Log of the #duraspace-ff channel on chat.freenode.net

Using timezone: Eastern Standard Time
* ksclarke joins01:29
* ksclarke leaves02:35
* scossu joins05:55
* scossu leaves
* scossu joins08:08
* scossu leaves
* dwilcox joins08:28
* tecoripa joins08:33
* tecoripa leaves
* benpennell2 joins09:11
* dwilcox leaves09:14
* dwilcox joins09:20
* tecoripa joins09:27
* tecoripa leaves09:28
* ajs6f joins10:12
* ajs6f1 joins10:14
* ajs6f leaves10:17
<pivotal-bot____>Andrew Woods edited "Separate Java constants now found in fcrepo-jcr into a new artifact" https://www.pivotaltracker.com/story/show/6317248210:18
* ajs6f1 leaves
* ajs6f joins10:19
* ajs6f1 joins
* mikeAtUVa joins10:20
* ajs6f leaves10:21
* ajs6f1 leaves10:24
* ajs6f joins
* ajs6f1 joins10:25
* ajs6f2 joins10:27
* ajs6f1 leaves10:28
* ajs6f leaves10:29
* ajs6f1 joins
* ajs6f2 leaves10:32
* ajs6f joins10:34
* ajs6f2 joins10:35
* ajs6f1 leaves
* fasseg joins10:36
* ajs6f leaves10:38
* ajs6f joins10:39
* ajs6f2 leaves10:43
* tecoripa joins10:48
* ajs6f leaves10:51
<pivotal-bot____>Andrew Woods edited "Demonstrate that we can predictably scale up a cluster to meet performance goals." https://www.pivotaltracker.com/story/show/6550515410:55
* ajs6f joins
<pivotal-bot____>Andrew Woods edited "Improve event reporting facilities for node property changes" https://www.pivotaltracker.com/story/show/6660307210:58
<benpennell2>https://wiki.duraspace.org/display/FF/Single-Node+Test+Results#Single-NodeTestResults-TransactionalversusNon-transactionalTestsResults%28Benchtool%2911:17
<fasseg>ajs6f: here's the ticket: https://www.pivotaltracker.com/s/projects/684825/stories/6651423611:19
<ajs6f>fasseg: thnx
<benpennell2>https://docs.google.com/spreadsheet/ccc?key=0Akmxs4dcuisQdG1jUW1JZ3ltUHYzclhoWUhpZHV4VGc&usp=sharing11:21
<awoods>gregjansen/tecoripa: Is there a plan to move the sprater puppet-scripts into the main git repo?11:29
<tecoripa>awoods: yes, I can do that
<awoods>tecoripa: is it ready?
<benpennell2>awoods: it looks like all the objects created by benchtool all end up at objects/, so perhaps updating benchtool to distribute its entries into subcontainers by run should be another ticket11:34
<awoods>benpennell2: sounds right
benpennell2: can you put in a ticket (slated for a subsequent sprint), but then work with daines on the AuthZ use case tickets?11:39
<benpennell2>sure
<ajs6f>awoods: 12:30— Hangout again?11:41
<awoods>ajs6f: perfect
<ajs6f>Really? "Perfect"? Isn't that quite a high bar? Shouldn't we be a little more conservative and say, "Really good." or "That'll do fine."11:42
Perfect is rare.
awoods: I'm gong to do the property event ticket.11:43
Because it's really straightforward and will make me feel better about getting nothing done last week.
<awoods>ajs6f: great. according to pivotal, you have already started.11:44
<ajs6f>I've started lots of things. Many, many things.
<pivotal-bot____>Benjamin Pennell added comment: "https://docs.google.com/spreadsheet/ccc?key=0Akmxs4dcuisQdEVRM3lZWk9jazF5WU9SUWh5Y2ItU0E&usp=sharing
h..." https://www.pivotaltracker.com/story/show/66103376
Benjamin Pennell finished "Performance tests with transactions" https://www.pivotaltracker.com/story/show/66103376
<ajs6f>awoods: Did you see @benpennell's results? You were clearly right about the savings.11:45
<awoods>ajs6f: yes, it is quite promising11:46
<benpennell2>let me know if there are other tests you'd liked to see filled in there11:48
<tecoripa>awwods: just talked with gregjansen: yes, the puppet-fcrepo work can be pulled into the futures repo for its new home, if you like...11:50
awoods: there see if I get your attention now
<awoods>tecoripa: I will fork and move the repo today.11:51
<tecoripa>awoods: to continue: the puppet-fcrepo module is in pretty rough shape... it may cause more frustration than help in the state it's in right now (though Greg and I are using it to manage our clusters)
my plan is to refactor it and test it between this sprint and the next time I'm on board, in May11:52
awoods: once you fork it, shall I make that my new master?11:53
<awoods>on a call
<pivotal-bot____>Benjamin Pennell added "Distribute objects created by benchtool into subfolders" https://www.pivotaltracker.com/story/show/6687092811:55
* ksclarke joins12:02
* gregjansen joins12:03
* fasseg leaves12:09
<pivotal-bot____>Scott Prater started "Tomcat process persists after shutdown in cluster environment" https://www.pivotaltracker.com/story/show/6598865612:13
Scott Prater added comment: "Testing a fix using @gregoryjansen PR in ticket #66454096" https://www.pivotaltracker.com/story/show/6598865612:16
* scossu joins12:26
<awoods>tecorpia: re:new master, yes12:30
<pivotal-bot____>Scott Prater added comment: "Still seeing same issue with @gregoryjansen fix: https://gist.github.com/sprater/9351649" https://www.pivotaltracker.com/story/show/6598865612:41
<scossu>ajs6f: Re: PSARQL test failing for restrictive types - I changed the tests to use HTTP API to create resource, but the same error persists.
^^SPARQL
Code is in https://github.com/scossu/fcrepo4/tree/primary-type12:42
<tecoripa>awoods: sounds good.12:43
* tecoripa leaves12:46
<ajs6f>scossu: Okay, I'll try to get a look at that this afternoon.
* fasseg joins12:55
<scossu>ajs6f: Thanks.13:00
* edInCo joins13:04
* ajs6f leaves13:05
<awoods>tecoripa/gregjansen: Please re-master off of: https://github.com/futures/puppet-fcrepo13:52
<gregjansen>awoods: will do13:58
best explanation of TCPPING thus far: http://livedocs.adobe.com/livecycle/8.2/programLC/programmer/lcds/help.html?content=services_clustering_5.html14:00
every host needs it's own jgroups file if you use TCPPING, with different initial_hosts
A bit harder to puppet, that way. I am going to configure it manually and try some tests.14:01
<awoods>gregjansen: good luck... and keep us posted.14:03
ksclarke: Do you have a sense of reasonable sprints on which to participate? https://wiki.duraspace.org/display/FF/Beta+Sprint+Schedule14:06
<ksclarke>hi, so I'll be at code4lib at the end of the month so not #1414:08
was debating #13 or #15
awoods: is it best to make the initial ones sequential (for getting into the flow, etc.)?14:09
also not 100% I can wrap up what I'm doing currently by #13
<awoods>ksclarke: sequential is probably preferred. It sounds like maybe 15, 16, & 17?14:15
<ksclarke>yes, I think that sounds good
<awoods>ksclarke: great. would you mind updating the wiki page?14:16
<ksclarke>sure, will do
* dwilcox leaves14:18
<gregjansen>awoods: aaand we're up.. adding some notes to the wiki14:20
<awoods>gregjansen: what!? is that a hint of good news?
* dwilcox_ joins14:23
* scossu leaves14:27
<gregjansen>awoods: yes14:28
awoods: cannot decide whether to test, document or run around in circles
<awoods>gregjansen: just take it in for a moment.14:29
<gregjansen>awoods: interesting side note.. you cannot comment out an element in a jgroups config file..
* tecoripa joins14:30
<gregjansen><!-- <TCP
jgroups will still use the TCP element in the above
it seems
* ajs6f joins14:33
<awoods>gregjansen: that is odd... given that it is xml.14:34
<tecoripa>gregjansen: your changes in https://www.pivotaltracker.com/story/show/66454096 fixed my tomcat problem in https://www.pivotaltracker.com/story/show/65988656
<pivotal-bot____>feature: Ensure a clean shutdown of the repository on container shutdown. (finished) / owner: Gregory Jansen
<tecoripa>nice work!
<gregjansen>thx!
* ajs6f leaves14:36
<pivotal-bot____>Scott Prater added comment: "Never mind. @gregoryjansen fixes in his PR also fix what I was seeing. I'll mark this ticket as finished,..." https://www.pivotaltracker.com/story/show/65988656
* ajs6f1 joins
<pivotal-bot____>Scott Prater finished "Tomcat process persists after shutdown in cluster environment" https://www.pivotaltracker.com/story/show/6598865614:37
* scossu joins14:48
* gregjansen leaves15:01
* gregjansen joins15:07
<pivotal-bot____>Scott Prater added comment: "Tested the fixes on my cluster, pulling and building fcrepo-webapp from @gregoryjansen branch, and verified..." https://www.pivotaltracker.com/story/show/6645409615:10
* ajs6f1 leaves15:11
* ajs6f joins15:15
<pivotal-bot____>A. "Torcrexilly" Soroka added comment: "@escowles , @ericjames : We need a more precise definition of the way in which you want multiple..." https://www.pivotaltracker.com/story/show/6660307215:28
<tecoripa>I think I need to go to the JGroups thread. I cannot get my cluster work. Same response every time after two nodes join:15:50
JGRP000032: FedoraNode2-14366: no physical address for 79f6e783-ed99-325e-d0b9-a2a54f944acf, dropping message15:51
<ajs6f>Urg. So it wasn't New Relic/
?
<tecoripa>JGroups forum, not thread
no, it's not. I turned off my firewall, disabled the new relic jars, used gregjansen's config, tweaked it here and there...15:52
and always the same results
which makes me think it may be something really dumb and low-level, like hostnames or IP addresses or some such.
I've tried using IP addresses instead of hostnames, but that doesn't make a difference15:53
<ajs6f>Maybe you're wearing the wrong kind of hat. Maybe you should try sitting differently. Maybe it'll work tomorrow with no change and for no apparent reason.
* dwilcox_ leaves
<tecoripa>and it's oddd that the first two nodes always connect okay, but the following three error out with the above message.
<gregjansen>tecoripa: is that message found on fedoranode2 or some other node?15:54
* dwilcox joins
<gregjansen>tecoripa: if you are seeing that on fedoranode2, then it may be b/c it is trying to connect to itself and failing
<tecoripa>in this case it's on fedoranode215:55
so should loopback be set to true?
I saw that others seemed to have that issue
<gregjansen>I fixed it on my nodes by using the IP and not including the local IP in initial hosts param15:56
<tecoripa>I did try that. no dice.
<ajs6f>Is this all related to the use of the TCP clustering mode? I'm still wondering why everyone has failed to get UDP working...15:57
<tecoripa>this is all using TCP. I ditched UDP last week.
<ajs6f>tecorpia: I know— I was wondering if this is _because_ we'r esing CP.
using TCP.
<tecoripa>same problem, though, which leads me to believe the solution lies elsewhere.15:58
I saw the same errors with both TCP and UDP
* dwilcox leaves
<ajs6f>tecoripa: Oh, okay. Never mind then.
<tecoripa>buuuuut... I'm hopeful that if I can get this problem solved, then I can also get UDP to work, (and use NewRelic)15:59
* gregjansen leaves16:10
* gregjansen joins16:13
* dwilcox joins16:54
* dwilcox leaves16:59
* mikeAtUVa leaves17:14
* gregjansen leaves17:28
* tecoripa leaves17:59
* tecoripa joins18:15
* fasseg leaves18:22
* tecoripa leaves18:25
* ksclarke leaves18:28
* fasseg joins18:44
* fasseg leaves18:54
* edInCo leaves18:58
* fasseg joins19:00
* scossu leaves19:12
* scossu joins
* benpennell2 leaves19:17
* scossu leaves19:24
* benpennell2 joins19:36
<pivotal-bot____>Andrew Woods accepted "Extend kernel API to include more interfaces" https://www.pivotaltracker.com/story/show/6446840019:54
Andrew Woods accepted "Separate Java constants now found in fcrepo-jcr into a new artifact" https://www.pivotaltracker.com/story/show/6317248219:55
Andrew Woods added comment: "This should have been two separate PRs (#65979744) ...next time." https://www.pivotaltracker.com/story/show/6598005820:00
Andrew Woods accepted "AccessRoles session logout when post of new roles fails validation" https://www.pivotaltracker.com/story/show/65980058
Andrew Woods accepted "Install puppet and load balancer on UNC cluster." https://www.pivotaltracker.com/story/show/6609683220:02
Andrew Woods added comment: "When the new maven-checkstyle-plugin is released, we will be able to discard https://github.com/futures/mav..." https://www.pivotaltracker.com/story/show/6614493020:17
* aawoods__ joins20:23
* aawoods__ leaves
* awoods leaves20:24
* awoods joins20:26
<pivotal-bot____>Andrew Woods accepted "Modify benchtool to support testing transactions" https://www.pivotaltracker.com/story/show/6634383220:29
* benpennell2 leaves22:39