrdKing | ooh, MaxThreads is different on the http conenctor, could that be part of the difference? |
16:10.44 | whartung | is the traffic at all different between the two servers LzrdKing |
16:11.34 | LzrdKing | there is no traffic |
16:11.42 | LzrdKing | except jmx |
16:11.58 | whartung | so this is just idle GCing? |
16:12.02 | LzrdKing | yup |
16:12.22 | LzrdKing | i can see eden filling up faster on the one in visualgc than the toher (and zabbix, too |
16:12.27 | LzrdKing | other* |
16:15.42 | *** join/#tomcat dknox (~dknox@66.109.209.21) |
16:16.55 | LzrdKing | i set both to the same number of threads, lets see if that matetrs |
16:16.58 | LzrdKing | matters* |
16:17.42 | whartung | it'll matter perhaps "a little", it depends on whether those threads wake up at all. If you had "twice as many" and they wake up regularly to "do nothing", then you'd have twice as much garbage. |
16:18.01 | LzrdKing | one was 200, the other was 450 |
16:18.35 | whartung | that could explain it. Mind, I think your thread numbers are ..um.. insane, but I don't know what kind of traffic you guys drive. |
16:18.55 | LzrdKing | isn't 200 the default? |
16:19.06 | whartung | I dunno -- seems awfully high to me. |
16:20.10 | whartung | My T7 install has the setting commented out. |
16:20.41 | LzrdKing | what, MaxThreads? |
16:20.46 | whartung | yes |
16:20.58 | LzrdKing | that is is using the default for sure :) |
16:21.01 | LzrdKing | it is* |
16:21.30 | whartung | well the commented out Executor tag has 150 assigned. I doubt the default is more than that, but I can't say. |
16:21.43 | whartung | I'd assume it was more like 10 |
16:22.03 | LzrdKing | blah, frist gc is about to happen on 01, 02 still has a way to go |
16:22.15 | *** join/#tomcat revattar (~revattar@50-0-117-34.dedicated.static.sonic.net) |
17:09.20 | LzrdKing | i think its looking good now... |
17:18.11 | LzrdKing | hah! 01 just GCed, but 02 hasn;t, though its ABOUT to |
17:20.04 | LzrdKing | or... not |
17:24.05 | LzrdKing | ok, it went |
17:25.07 | xortim | why are you so concerned about the appserver GCing? that's usual behavior for the JVM |
17:25.38 | xortim | if you're not geting out of memory exceptions and/or cpu starvation, it's suppose to GC |
17:33.41 | LzrdKing | i know, but why is one GCing more frequently than another identical one? |
17:34.18 | LzrdKing | 01 just GCed again, its got 5 to 02's 3 |
17:34.22 | xortim | one is busier than the other ;o) |
17:34.36 | LzrdKing | no, they are both equally idle |
17:34.59 | xortim | diff the server.xml |
17:35.06 | xortim | and diff the configuration files |
17:35.13 | LzrdKing | they are identival not; one had more threads before but i changed that |
17:35.17 | LzrdKing | identical now* |
17:36.08 | xortim | what OS? |
17:36.31 | LzrdKing | rhel 5.5 |
17:37.33 | xortim | using the tomcat package from RHN or from the apache foundation? and also which Java versions are you using? |
17:37.40 | xortim | are the ulimits the same for each user? |
17:37.52 | LzrdKing | using a tarball from apache |
17:39.11 | LzrdKing | same jvm, same ulimits |
17:39.25 | LzrdKing | java version "1.6.0_22" |
17:39.26 | LzrdKing | Java(TM) SE Runtime Environment (build 1.6.0_22-b04) |
17:39.26 | LzrdKing | Java HotSpot(TM) 64-Bit Server VM (build 17.1-b03, mixed mode) |
17:40.57 | xortim | same kernel versions? |
17:41.15 | LzrdKing | indeed |
17:44.16 | xortim | well. I don't think it's something to lose any sleep over. It sounds like you've done some settings changes between the two and are wondering why one is different than the other. |
17:44.24 | xortim | try taking coredumps (kill -3) |
17:44.33 | xortim | and see what's running at the time of the GC |
17:45.46 | LzrdKing | 01 has a few more threads |
17:47.31 | xortim | which threads? |
17:47.55 | LzrdKing | live and daemon |
17:49.38 | xortim | what's the command being executed for each? |
17:49.40 | xortim | ps -ef | grep tomcat |
17:51.59 | LzrdKing | svcgeo 19960 19959 0 12:55 ? 00:00:06 jsvc.exec -user svcgeo -home /usr/java/default -Dcatalina.home=/opt/geocoding/tomcat-geocode -Dcatalina.base=/opt/geocoding/tomcat-geocode/ -Djava.io.tmpdir=/var/tmp -wait 10 -pidfile /var/run/tomcat.geocode.pid -outfile /opt/geocoding/tomcat-geocode/logs/catalina.out -errfile &1 -Xmx4096m -XX:MaxPermSize=512m -Dcom.sun.management.jmxremote -Djava.rmi.server.hostname=serverFQDN -Dcom.sun.management. |
17:52.08 | LzrdKing | damn, i'm sure that got cut off |
17:52.15 | LzrdKing | but in any case, they are identical |
17:54.11 | LzrdKing | tomcat-users.xml has a user on one, but no manager application for it to matter |
17:55.19 | LzrdKing | the 02 server hasn't GCed since 1:15, but the 02 has GCed 2 more times since then |
17:57.08 | LzrdKing | oooh they both just full-GCed, maybe that will help the one |
17:57.23 | LzrdKing | doesn;t look it |
17:58.31 | LzrdKing | well, i will continue to monitor but i won;t "lose any sleep over" it |
18:31.47 | LzrdKing | where do i get jsvc from? |
18:32.33 | xortim | http://commons.apache.org/proper/commons-daemon/jsvc.html ? |
18:32.44 | whartung | it's a source tar ball in the bin directory of the binary distribution |
18:32.55 | LzrdKing | yeah, i'm not seeing it |
18:33.16 | LzrdKing | commons-daemon-native.tar.gz? |
18:33.22 | whartung | yes |
18:33.31 | LzrdKing | i'd prefer to not compile it |
18:33.44 | whartung | guess you're out of luck then |
18:34.05 | whartung | have you tried it? |
18:34.06 | LzrdKing | is tomcat 7's very different than tomcat 6's? because i could just copy over the tomcat 6 version |
18:34.13 | whartung | I imagine they're identical |
18:35.24 | LzrdKing | also, http://www.apache.org/dist/commons/daemon/binaries/ |
18:35.59 | whartung | binaries for what platforms? |
18:36.32 | LzrdKing | linux |
18:36.40 | LzrdKing | but yeah, thats not the right thing |
18:46.23 | xortim | define "thing" |
18:47.07 | LzrdKing | executable |
18:48.11 | xortim | how so? |
18:48.16 | xortim | they're executable jar files |
18:48.57 | LzrdKing | what i copied from a tomcat 6 box (jsvc and commons-daemon.jar) isn;t working |
18:49.09 | LzrdKing | jsvc.exec error: Service exit with a return value of 3 |
18:49.13 | xortim | or it's a compiled library for you to use in your application |
18:49.48 | xortim | check the javadoc/usage for this particular project |
18:50.08 | whartung | jsvc is not java xortim |
18:50.14 | LzrdKing | i have a script that calls it |
18:50.17 | whartung | it's a c program to run java as a service |
18:50.32 | xortim | looks like a bunch of jar files to me ;o) |
18:50.33 | whartung | I've never used it |
18:50.48 | xortim | why not use tanuki java service wrapper? |
18:50.59 | *** join/#tomcat jahor (~jahor@89.185.248.105) |
18:51.42 | LzrdKing | why not do a lot of things? |
18:54.00 | xortim | yeah, it looks that commons-daemon that you linked to isn't jsvc |
18:54.55 | xortim | Daemon is made of 2 parts. One written in C that makes the interface to the operating system and the other in Java that provides the Daemon API. |
18:55.13 | xortim | so like i said. that commons-daemon is a series of jar files for you to use in your application |
18:55.30 | xortim | and the other part is procrun (win) or jsvc (*nix) |
18:56.51 | LzrdKing | so where do i get the jsvc part from? |
18:57.23 | whartung | why don't you just build it from the tar.gz file? |
18:57.42 | xortim | you build it |
18:58.26 | LzrdKing | i don;t have a build env |
18:58.44 | xortim | do you have a linux machine? |
18:59.26 | LzrdKing | yes, i suppose i could do it there |
19:05.37 | LzrdKing | fgah, have to install gcc and friends |
19:06.59 | LzrdKing | why don't jsvc binaries exist? |
19:08.46 | xortim | they do. once they're built ;o) |
19:09.22 | xortim | my speculation is that it wasn't designed to exist as a standalone product |
19:09.22 | LzrdKing | 1) thats not what i mean, 2) you know what i mean, and 3) i can't build jsvc because the java path is different on the servers than on my workstation |
19:09.43 | xortim | rather as something that can be included with a particular product |
19:10.14 | LzrdKing | its pissing me off |
19:10.28 | xortim | do you not have a valid RHN subscription? |
19:12.16 | LzrdKing | who said? |
19:13.14 | xortim | why not install gcc? |
19:13.29 | *** join/#tomcat yassine (~yassine@unaffiliated/yassine) |
19:13.31 | LzrdKing | on a server? |
19:13.40 | LzrdKing | now you're just being silly |
19:13.41 | xortim | remove it when you're done |
19:13.57 | *** join/#tomcat franck (~franck@tibco-5.tibco.com) |
19:14.16 | LzrdKing | i just made a link so i could specify the same location for java |
19:16.10 | LzrdKing | /usr/local/prod/server/tomcat-proxy/bin/jsvc: /lib64/libc.so.6: version `GLIBC_2.14' not found (required by /usr/local/prod/server/tomcat-proxy/bin/jsvc) |
19:16.18 | LzrdKing | bugger |
19:16.31 | xortim | need the dev libraries |
19:17.10 | LzrdKing | its already compiled |
19:18.52 | LzrdKing | Package glibc-devel-2.5-81.el5_8.1.x86_64 already installed and latest version |
19:19.18 | LzrdKing | i guess i'll have to put gcc on the server |
19:25.46 | LzrdKing | wtf! the point of a binary distro was to avoid crap like "failed loading capabilities library -- /usr/lib/libcap.so: cannot open shared object file: No such file or directory." |
19:26.20 | xortim | not really |
19:26.33 | xortim | you just don't have all the libraries installed to compile it |
19:26.37 | xortim | same as everything else |
19:26.40 | LzrdKing | no, it compiled fine |
19:26.53 | LzrdKing | twice now, once on my local machine and once on the server; problems both times |
19:26.58 | LzrdKing | runtime problems, that is |
19:27.03 | xortim | ldd /path/to/jsvc |
19:27.12 | LzrdKing | yum install jsvc |
19:27.22 | xortim | so there was a package for it all along? |
19:27.38 | LzrdKing | no, i wish |
19:27.42 | LzrdKing | why isn;t there? |
19:28.00 | xortim | because of situations like you're having |
19:28.20 | xortim | if it requires the path to java installed, which can vary between applications and distribution methods. you run into issues with it |
19:28.29 | LzrdKing | oh jeez |
19:28.31 | xortim | that said. try tanuki java service wrapper |
19:28.32 | LzrdKing | are you kidding me? |
19:28.39 | LzrdKing | redhat has a standar java location |
19:28.53 | xortim | yet yours are different between the two servers? |
19:29.01 | LzrdKing | no |
19:29.09 | LzrdKing | fedora desktop, rhel server |
19:29.31 | LzrdKing | so ldd gave me 5 libraries each with an address |
19:29.51 | xortim | oh, i misunderstood your previous comment about java locations then |
19:30.02 | xortim | ldconfig run recently? |
19:30.10 | LzrdKing | dunno |
19:30.16 | xortim | permission to that library? |
19:30.19 | LzrdKing | its not even looking for libcap.so |
19:30.26 | xortim | hit the low hanging fruit first |
19:30.32 | xortim | so that's why it compiled properly |
19:30.41 | xortim | it wasn't required to building, but required during runtime |
19:30.48 | LzrdKing | no, i mean ldd isn;t looking for it |
19:31.23 | xortim | yeah, means that jsvc wasn't compiled with a link to that library, but it's trying to use it anyway |
19:31.26 | LzrdKing | goddamn! its at /lib/libcap.so |
19:31.35 | xortim | booom |
19:31.49 | xortim | the ./configure should have options for specifying the library directory |
19:32.14 | LzrdKing | nice to see compiling code has gotten HARDER since 1996 |
19:32.23 | xortim | it really hasn't ;o) |
19:32.37 | LzrdKing | i never had this much trouble back then |
19:32.53 | xortim | it's always a pain because the author probably don't know about or give a crap about where the various spots are, defaults will be for his/her system |
19:33.13 | xortim | hence the ./configure options |
19:33.58 | LzrdKing | why am i bothering? failed loading capabilities library -- /usr/lib/libcap.so: wrong ELF class: ELFCLASS32. |
19:34.48 | xortim | ./configure --help |
19:35.14 | xortim | and see if it's even 64bit compatible, otherwise you'll need to install the ia32-libs and point it there |
19:35.44 | xortim | or vis-versa, not sure how your system is set up |
19:36.56 | LzrdKing | doesn;t say |
19:38.03 | xortim | uname -p |
19:38.46 | LzrdKing | well, we KNOW the sierver is 64 bit |
19:38.48 | LzrdKing | server |
19:39.10 | xortim | so see if there's a ./configure option to compile it for 64bit |
19:43.46 | LzrdKing | there is not |
19:43.51 | LzrdKing | this is total BS |
19:44.16 | LzrdKing | tomcat will run on 8080, i don;t even need this |
19:45.36 | xortim | i believe they still recommend running apache infront of tomcat |
19:46.55 | LzrdKing | ok, still no need to start as rioot and drop privs, like jsvc lets you; i can just start tomcat |
20:07.58 | LzrdKing | there, now tomcat starts up; damn that jsvc crap |
20:34.40 | LzrdKing | is the only way to monitor two different JVMs on on server to use two JMX ports? |
20:35.15 | LzrdKing | on one server |
20:36.26 | whartung | I would imagine LzrdKing how else could it be done? |
20:37.09 | LzrdKing | something other than the port |
20:37.20 | xortim | O_o |
20:37.46 | *** join/#tomcat namredips (~namredips@76-224-212-245.lightspeed.iplsin.sbcglobal.net) |
20:37.54 | LzrdKing | since i had to submit a firewall request to have the port opened... i would like to be able to monitor an arbitrary number of jvms through the one port |
20:38.37 | *** join/#tomcat revattar (~revattar@50-0-117-34.dedicated.static.sonic.net) |
20:39.01 | xortim | you always use ssh tunnels |
20:39.12 | xortim | *could always |
20:39.30 | *** join/#tomcat bkfitz (~bkfitz@218.sub-70-197-162.myvzw.com) |
20:40.44 | LzrdKing | what set up is better, to have multiple tomcats/JVMs eachrunning one or two apps, or to throw all the apps into one tomcat/JVM? |
20:41.05 | LzrdKing | well, maybe more than 2 apps |
20:41.06 | whartung | depends |
20:41.15 | LzrdKing | in what? |
20:42.11 | xortim | vertical scaling isn't really necessary with 64bit |
20:42.13 | whartung | how often they change, what impact the necessity to reboot the server because of some change in one app, even if the others have not. Need to balance CPU loads between apps |
20:42.41 | whartung | benefits are lower memory footprint. |
20:43.07 | LzrdKing | so basically more tomcats means greater flexibility, but more memory is needed? |
20:43.58 | xortim | that a bit of an over simplification. it depends on your needs |
20:44.33 | whartung | well you also have to deal with port conflicts and such, but you could mitigate that with extra IPs and interfaces if that was an issue. |
20:44.35 | xortim | if you have a front end app that gets a code deploy very frequently and a backend app that cannot go down outside of maintenance windows, then you'll want two tomcats |
21:01.44 | LzrdKing | got it |
21:03.03 | LzrdKing | so here's whats going on with these two "identical" tomcats: one has done a full GC once per hour for the past 3 hours and thats it; the other has done a full gc at the same times, but also a scavenge 3-4 times in the hour BEFORE the full GC |
21:03.22 | LzrdKing | by full GC, i mean a Scavenge AND a MarkSweep |
21:25.39 | *** join/#tomcat clajo04 (~clajo04@pool-96-224-200-185.nycmny.fios.verizon.net) |
21:56.39 | *** join/#tomcat AJ_Z0 (~AJ_Z0@pdpc/supporter/professional/ajz0) |
22:42.51 | *** join/#tomcat nevada-smith (~nevada-sm@S0106c8fb2641d848.vs.shawcable.net) |
22:54.58 | *** join/#tomcat bzb (~bzb@108.175.226.248) |