20:03:12 #startmeeting better host tracker interns 20:03:12 Meeting started Mon Jul 7 20:03:12 2014 UTC. The chair is colindixon. Information about MeetBot at http://ci.openstack.org/meetbot.html. 20:03:12 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 20:03:12 The meeting name has been set to 'better_host_tracker_interns' 20:03:18 #topic status update 20:03:21 so, what’s up? 20:03:22 :p 20:03:32 updated the databroker stuff 20:03:34 from the deprecated 20:03:55 everything should import to eclipse 20:04:00 at least I get no errors on my laptop 20:04:17 #info we’ve moved to the new data broker interfaces 20:04:41 evanz_ I am not, so you shouldn't have a problem with that 20:04:51 on my side everything works as well 20:05:12 and colindixon: I put in an executor where you mentioned it, but do we need it now that the databroker can do async tx? 20:06:20 …so, yes, you can’t use a DataChangeNotification thread to then do a transaction 20:06:22 you’ll get a deadlock 20:06:24 as I learned 20:06:26 :p 20:07:19 interesting, so my mdhosttracker stuff is broken 20:07:39 so, I’ll just remove it from the root pom file 20:07:42 yeah it can't resolve a bunch of Task types 20:07:48 yeah 20:07:50 that’s my fault 20:08:16 btw if we need to share screen or whatever, I think I've fixed my microphone issues 20:08:35 there is one other person in the lab but I don't think it will bother him 20:09:02 I just pushed the change to the pom file 20:11:00 huh 20:11:04 I get errors... 20:11:13 in Eclipse 20:11:33 I only have errors for mdhosttracker 20:11:45 which btw should we change betterhosttracker to "mdhosttracker" 20:11:46 they’re all @override saying that they must override things 20:11:48 but they appare to be 20:11:48 probably a better name 20:12:39 ok deleting an reimporting fixed it 20:12:50 so, I think mdhosttracker is probably a better name 20:12:59 yeah :) 20:13:08 #action evanz_ or djx to rename betterhosttracker to mdhostracker (at somet point) 20:13:36 #link https://github.com/nilok/betterhosttracker/ the current code just for posterity 20:14:07 #topic next steps 20:14:36 we lost djx 20:14:37 uh oh 20:15:04 the good news is when he gets back the logs are here: https://meetings.opendaylight.org/opendaylight-intern/2014/better_host_tracker_interns/opendaylight-intern-better_host_tracker_interns.2014-07-07-20.03.log.txt 20:15:05 do you guys said anything afte "colindixon> that’s my fault"? 20:15:20 I never saw that :p 20:15:26 the logs so far are at that link 20:15:43 cool 20:15:46 basically just said at some point we will rename betterhosttracker to mdhosttracker 20:15:46 so, the next step I think is to actually wire this up so we can test things 20:16:03 yeah we need to cause a data change event 20:16:06 I think? 20:16:13 #info the next steps are basically to wire this up with something that will generate Addresses objects in the MD-SAL to cause our data change events 20:16:15 yes 20:16:32 the l2switch guys claim their addresstracker is working 20:16:55 ah right 20:17:04 ok, I was trying to build l2switch but didn't work 20:17:26 #info the l2switch project claims to have the address tracker working, I was just able to build the addresst tracker part at least 20:17:51 ah I tried to build the whole project 20:17:51 building l2switch right now 20:17:59 and then, as expected I get implementation/target/addresstracker-impl-0.1.0-SNAPSHOT.jar and model/target/addresstracker-model-0.1.0-SNAPSHOT.jar 20:18:03 I just build the address tracker 20:18:05 nothing else 20:18:15 okay 20:18:22 nop, it fails 20:18:41 djx: huh 20:18:43 error message? 20:18:46 l2switch/addresstracker/model/src/main/yang-gen-code/org/opendaylight/yang/gen/v1/urn/opendaylight/address/tracker/rev140617/address/node/connector/AddressesKey.java:[9,24] type argument org.opendaylight.yang.gen.v1.urn.opendaylight.address.tracker.rev140617.address.node.connector.Addresses is not within bounds of type-variable T 20:19:04 did you do a git pull? 20:19:13 yup 20:19:14 29a220718daaf6ef03f205bfe6c9d918a787a436 20:19:29 ugh 20:19:31 I hadn’t :p 20:19:36 building now 20:19:44 it still built 20:19:46 Reactor Summary: 20:19:46 [INFO] 20:19:47 [INFO] addresstracker-model .............................. SUCCESS [4.284s] 20:19:47 [INFO] addresstracker-impl ............................... SUCCESS [0.305s] 20:19:47 [INFO] addresstracker.aggregator ......................... SUCCESS [0.106s] 20:20:06 what version do you have from git log? 20:20:21 spectre-2:addresstracker ckd$ git br 20:20:21 * master 20:20:23 spectre-2:addresstracker ckd$ git log -1 --pretty=oneline 20:20:24 29a220718daaf6ef03f205bfe6c9d918a787a436 Fix a bug in flooding in sendPacketOut. 20:20:25 spectre-2:addresstracker ckd$ git st 20:20:26 spectre-2:addresstracker ckd$ 20:20:42 do you have any unmerged changes 20:20:43 git status 20:20:53 Your branch is up-to-date with 'origin/master' 20:20:58 nothing to commit, working directory clean 20:21:06 git log -1 --pretty=oneline 20:21:06 29a220718daaf6ef03f205bfe6c9d918a787a436 Fix a bug in flooding in sendPacketOut. 20:22:26 oh no I imported l2switch and it broke betterhosttracker 20:22:45 … 20:22:48 that’s no good 20:22:58 if I close the l2switch project it works 20:23:00 what if you change to l2switch/packethandler and try doing mvn claen install? 20:23:01 :/ 20:23:22 that was for colin 20:23:35 djx: I didn’t change anything 20:23:38 totally clean 20:24:40 mvn clean install right? 20:25:35 yeah 20:26:18 I don't know what might be, I will pull the controller and build it 20:26:52 yeah… that’s weird 20:26:58 ok, I can rename the betterhosttracker to mdhosttracker 20:27:38 basically everything should work on the distribution build of l2switch, no? 20:28:05 #info if we import the l2switch, we get errors in betterhosttracker, I think this is because their pom file sucks and doesn’t list the yang generated source directories correclty 20:28:14 either I or one of you can try to take a stab at that 20:29:00 #info the l2switch project has a distribution, which should work for us, we’ll just need to copy in the two addresstracker jar files and the betterhosttracker jar files 20:29:05 djx: yes 20:30:03 the weirdest part is that I can build the distribution part 20:31:40 what are the TaskProvider.java? 20:31:43 stuff 20:31:52 just ignore it 20:32:02 use the betterhosttracker directories for now 20:32:13 I remvoed the mdhosttracker directory from the pom file so it doesn’t build 20:32:29 if you want to pull the most recent from github 20:33:13 ok, build success 20:33:26 I didn't get the errors evanz_ was talking earlier 20:33:36 which ones? 20:34:02 did you import the l2switch project? 20:34:13 I think he’s talking about the errors if you have both imported into Eclipse 20:34:22 I have imported l2switch and mdhosttracker 20:34:23 yeah I saw that know 20:34:26 now* 20:34:27 forget it 20:34:29 l2switch has no errors in addresstracker 20:34:33 yeah 20:34:42 it just causes errors in betterhosttracker 20:34:44 but betterhosttracker can't find "Addresses" 20:35:01 evanz_ I don't have that issu 20:35:01 and when I run maven in eclipse I get a bunch of dependency "must be unique" errors 20:35:02 e 20:35:26 yeah, and that’s just becaues it knows it *should* be able to find it in the imported bundle, but the bundle’s pom file doesn’t have the generated source directory set up right 20:35:27 I think 20:35:38 ok 20:35:49 I’ve built the l2switch distribution 20:35:53 and I’m now trying to run it 20:36:58 #info just as a heads up, when runnign run.sh, you probably want to add -Xmx2048m -XX:MaxPermSize=512m so “./run.sh -Xmx2048m -XX:MaxPermSize=512m” so that it doesn’t run out of memory 20:37:07 even 2048m is coming up short sometiems 20:37:24 I always used 1G 20:37:46 great 20:37:52 and I run and get a nullpointerexception 20:38:50 with our project? 20:39:00 no when I run the l2switch distribution 20:39:33 ah phew, you should pull from git our project btw 20:40:20 push now 20:41:45 colindixon see if the nullpointexception isn't from l2switch/addresstracker/model/src/main/yang-gen-code/org/opendaylight/yang/gen/v1/urn/opendaylight/address/tracker/rev140617/address/node/connector/AddressesKey.java:[9,24] 20:42:12 …. 20:42:14 let me look 20:42:32 because that's where I'm getting my error while compiling 20:43:14 https://gist.github.com/nilok/0005658677c71bb3a788 20:44:36 packethandler.decoders.EthernetDecoder.decode(EthernetDecoder.java:62) 20:44:56 yep I'm getting another compilation error on packethandler as well 20:45:55 l2switch/packethandler/implementation/src/main/java/org/opendaylight/l2switch/packethandler/decoders/EthernetDecoder.java:[131,16] cannot find symbol 20:48:58 huh 20:49:02 that’s really weird 20:49:18 so you did cd l2switch; git pull; mvn clean install; 20:49:32 yups 20:50:15 I’m doing that now 20:50:47 trying to build everything not just the address tracker 20:51:08 you're saying that to me? 20:51:15 yeah 20:51:22 yeah 20:51:25 that's what I'm trying to do in the last half an hour 20:51:27 I get a failure on main-impl 20:51:39 [INFO] l2switch-parent ................................... SUCCESS [0.178s] 20:51:39 [INFO] packethandler-model ............................... SUCCESS [8.045s] 20:51:39 [INFO] packethandler-impl ................................ FAILURE [0.289s] 20:51:39 [INFO] packethandler.aggregator .......................... SKIPPED 20:51:39 [INFO] addresstracker-model .............................. SKIPPED 20:51:40 [INFO] main-impl ......................................... SKIPPED 20:51:42 [INFO] Opendaylight L2 Switch Base Edition ............... SKIPPED 20:51:44 [INFO] addresstracker-impl ............................... SKIPPED 20:51:46 [INFO] addresstracker.aggregator ......................... SKIPPED 20:51:48 [INFO] l2switch-controller-config ........................ SKIPPED 20:51:50 [INFO] l2switch .......................................... SKIPPED 20:51:51 [INFO] l2switch-parent .................................... SUCCESS [ 0.258 s] 20:51:52 [INFO] packethandler-model ................................ SUCCESS [ 14.559 s] 20:51:54 [INFO] packethandler-impl ................................. SUCCESS [ 2.809 s] 20:51:55 [INFO] packethandler.aggregator ........................... SUCCESS [ 0.109 s] 20:51:57 [INFO] addresstracker-model ............................... SUCCESS [ 1.307 s] 20:51:58 [INFO] main-impl .......................................... FAILURE [ 0.622 s] 20:52:00 [INFO] Opendaylight L2 Switch Base Edition ................ SKIPPED 20:52:00 lol 20:52:01 [INFO] addresstracker-impl ................................ SKIPPED 20:52:03 [INFO] addresstracker.aggregator .......................... SKIPPED 20:52:04 [INFO] l2switch-controller-config ......................... SKIPPED 20:52:05 [INFO] l2switch ........................................... SKIPPED 20:52:06 that’s what I get 20:52:07 so, I make it further 20:52:16 that's not fair 20:53:03 uh 20:53:11 [INFO] Reactor Summary: 20:53:12 [INFO] 20:53:12 [INFO] l2switch-parent ................................... SUCCESS [0.244s] 20:53:12 [INFO] packethandler-model ............................... SUCCESS [9.541s] 20:53:12 [INFO] packethandler-impl ................................ SUCCESS [2.047s] 20:53:13 [INFO] packethandler.aggregator .......................... SUCCESS [0.081s] 20:53:14 [INFO] addresstracker-model .............................. SUCCESS [0.724s] 20:53:16 [INFO] main-impl ......................................... SUCCESS [0.728s] 20:53:18 [INFO] Opendaylight L2 Switch Base Edition ............... SUCCESS [18.822s] 20:53:20 [INFO] addresstracker-impl ............................... SUCCESS [0.363s] 20:53:22 [INFO] addresstracker.aggregator ......................... SUCCESS [0.004s] 20:53:26 [INFO] l2switch-controller-config ........................ SUCCESS [1.204s] 20:53:27 evanz 20:53:28 lol 20:53:28 [INFO] l2switch .......................................... SUCCESS [0.004s] 20:53:30 [INFO] ------------------------------------------------------------------------ 20:53:32 [INFO] BUILD SUCCESS 20:53:33 you won 20:53:59 why is maven so strange 20:54:04 sheesh 20:54:08 I just did an rm -rf ~/.m2/ 20:54:10 should I delete the folder and pull everything? 20:54:13 and I’m trying to rebuild 20:54:16 and see what happens 20:54:16 omg... 20:54:16 yeah rm .m2 is a good idea 20:54:23 well, it’s a shitty idea 20:54:27 but sometimes it’s the only one that works 20:54:32 lol o 20:54:35 rm .m2 is like reinstalling the OS 20:54:42 maybe mine only builds because theres something it needs sitting in .m2 20:54:45 that will be gone if I delete it 20:55:48 P.S. they still don't play nicely in eclipse 20:55:51 colin tell me if that works because I don't like deleting .m2 20:56:30 yeah 20:56:42 I’m working on it 20:56:49 no 20:56:54 mine fails in the same place 20:57:40 running distribution 20:58:58 thanks colin for sending that email 21:00:04 yeah 21:02:09 well the controller didn't exactly work 21:02:34 evanz_ what happen? 21:03:20 not sure 21:03:43 do you get exceptions 21:03:51 I wonder how it is that you can build it evan 21:03:54 are you on Linux? 21:03:57 isn’t djx also Linux 21:03:58 indeed 21:04:05 I'm on linux 21:04:10 djx is on fedora? 21:04:12 is that right? 21:04:18 yup 21:04:26 #info we’re running into problems building and/or running the l2switch project 21:04:37 just to be clear 21:04:43 mvn --version 21:04:44 Apache Maven 3.0.5 (r01de14724cdef164cd33c7c8c2fe155faf9602da; 2013-02-19 13:51:28+0000) 21:04:44 Maven home: /usr/share/maven 21:04:44 Java version: 1.7.0_51, vendor: Oracle Corporation 21:04:48 Java home: /usr/java/jdk1.7.0_51/jre 21:04:52 Default locale: en_US, platform encoding: UTF-8 21:04:56 OS name: "linux", version: "3.14.9-200.fc20.x86_64", arch: "amd64", family: "unix" 21:05:01 mvn -v 21:05:01 * djx should update it's java version 21:05:02 Apache Maven 3.2.2 (45f7c06d68e745d05611f7fd14efb6594181933e; 2014-06-17T08:51:42-05:00) 21:05:03 Maven home: /usr/local/Cellar/maven/3.2.2/libexec 21:05:05 Java version: 1.7.0_51, vendor: Oracle Corporation 21:05:06 Java home: /Library/Java/JavaVirtualMachines/jdk1.7.0_51.jdk/Contents/Home/jre 21:05:07 Default locale: en_US, platform encoding: UTF-8 21:05:08 OS name: "mac os x", version: "10.9.4", arch: "x86_64", family: "mac" 21:05:11 evan@evan-ThinkPad ~/git_reps/l2switch/distribution/base/target/distributions-l2switch-base-0.1.0-SNAPSHOT-osgipackage/opendaylight $ mvn --version 21:05:12 Apache Maven 3.0.4 21:05:12 Maven home: /usr/share/maven 21:05:12 Java version: 1.7.0_25, vendor: Oracle Corporation 21:05:12 Java home: /usr/lib/jvm/java-7-openjdk-amd64/jre 21:05:13 Default locale: en_US, platform encoding: UTF-8 21:05:15 OS name: "linux", version: "3.8.0-35-generic", arch: "amd64", family: "unix" 21:05:44 well, we have different java versions, and different maven versions 21:06:00 worst 21:06:22 evanz_ I would advice you to work with proprietary java 21:07:30 colin but this shouldn't be a problem 21:08:10 should I rm .m2 21:08:15 nope 21:08:16 well 21:08:33 mv ~/.m2 ~/old.m2 21:08:36 or something 21:08:40 so you can move it back 21:09:09 #info colindixon e-mailed the l2swtich team to ask for help with the fact that things aren’t compiling 21:09:17 ok, I’m going to have to call it a day for this 21:09:21 :-/ 21:10:05 does the addresstracker depend on anything that’s not building 21:10:06 ? 21:10:10 #topic fallback plans 21:10:40 #info so, while we wait for them to fix the l2switch stuff, we could take the base distribution from openflowplugin and use it 21:10:59 #info we’ll then need to copy in the address tracker jars, their dependencies and the host tracker jars 21:11:05 I can't build the addresstracker too 21:11:30 #info and see if we can get that to work and generate data change events 21:11:44 yes 21:12:14 #info I recommend using log.error() or system.out.println() messages to figure out if you’re code is hitting because the logging is often configured to only print error messages and above 21:12:59 ugh I have no idea why stuff isn’t building 21:13:01 the logger is there ;) 21:13:12 ok, well lets touch base with things again tomorrow morning 21:13:20 maybe djx will wake up having a fix 21:13:24 :-) 21:13:27 #endmeeting