Perl 6 - the future is here, just unevenly distributed

IRC log for #metacpan, 2016-09-08

| Channels | #metacpan index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
01:53 metacpan joined #metacpan
01:53 metacpan [metacpan-api] oalders pushed 1 new commit to master: https://git.io/viWuq
01:53 metacpan metacpan-api/master 848b46e Olaf Alders: Merge pull request #517 from metacpan/mickey/GH515...
01:53 metacpan left #metacpan
01:53 metacpan joined #metacpan
01:53 metacpan [metacpan-api] oalders deleted mickey/GH515 at 0aad81b: https://git.io/viWum
01:53 metacpan left #metacpan
02:23 oalders ranguard: https://metacpan.org/pod/release/OALDERS/MetaCPAN-Moose-0.000003/lib/MetaCPAN/Moose.pm
03:00 jibsheet joined #metacpan
03:00 rblackwe joined #metacpan
03:00 jdv79 joined #metacpan
03:00 ranguard joined #metacpan
03:00 haarg joined #metacpan
03:00 moltar joined #metacpan
03:00 Kundun joined #metacpan
03:00 arc joined #metacpan
03:00 simcop2387 joined #metacpan
03:00 rwstauner joined #metacpan
03:00 dg joined #metacpan
03:00 trs joined #metacpan
03:00 cjm joined #metacpan
03:00 osfameron joined #metacpan
03:00 harleypig_ joined #metacpan
03:00 dolmen_ joined #metacpan
03:00 zostay joined #metacpan
03:00 Exodist joined #metacpan
03:00 chansen_ joined #metacpan
03:01 punytan_ joined #metacpan
03:01 ingy joined #metacpan
03:01 oalders joined #metacpan
03:01 ribasushi joined #metacpan
03:01 karjala joined #metacpan
03:01 Tempesta joined #metacpan
03:01 Grinnz joined #metacpan
03:01 TBSliver joined #metacpan
03:02 bjakubski joined #metacpan
03:49 y1mmm joined #metacpan
05:45 oiami joined #metacpan
06:48 damog joined #metacpan
07:31 neilb joined #metacpan
08:10 neilb joined #metacpan
08:12 Relequestual joined #metacpan
08:19 ranguard oalders++
08:34 edward joined #metacpan
09:08 GumbyNET7 joined #metacpan
09:52 GumbyNET5 joined #metacpan
10:03 mickey ranguard: good thinking... this may explain the problem
10:17 uree joined #metacpan
10:20 ranguard mickey: maybe when stuff's added to the queue add a 1 min delay and also a try again after 1 fail in 10 mins?
10:21 ranguard also if that 2nd one fails we should record / alert / do something so we can report/restart it easily - without waiting for users to tell us of the issue
10:34 mickey the last one is what i tried to come up with but didn't finish yet... delay always feels wrong
10:35 mickey general question: why do we rsync to the other machines? any reason not to have a shared fs between them?
11:09 Relequestual joined #metacpan
11:14 metacpan joined #metacpan
11:14 metacpan [metacpan-api] mickeyn created mickey/GH517 (+1 new commit): https://git.io/vilqR
11:14 metacpan metacpan-api/mickey/GH517 8f2303c Mickey Nasriachi: added tests for download_url controller
11:14 metacpan left #metacpan
11:16 metacpan joined #metacpan
11:16 metacpan [metacpan-api] mickeyn deleted mickey/GH517 at 8f2303c: https://git.io/vilqa
11:16 metacpan left #metacpan
11:16 metacpan joined #metacpan
11:16 metacpan [metacpan-api] mickeyn created mickey/GH519 (+1 new commit): https://git.io/vilqR
11:16 metacpan metacpan-api/mickey/GH519 8f2303c Mickey Nasriachi: added tests for download_url controller
11:16 metacpan left #metacpan
11:52 metacpan joined #metacpan
11:52 metacpan [metacpan-api] ranguard deleted mickey/GH519 at 8f2303c: https://git.io/vilqa
11:52 metacpan left #metacpan
13:49 Lee joined #metacpan
14:18 oalders ranguard, mickey: i wouldn't want the 1 minute delay to start with
14:19 oalders i'd say index what we have but maybe add a second job to re-index after x minutes
14:19 oalders people get uptight if they don't see their upload in 30 seconds
14:23 mickey oalders: i didn't look into the details yet, but i'm assuming this is only watcher related indexing - so maybe we can just not use the queue from the watcher ?
14:24 oalders mickey: we could, but it would be nice to be consistent and use the queue for everything
14:25 oalders and we'll have to do something like this anyway, since 02packages doesn't update at exactly the same time as we get the new tarball
14:25 oalders if we use the queue, we can scale it to all the machines in the cluster
14:26 oalders and once we have good statistics for the queue, we can get a better overview of what's going on if everything runs through it
14:27 mickey sure, but then we have to have a synchronized data source for all the machines running workers
14:27 oalders we do. i think we want that anyway, don't we?
14:28 oalders but in the meantime, we can just run workers on the one production box
14:28 mickey if we depend on rsync between the servers and can expect a "1 min delay" that's not so synched
14:29 oalders where is the rsync delay?
14:29 oalders ie specifically which delay are we talking about
14:29 oalders we can talk multiple machines out of the conversation for now, that's complicating the problem
14:30 mickey 23:36:07 � ranguard� mickey: just thought - indexing issue could be because now via the queue, the rsync might be a min or so diff in the actual worker box
14:30 oalders ie if we only index on the machine that has the CPAN what's the main problem?
14:30 oalders right, so either let's only index on the one box or we can fail and retry if the tarball is not yet on the worker machine
14:31 mickey i would assume we wouldn't want to lose worker power by limiting the queue to one machine
14:32 oalders for most things it shouldn't matter too much, but it's nice to have.  for the full re-index it obviously makes a huge difference
14:32 mickey but that will indeed solve this problem (if it's actually the sourse of it)
14:32 oalders is there a specific ticket we're talking about?
14:32 mickey full re-index is controlled and therefore we can manually add the other servers with workers when needed
14:33 mickey not sure, i just woke up this morning to see ranguard's message :)
14:34 mickey so limiting the queue to one host is something we can try as a quick fix
14:34 mickey ranguard: ^^ ?
14:42 ranguard all boxes use rrr-watcher to keep insync with the master CPAN mirror
14:43 ranguard I don't know what - if any delay - there actually is, but I think we have to deal with it being a posibility
14:47 ranguard FYI: setting up a new box, I do a full rsync from of our existing boxes (as we are a backpan) then start rrr-watcher
14:48 ranguard I have been told there is an alternative to rrr-watches, but haven't looked at it
14:48 ranguard we also do a nightly full rsync from the master CPAN mirror for any glitches
14:48 ranguard this happens on all 6 of our production boxes
14:49 ranguard Could just set the worker to sleep for X and try find the file again before actually failing (IF that's what's happening) :)
15:00 ranguard http://www.staff.science.uu.nl/~penni101/iim/ is the alternative
15:01 ranguard I'll have a look at that in anycase
15:07 oalders neat.  did not know about iim
15:08 oalders i don't think we want to tie up the worker with a sleep.  the job can just go back into the queue
17:37 neilb joined #metacpan
17:38 ranguard oalders: sure, so maybe that's all we need to do, requeue if not found on disk and then try again
17:39 oalders should be easy to write a test for that
17:42 reyjrar joined #metacpan
18:12 ranguard https://v1.metacpan.org/release/LLAP/MetaCPAN-Role-0.01 - duff changes file but otherwise it's ok :)
18:13 ranguard Added OALDERS to co-maintainers of CatalystX::Fastly::Role::Response.
18:13 ranguard Added OALDERS to co-maintainers of MetaCPAN::Role.
18:13 ranguard Added OALDERS to co-maintainers of MetaCPAN::Role::Fastly.
18:13 ranguard Added OALDERS to co-maintainers of MetaCPAN::Role::Fastly::Catalyst.
18:13 ranguard Added OALDERS to co-maintainers of MooseX::Fastly::Role.
18:13 ranguard :)
18:14 ranguard mickey: let me know if you are interested in any of those :)
18:26 mickey sure, give me co-maint so i can break stuff :)
18:37 ranguard mickey: done
18:39 oalders so much co-maint!
18:39 oalders ranguard++ # bus number
18:41 mickey ranguard++
18:45 metacpan joined #metacpan
18:45 metacpan [metacpan-api] ranguard created leo/fastly_purge_and_headers (+6 new commits): https://git.io/vi8nQ
18:45 metacpan metacpan-api/leo/fastly_purge_and_headers ae3be4e Leo Lapworth: update to latest Net::Fastly
18:45 metacpan metacpan-api/leo/fastly_purge_and_headers 4fc761c Leo Lapworth: switch to calling all fastly services, will be www and fastapi
18:45 metacpan metacpan-api/leo/fastly_purge_and_headers 0184602 Leo Lapworth: make Fastly role compatible with MC::Web
18:45 metacpan left #metacpan
18:48 ranguard oalders: 42 :)
18:49 oalders :)
18:54 metacpan joined #metacpan
18:54 metacpan [metacpan-web] ranguard force-pushed leo/fastly_add_keys_for_caching from c1c37d2 to cf6c389: https://git.io/v6hKr
18:54 metacpan metacpan-web/leo/fastly_add_keys_for_caching 3fd66fb Leo Lapworth: remove merge or copy/paste dupe line
18:54 metacpan metacpan-web/leo/fastly_add_keys_for_caching 4b97d49 Leo Lapworth: add a browser_max_age()
18:54 metacpan metacpan-web/leo/fastly_add_keys_for_caching 7b9c38f Leo Lapworth: add cache headers to /feed
18:54 metacpan left #metacpan
19:03 reyjrar joined #metacpan
19:37 ranguard https://github.com/metacpan/metacpan-web/pull/1765 ready for review if anyone wants to look - can't merge until I find this API issue though
20:02 oalders ranguard: good stuff.  i'll look it over this evening
20:08 metacpan joined #metacpan
20:08 metacpan [metacpan-api] ranguard pushed 1 new commit to leo/fastly_purge_and_headers: https://git.io/vi8V4
20:08 metacpan metacpan-api/leo/fastly_purge_and_headers 09c9edd Leo Lapworth: fix tests
20:08 metacpan left #metacpan
20:11 ranguard oalders: https://github.com/metacpan/metacpan-api/pull/521 also ready for review (think tests will pass now and everything) :)
20:11 oalders ranguard: great! i'll add it to my list
20:16 ranguard reyjrar: do you have that script for rsyslog->ES for Kibana :) ?
21:00 neilb_ joined #metacpan
21:03 reyjrar ranguard: I got sidetracked on graphite/statsd stuff and didn't finish it.. I plan on hacking on it tonight while the wife is in class.. Thu is my hack night :)
21:06 ranguard reyjrar: sweet :)
21:07 ranguard I'd also like to ask you to look at ES snapshot / backup stuff at some point if thats ok
21:08 ranguard but can discuss another day - nn :)
21:11 reyjrar k, haven't played much with the es snapshot stuff.. it was terribly unstable in the 1.x release.. everything is so much better in the 2.x release
21:11 reyjrar It's like amazingly stable and dies like orders of magnitude less.
21:32 oiami joined #metacpan
22:04 Timbus joined #metacpan
22:08 Timbus Hey dudes & dudettes. I've setup a metacpan-api for internal modules at $work, but after getting it all running it doesn't appear to be working specifically with cpanm's use of the 'source' query parameter? Normal queries seem to be working.. Is there a setting I missed?
22:08 Timbus Specifically, I get `Unknown param (source) in (search) request.`
22:09 oalders Timbus: this is using the latest master branches?
22:11 oalders I don't know if it has been properly tested with cpanm.  not sure if mickey can say more about that
22:12 oalders I'm going afk, but if you can possibly put the full cpanm query and the full error message into a github issue for metacpan-api, that would be helpful
22:12 Timbus This is using the metacpan-developer repo with the vagrant setup..
22:13 oalders Timbus: ok, that would be the latest master, which is v1 of the API.  that's not strictly live yet, so there are still some rough patches
22:13 Timbus the query is fairly, it's just that the ?source='meh' doesn't seem to work
22:13 Timbus ooh
22:13 Timbus fairly simple*
22:14 oalders Timbus: you could test it using https://fastapi.metacpan.org/v1
22:14 oalders To confirm the bug
22:14 Timbus awesome, I'll try that
22:14 oalders thanks!
22:15 oiami left #metacpan
22:15 Timbus yes, same issue
22:16 oalders Timbus: thanks. if you can create an issue, we can track it there
22:17 Timbus on gitlab?
22:17 Timbus github*
22:18 oalders Timbus: please
22:19 Timbus On it. And I'll see about checking out an older version so I have something going. Thanks for your help!
23:03 ThePostar614 joined #metacpan
23:03 ThePostar614 Best 3D XXX Games And More http://www.mysexgames.org/
23:03 ThePostar614 left #metacpan
23:34 HappyGilmore760 joined #metacpan
23:34 HappyGilmore760 left #metacpan
23:39 HappyGilmore120 joined #metacpan
23:39 HappyGilmore120 left #metacpan

| Channels | #metacpan index | Today | | Search | Google Search | Plain-Text | summary