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

Using timezone: Eastern Standard Time
* ff_logger_ joins04:00
* dwilcox joins07:31
* ff_logger_ leaves07:35
* dhlamb joins08:02
* peichman joins09:00
* whikloj joins09:06
* yamil joins09:27
* mcritchlow joins10:12
* mcritchlow leaves10:17
* kestlund joins10:18
* mikeAtUVa joins10:32
* coblej joins10:37
* mikeAtUVa leaves10:48
<whikloj>https://wiki.duraspace.org/display/FF/2017-03-09+-+Fedora+Tech+Meeting10:58
* barmintor joins11:04
* kefo joins11:05
<ruebot>dhlamb++
<kefo>That was me joining, late. Apologies.11:06
<whikloj>https://github.com/fcrepo/fcrepo-specification/issues/40
<ruebot>if barmintor is here, are we good to close https://github.com/fcrepo/fcrepo-specification/issues/41 ?
<barmintor>ruebot: IDK11:07
ruebot: I'm serious when I say that
* ruebot snorts
<barmintor>ruebot: IDK if the spec is usably complete if we don't have an idea of how to do it
<whikloj>https://github.com/fcrepo/fcrepo-specification/issues/4111:08
<dhlamb>getting a bit max headroom, but i think i'm following you11:10
<ruebot>barmintor _is_ max headroom
<barmintor>I thik it also gets into the weirdness of the external-body MIME11:11
whch we have a very editorial interpretation of
we have to decide whether it's in scope, but we need the editors to explicitly ackowledge what deciding it's out of scope means11:13
<ruebot>barmintor++11:16
<dhlamb>+1
<barmintor>done!
I will harrass hydra about it tout suite.11:17
<dhlamb>i have an interest, but zero time. i know that's not helpful.
<barmintor>also I'm sorry if I sounded mean. I don't like sounding mean.11:18
1976!
<whikloj>https://jira.duraspace.org/browse/FCREPO-1976
<dwilcox>You never sound mean, barmintor
<barmintor>whoa
I think my aassignment of 1976 was "ben will raise this issue on the ldp list"11:19
<whikloj>https://wiki.duraspace.org/display/FF/Many+Members+Performance+Testing11:20
https://github.com/dbernstein/fcrepo4/compare/master...dbernstein:daa11f311:23
<barmintor>you're very faint on the line, but the gist that I understand was that you were able to acheive a significant performance gain on hasMember by being less invested in readably ordered triples?11:29
oh, I see. That would line up with some of what PUL observed, too11:30
so there's a hypothesis that when you use a RDBMS backend, MODE doesn't use a cache, and also that the default size of the cache has a performance inflection around 10000 nodes11:31
<kefo>barmintor: That's what I understand11:32
<barmintor>kefo++ // thanks, my audio set up is perpetually busted
<kefo>Here's dbernstein's commit: https://github.com/dbernstein/fcrepo4/commit/f453a1da868c899ac6c71b901b87a8231ccb8c0711:34
I can say that I'll be up-ing the modeshape cache size in our test system, which several hundred thousand resources, and give this a whorl.11:36
<whikloj>kefo++
<dhlamb>sounds promising11:37
<ruebot>whikloj++11:38
<barmintor>IT HAS BEEN MET
<dhlamb>good form whikloj
* coblej leaves11:49
* mikeAtUVa joins11:57
* mikeAtUVa leaves12:04
* coblej joins12:43
* coblej leaves13:00
* coblej joins13:01
* coblej leaves
* coblej joins
* dwilcox leaves13:04
* kestlund leaves13:11
* dwilcox joins13:22
* dwilcox leaves15:18
* dwilcox joins15:31
* coblej leaves15:51
* coblej joins15:53
* coblej leaves15:57
* coblej joins16:19
* coblej leaves16:21
* coblej joins16:22
* acoburn joins16:52
* mcritchlow joins
* yamil leaves16:58
* acoburn leaves17:03
* coblej leaves17:05
* peichman leaves17:20
* mcritchlow leaves17:33
* whikloj leaves18:00
* peichman joins18:20
* peichman leaves18:24
* mcritchlow joins18:37
* mcritchlow leaves19:17
* peichman joins19:31
* mcritchlow joins19:32
* peichman leaves19:36
* peichman joins19:54
* peichman leaves20:06
* kefo leaves20:11
* mcritchlow leaves20:34
* mcritchlow joins21:59
* mcritchlow leaves
* mcritchlow joins
* dwilcox leaves23:01
* mcritchlow leaves23:20
* mcritchlow joins23:21
* mcritchlow leaves23:25
* mcritchlow joins23:49

Generated by Sualtam