Log of the #fcrepo channel on chat.freenode.net

Using timezone: Eastern Standard Time
* dhlamb joins08:48
* dwilcox joins09:00
* bseeger joins09:02
* acoburn joins09:15
* peichman joins09:44
* mohamedar joins09:52
* mohamedar1 joins09:58
* mohamedar1 leaves
* mohamedar1 joins
* mohamedar leaves10:00
* jrgriffiniii joins10:08
* bseeger leaves10:12
* jrgriffiniii_ joins10:20
* jrgriffiniii leaves10:22
* dwilcox leaves10:25
* jrgriffiniii_ joins
* jrgriffiniii leaves10:26
* github-ff joins10:29
[fcrepo4] yinlinchen closed pull request #953: fcr:backup error response (master...fcrepo-1788) http://git.io/v4ItP
* github-ff leaves
* jrgriffiniii_ leaves10:33
* whikloj joins10:39
* bseeger joins10:41
* jrgriffiniii joins
* jrgriffiniii leaves10:46
* cmills joins
* jrgriffiniii joins10:47
* cmmills leaves10:49
* chadmills leaves
* cmmills joins
* jrgriffiniii leaves10:52
* jrgriffiniii joins
* dwilcox joins10:57
<ruebot>is there a performance and scaling call today?10:58
<jrgriffiniii>I assumed so11:04
ruebot: (I'm currently on the call, at least one other is in attendance)
* mohamedar1 leaves11:06
<ruebot>jrgriffiniii: ok. i'll jump on then.11:07
* aelkiss joins11:10
* dwilcox leaves
<ruebot>https://github.com/fcrepo4-labs/fcrepo4-jmeter
https://wiki.duraspace.org/display/FF/Performance+and+Scalability+Test+Plans11:12
ajwagner: are you joining us today?
* jrgriffiniii leaves11:13
* dwilcox joins11:15
* jrgriffiniii joins11:19
<ruebot>jrgriffiniii, aelkiss: how's this for notes? https://wiki.duraspace.org/display/FF/2016-01-11+Performance+-+Scale+Meeting11:22
* jrgriffiniii_ joins11:23
<aelkiss>Looks good to me
* jrgriffiniii leaves11:24
* aelkiss leaves11:25
* aelkiss joins11:27
* jrgriffiniii_ leaves
* bseeger leaves11:29
* jrgriffiniii joins11:35
ruebot: Sorry for the disconnection, and thank you again for drafting and distributing the notes11:36
<ruebot>jrgriffiniii: no worries!11:50
* jrgriffiniii leaves11:53
* jrgriffiniii joins11:54
* jrgriffiniii leaves12:01
* esm_ joins
* esm_ leaves12:10
* jrgriffiniii joins12:42
* esm_ joins13:06
* jrgriffiniii leaves13:19
* jrgriffiniii joins13:24
* jrgriffiniii leaves13:29
* jrgriffiniii joins13:38
* jrgriffiniii leaves13:45
* bseeger joins13:47
* jrgriffiniii joins13:51
* jrgriffiniii_ joins13:56
* jrgriffiniii leaves
* jrgriffiniii_ leaves14:00
* jgpawletko joins
* bseeger leaves14:08
* esm_ leaves14:10
* ajs6f joins14:13
* jrgriffiniii joins14:19
* jrgriffiniii leaves14:43
* bseeger joins14:46
<whikloj>acoburn: ping14:52
<acoburn>whikloj: pong14:53
<whikloj>acoburn: can you point me in the direction of a good way to check memory usage in a Camel route?
if such a thing exists
<acoburn>whikloj: have you ever used hawt.io?
<whikloj>acoburn: yes using that now, but not getting memory info, or maybe not looking in the right place..14:54
<acoburn>whikloj: I'll check, just a sec
whikloj: you mean a _particular_ camel route, right?15:00
whikloj: as opposed to the entire camel runtime?
<whikloj>acoburn: yes
acoburn: though the camel runtime on my particular machine might be okay too
<ajs6f>whikloj:acoburn: Sounds like a job for JMX.
<whikloj>ajs6f: yeah I looked but the JMX "properties" don't seem to cover memory15:01
ajs6f: mostly thread size, route received, processed, failed, etc
<ajs6f>whikloj: No, not JMX on the individual oibjects, JMX on the heap.
whikloj: Open jconsole on your JVM.15:02
<whikloj>ajs6f: oh, yes I have to wait and restart Tomcat with the jmx remote. But that is my next step. I was trying to see if I could do it right now.15:03
^^ jmx remote enabled
<ajs6f>whikloj: Are you not going to run jconsole on the same machine?
<acoburn>whikloj: the "dashboard" portion of hawt.io shows memory info about the heap
<whikloj>ajs6f: Will it run in a terminal environment15:06
acoburn: thanks, I guess that is probably my best option
<ajs6f>whikloj: Is the JVM hosting Camel on the same machine as the one on which you are going to run jconsole (or other JMX probe)?
<whikloj>ajs6f: normally no, I run jconsole on my workstation.15:07
<ajs6f>whikloj: Oh, okay. If they are on the same machine you can avoid tinkering with the remoting stuff.
whikloj: Might be worth using X/Windows. That's what I do.15:08
<whikloj>ajs6f: yeah, feels like I am building a mountain of work just to confirm that my camel route is more memory efficient than ol' Gsearch15:09
or maybe https://github.com/cjmx/cjmx15:11
* jrgriffiniii joins
<ajs6f>whikloj: A) Efficient != always is using less memory. Efficiency is a ratio. What is your numerator and denominator? Instant memory usage per instant throughput? Average memory usage per average throughput? Memory per indexing action? These are all different. B) Suppose Camel is _not_ more efficient. What does that mean?
<whikloj>ajs6f: more on a per action basis. If I index 1000 documents with GSearch and then with my camel-route, which a) is faster and b) uses less memory to create the same result.15:12
* aelkiss leaves15:14
<ajs6f>whikloj: What is "uses less memory"? Instantaneous maximum total heap? Average total heap? Instantaneous maximum heap per document? Average heap per document?
<whikloj>ajs6f: I'd probably say average heap per document or possibly average total heap, but my understanding of the whole java heap concept is limited15:15
<ajs6f>whikloj: Heap is almost all memory allocated to state.15:16
whikloj: Total heap is going to include overhead from various sources, so the comparison will be harder, but on the other hand, heap per document is going to be harder to measure, because it's a more transient observation.15:17
whikloj: Choose you poison.
<whikloj>ajs6f: arsenic....oh you meant...I got it.15:18
* jrgriffiniii leaves
<ajs6f>whikloj: No, I was thinking of http://www.last.fm/music/Poison:: "There are at least 6 bands with the name Poison:"15:21
<whikloj>ajs6f: Ahhhh yeah, as a child of the 80s I appreciate Bret Michaels' sweet lyrics.15:22
* jrgriffiniii joins15:23
<ajs6f>whikloj: I think we should start including album covers with releases. This is inspiring: http://www.amazon.com/poison-album-list-order-release/lm/R3BSZEH0CI3Q7815:24
* jrgriffiniii leaves15:28
<whikloj>ajs6f++15:30
<ajs6f>This is actually a really good idea. It would invite participation from the community. It would enable contribution from non-programmers.15:31
* jrgriffiniii joins15:33
<whikloj>We could have a battle of the bands. Poison vs Cinderella, Dokken vs Warrant15:44
* jrgriffiniii_ joins15:46
* jrgriffiniii leaves15:47
* dwilcox leaves15:49
* jrgriffiniii leaves15:52
* dhlamb leaves15:53
* jrgriffiniii joins16:01
* jrgriffiniii leaves16:09
* jrgriffiniii joins16:12
* jrgriffiniii leaves16:16
* acoburn leaves16:21
* jrgriffiniii joins
* jrgriffiniii_ joins16:27
* jrgriffiniii leaves16:28
* jrgriffiniii leaves16:32
* ajs6f leaves16:44
* peichman leaves17:21
* bseeger leaves17:38
* dwilcox joins17:55
* dwilcox leaves17:58
* whikloj leaves17:59
* ksclarke leaves18:06
* dwilcox joins18:11
* dwilcox leaves
* dwilcox joins18:13
* dwilcox leaves18:29
* dhlamb joins19:12
* mohamedar joins
* mohamedar leaves19:17
* dwilcox joins19:32
* dhlamb leaves19:54
* dhlamb joins
* dwilcox leaves19:58
* dwilcox joins20:01
* dwilcox leaves20:07
* ksclarke joins20:17
* dwilcox joins20:18
* ksclarke leaves20:21
* ksclarke joins20:31
* ksclarke leaves20:35
* dwilcox leaves20:47
* ksclarke joins20:53
* dhlamb leaves21:04
* dhlamb joins22:04
* jgpawletko leaves22:09
* dhlamb leaves23:35