Displaying #traffic-server/2015-12-14.log:

Mon Dec 14 00:48:21 2015  biilmann:Joined the channel
Mon Dec 14 00:56:50 2015  pquerna:zwoop: http://curl.haxx.se/mail/lib-2015-12/0053.html
Mon Dec 14 01:05:02 2015  biilmann:Joined the channel
Mon Dec 14 01:11:07 2015  biilmann:Joined the channel
Mon Dec 14 01:16:07 2015  biilmann:Joined the channel
Mon Dec 14 02:00:31 2015  PSUdaemon:pquerna: we don't support H2 in 5.3.2 officially
Mon Dec 14 02:00:44 2015  PSUdaemon:that should work against a 6.0 box with H2 enabled
Mon Dec 14 02:06:07 2015  psp:Joined the channel
Mon Dec 14 03:25:33 2015  ibrezac:Joined the channel
Mon Dec 14 03:58:34 2015  dxu:2/cr
Mon Dec 14 03:58:44 2015  dxu:whoops
Mon Dec 14 04:10:18 2015  oag:Joined the channel
Mon Dec 14 04:13:19 2015  oag:I have a question about plugins and threads.
Mon Dec 14 04:13:57 2015  oag:Up until just now I have been finding that my plugin (an atscppapi InterceptPlugin) gets all its calls in the same thread.
Mon Dec 14 04:14:25 2015  oag:I've been using AsyncHttpFetch and that seems always to return on the thread it was invoked on.
Mon Dec 14 04:15:29 2015  oag:But it seems when I do https termination, the plugin is initially invoked on a different thread, but gets any fetch results back on the thread that handles all http traffic.
Mon Dec 14 04:16:30 2015  oag:I know, one option would be to make the plugin fully thread-safe but that might not be easy.
Mon Dec 14 04:18:56 2015  oag:I'd prefer to somehow arrange for each plugin instance to always operate on the thread that instantiated it.
Mon Dec 14 04:20:29 2015  oag:I would naively guess that each thread is associated with some select/poll loop somewhere.
Mon Dec 14 04:21:45 2015  oag:Is there any mechanism for taking a call or event from one thread and turning it into an event on a different specified thread?
Mon Dec 14 04:25:06 2015  oag:I suppose another option would be to set up a dedicated thread pool for the work my plugin does, and have the TS plugin just be an adapter that passes messages to and from my workers.
Mon Dec 14 04:37:24 2015  maskit_:Joined the channel
Mon Dec 14 04:37:46 2015  kichan:Joined the channel
Mon Dec 14 04:38:47 2015  bcall:Joined the channel
Mon Dec 14 04:41:19 2015  briang:Joined the channel
Mon Dec 14 04:41:20 2015  briang:Joined the channel
Mon Dec 14 04:41:49 2015  sudheerv:Joined the channel
Mon Dec 14 04:42:19 2015  jpeach:Joined the channel
Mon Dec 14 06:03:45 2015  dstates:Joined the channel
Mon Dec 14 06:29:41 2015  _mjones:Joined the channel
Mon Dec 14 06:32:13 2015  felicity:Joined the channel
Mon Dec 14 07:33:08 2015  mturk:Joined the channel
Mon Dec 14 07:33:09 2015  mturk:Joined the channel
Mon Dec 14 07:34:53 2015  briang:HostDB.cc has the following code:
Mon Dec 14 07:34:59 2015  briang: if (!ttl_seconds)
Mon Dec 14 07:34:59 2015  briang: ttl_seconds = 1; // www.barnsandnobel.com is lame
Mon Dec 14 07:35:00 2015  briang:lol
Mon Dec 14 07:35:20 2015  briang:given that no one uses banesandnoble.com anymore is it safe to remove this? lol
Mon Dec 14 07:56:43 2015  _mjones:I'm not a ts dev but my inclination would be to 'git blame' that and look for an issue number.
Mon Dec 14 08:36:16 2015  oag:Joined the channel
Mon Dec 14 08:58:09 2015  biilmann:Joined the channel
Mon Dec 14 09:26:56 2015  Lethalman:Joined the channel
Mon Dec 14 09:59:14 2015  biilmann:Joined the channel
Mon Dec 14 10:59:57 2015  biilmann:Joined the channel
Mon Dec 14 11:46:05 2015  JSeymour:Joined the channel
Mon Dec 14 12:00:41 2015  biilmann:Joined the channel
Mon Dec 14 12:28:24 2015  mturk:Joined the channel
Mon Dec 14 12:28:24 2015  mturk:Joined the channel
Mon Dec 14 13:01:29 2015  biilmann:Joined the channel
Mon Dec 14 13:59:01 2015  ibrezac:Joined the channel
Mon Dec 14 14:02:12 2015  biilmann:Joined the channel
Mon Dec 14 14:10:45 2015  shinrich1:Joined the channel
Mon Dec 14 14:35:02 2015  esproul:Joined the channel
Mon Dec 14 14:57:18 2015  zwoop:briang seems fine to remove, if BarnesAndNoble don't approve, they can file a Jira
Mon Dec 14 14:58:17 2015  felicity:some Cisco routers returning 0 TTL for CNAMEs, in my experience. not sure if that's a bug or a feature
Mon Dec 14 14:59:29 2015  zwoop:probably a feature, try to prevent any caching I'd imagine ?
Mon Dec 14 14:59:32 2015  jrushford:Joined the channel
Mon Dec 14 15:03:00 2015  biilmann:Joined the channel
Mon Dec 14 15:05:45 2015  davet_:Joined the channel
Mon Dec 14 15:22:26 2015  yamc:Joined the channel
Mon Dec 14 15:33:57 2015  zwoop:maskit did you get your ASF account setup ?
Mon Dec 14 15:36:54 2015  maskit_:no, it seems the account request is still in the queue
Mon Dec 14 15:36:57 2015  maskit_:https://whimsy.apache.org/new-account-request-queue
Mon Dec 14 15:37:40 2015  zwoop:ok, cool
Mon Dec 14 15:42:25 2015  zwoop:alright, heading out to CA, don't do anything crazy while I'm gone. yamc !
Mon Dec 14 15:50:24 2015  yamc:Too late.
Mon Dec 14 16:00:29 2015  jumby:Joined the channel
Mon Dec 14 16:03:43 2015  biilmann:Joined the channel
Mon Dec 14 16:19:55 2015  biilmann:Joined the channel
Mon Dec 14 16:55:25 2015  meeramn:Joined the channel
Mon Dec 14 17:01:19 2015  ibrezac:Joined the channel
Mon Dec 14 17:06:59 2015  shinrich1:Joined the channel
Mon Dec 14 17:07:16 2015  es:Joined the channel
Mon Dec 14 17:09:05 2015  es1:Joined the channel
Mon Dec 14 17:11:36 2015  dxu_:Joined the channel
Mon Dec 14 17:11:54 2015  blattj:Joined the channel
Mon Dec 14 17:34:37 2015  meeramn:Joined the channel
Mon Dec 14 17:39:28 2015  _mjones:Routers returning 0 TTL? Maybe firewalls (ASA) or load-balancers even (Ace), but I rather doubt routers.
Mon Dec 14 17:39:47 2015  gancho:Joined the channel
Mon Dec 14 17:40:11 2015  felicity:_mjones: i have a router here that does it
Mon Dec 14 17:41:02 2015  felicity:although it only seems to happen sometimes, i haven't worked out the reason
Mon Dec 14 17:42:33 2015  _mjones:felicity: non-deterministic altering of DNS responses? Cisco firewalls commit some great crimes in their default config, but I can't say I've seen the same from routers.
Mon Dec 14 17:52:44 2015  blattj:Joined the channel
Mon Dec 14 17:53:42 2015  shinrich2:Joined the channel
Mon Dec 14 18:34:29 2015  meeramn:Joined the channel
Mon Dec 14 18:36:47 2015  es:Joined the channel
Mon Dec 14 18:53:08 2015  biilmann:Joined the channel
Mon Dec 14 19:00:33 2015  blattj:Joined the channel
Mon Dec 14 19:00:42 2015  PSUdaemon:jpeach: ping
Mon Dec 14 19:00:48 2015  jpeach:y
Mon Dec 14 19:01:01 2015  PSUdaemon:are we merging our own PR's
Mon Dec 14 19:01:06 2015  PSUdaemon:or does someone else have to do it/
Mon Dec 14 19:01:15 2015  PSUdaemon:once they have review
Mon Dec 14 19:03:25 2015  zwoop:merge your own
Mon Dec 14 19:04:06 2015  PSUdaemon:k
Mon Dec 14 19:04:09 2015  PSUdaemon:will do
Mon Dec 14 19:04:11 2015  zwoop:and, you don't have to wait indefinitely, if there's no review, it's an implicit "+1". What we sort of failed to agree on is how long to let a PR sit for (possible) review. I'd suggest use your judgement, but 2-3 days seems reasonable)
Mon Dec 14 19:04:46 2015  zwoop:we're still in an official CTR policy, with the addition of tracking / requesting reviews via PRs
Mon Dec 14 19:06:05 2015  jpeach:2 - 3 days is short if the PR is non-trivial
Mon Dec 14 19:10:51 2015  zwoop:agreed
Mon Dec 14 19:10:55 2015  zwoop:use your judgement :)
Mon Dec 14 19:11:23 2015  zwoop:also, if you (someone) are going to review it, but not right now, a comment on the PR is prudent ("I will review this on Tuesday").
Mon Dec 14 19:20:57 2015  ibrezac:Joined the channel
Mon Dec 14 19:25:18 2015  yamc:Joined the channel
Mon Dec 14 19:28:39 2015  swoc:Joined the channel
Mon Dec 14 19:33:11 2015  gancho:Joined the channel
Mon Dec 14 19:35:48 2015  gancho:Joined the channel
Mon Dec 14 19:38:27 2015  blattj:Joined the channel
Mon Dec 14 19:39:49 2015  gancho:Joined the channel
Mon Dec 14 19:46:13 2015  zwoop:woot
Mon Dec 14 19:46:26 2015  zwoop:PSUdaemon did you add docs for the new option too ?
Mon Dec 14 19:46:40 2015  jpeach:PSUdaemon: can you have rushford rebase?
Mon Dec 14 19:47:20 2015  PSUdaemon:zwoop: there was no mention of docs in the PR!!
Mon Dec 14 19:47:22 2015  PSUdaemon:i can add docs...
Mon Dec 14 19:47:27 2015  PSUdaemon:jpeach: just about to
Mon Dec 14 19:47:34 2015  jpeach:thanks :)
Mon Dec 14 19:47:48 2015  zwoop:PSUdaemon I think you should add a line or two on the new option ;)
Mon Dec 14 19:48:41 2015  PSUdaemon:zwoop: will do
Mon Dec 14 19:49:16 2015  zwoop:I should have said something on the PR, so blame me
Mon Dec 14 19:50:20 2015  PSUdaemon:you don't have to say anything. i'll blame you anyway.
Mon Dec 14 19:51:25 2015  zwoop:+1
Mon Dec 14 19:51:27 2015  zwoop:everyone does
Mon Dec 14 19:51:38 2015  gancho:Joined the channel
Mon Dec 14 20:02:33 2015  blattj:Joined the channel
Mon Dec 14 20:08:29 2015  gancho:Joined the channel
Mon Dec 14 20:16:38 2015  shinrich1:Joined the channel
Mon Dec 14 20:34:08 2015  blattj1:Joined the channel
Mon Dec 14 20:35:53 2015  blattj:Joined the channel
Mon Dec 14 20:37:32 2015  gancho:Joined the channel
Mon Dec 14 20:39:46 2015  meeramn:Joined the channel
Mon Dec 14 21:12:22 2015  reveller:Joined the channel
Mon Dec 14 21:19:58 2015  biilmann:Joined the channel
Mon Dec 14 21:32:59 2015  jrushford:jpeach: I rebased that pull request
Mon Dec 14 22:09:02 2015  blattj:Joined the channel
Mon Dec 14 22:25:43 2015  jrickman:Joined the channel
Mon Dec 14 22:33:47 2015  psp:Joined the channel
Mon Dec 14 22:35:32 2015  psp1:Joined the channel
Mon Dec 14 22:58:22 2015  meeramn:Joined the channel
Mon Dec 14 23:08:51 2015  blattj1:Joined the channel
Mon Dec 14 23:35:33 2015  \ask:Joined the channel

Comments