00:19:34  * st_lukejoined
00:27:19  * ralphtheninjaquit (Ping timeout: 246 seconds)
00:49:38  * b4bofitosjoined
00:49:53  * babof1tosquit (Ping timeout: 240 seconds)
01:01:33  * thl0joined
01:40:39  * chiltsquit (Ping timeout: 240 seconds)
01:43:03  * chiltsjoined
01:47:53  * chiltsquit (Ping timeout: 240 seconds)
01:53:42  * chiltsjoined
01:54:55  * dominictarrquit (Quit: dominictarr)
01:58:08  * chiltsquit (Ping timeout: 260 seconds)
01:59:12  * chiltsjoined
02:04:04  * chiltsquit (Ping timeout: 276 seconds)
02:09:40  * st_lukequit (Remote host closed the connection)
02:17:04  * chiltsjoined
02:22:23  * chiltsquit (Ping timeout: 240 seconds)
02:47:44  * chiltsjoined
03:02:25  * thl0quit (Remote host closed the connection)
03:21:27  * timoxleyjoined
04:18:54  * timoxleyquit (Quit: Computer has gone to sleep.)
04:26:14  * eugenewarepart
04:26:23  * eugenewarejoined
05:18:03  * ramitosquit (Ping timeout: 245 seconds)
06:03:54  * timoxleyjoined
06:16:23  * timoxleyquit (Ping timeout: 240 seconds)
06:29:52  * timoxleyjoined
07:12:39  * timoxleyquit (Ping timeout: 260 seconds)
07:33:29  * dominictarrjoined
07:52:51  * babof1tosjoined
07:54:45  * b4bofitosquit (Ping timeout: 248 seconds)
08:01:19  * dominictarrquit (Quit: dominictarr)
08:21:13  * no9quit (Ping timeout: 246 seconds)
08:34:40  * no9joined
08:40:45  * dominictarrjoined
08:43:02  * ChrisPartridgequit (Ping timeout: 240 seconds)
09:03:46  * ralphtheninjajoined
09:20:16  * dominictarrquit (Quit: dominictarr)
09:22:35  * timoxleyjoined
09:25:02  * ChrisPartridgejoined
10:10:32  * werlejoined
10:50:46  * thl0joined
11:17:50  * dominictarrjoined
11:48:42  * owen1quit (Ping timeout: 264 seconds)
11:50:07  * owen1joined
12:26:27  * ChrisPartridgequit (Read error: No buffer space available)
12:26:40  * ChrisPartridgejoined
13:08:34  * b4bofitosjoined
13:09:10  * ChrisPartridgequit (Disconnected by services)
13:09:11  * ChrisPartridgeCljoined
13:10:44  * babof1tosquit (Ping timeout: 264 seconds)
13:15:41  * chilts_joined
13:18:51  * chiltsquit (Write error: Broken pipe)
13:20:04  * ChrisPartridgeClquit (Ping timeout: 276 seconds)
13:20:28  * ChrisPartridgejoined
13:20:51  * no9quit (Ping timeout: 273 seconds)
13:21:32  * Guest82926changed nick to weetabeex
13:21:53  * no9joined
13:22:09  <weetabeex>is anyone aware of segfaults under leveldb::TableBuilder::WriteRawBlock() when the disk runs out of available space?
13:22:46  <weetabeex>or is it common for leveldb to segfault under those circumstances?
13:28:52  * thl0quit (Ping timeout: 246 seconds)
13:38:56  * werle1joined
13:40:42  * werlequit (Ping timeout: 256 seconds)
13:44:34  <dominictarr>weetabeex: you should ask on the leveldb mailing list, the authors of leveldb are on there, and can probably answer such questiots
13:45:04  <weetabeex>sure
13:45:45  <weetabeex>just though that someone around could have an inkling :)
13:55:13  * Acconutjoined
13:58:01  * Acconutquit (Client Quit)
14:08:06  * thl0joined
14:24:06  * no9quit (Ping timeout: 245 seconds)
14:24:35  * Acconutjoined
14:25:35  <juliangruber>dominictarr: i just figured out how to do zero downtime redeploy for processes owning a leveldb
14:25:56  <juliangruber>because for that deployment feature you need to have 2 processes running for a short period of time
14:26:03  <juliangruber>and they can't both own the leveldb
14:26:27  <juliangruber>so what you do is: the process that own's the leveldb opens a multilevel server
14:26:34  <juliangruber>and then you redeploy
14:26:40  <juliangruber>and the new process tries to take the db
14:26:42  <juliangruber>but it can't
14:26:51  <juliangruber>so it connects to the old process's multilevel
14:26:53  <juliangruber>and uses that
14:27:00  <juliangruber>until the old process has been killed
14:27:08  <juliangruber>then it will take ownership of the real db
14:27:12  <dominictarr>right
14:27:18  <dominictarr>hmm
14:28:05  * st_lukejoined
14:28:17  <Acconut>Does this means I can't have a nodejs cluster connecting to one leveldb?
14:29:35  <juliangruber>Acconut: just use multilevel then
14:29:50  <Acconut>juliangruber: Ok, thanks
14:29:52  <juliangruber>my strategy doesn't work if you always have like 4 processes or something
14:29:57  * Acconutquit (Quit: Acconut)
14:31:51  <dominictarr>juliangruber: there is a very similar problem with having open multiple browser tabs
14:32:06  <dominictarr>actually, you could say "precisely similar"
14:32:26  <juliangruber>explain please
14:33:12  <dominictarr>well, for plugins to work, you need to have just one process responsible for them
14:33:33  <dominictarr>indexeddb doesn't emit events :(
14:34:22  <dominictarr>so, if you had multiple tabs open, then they would need to talk to the tab that owns the database.
14:35:05  * werle1quit (Quit: Leaving.)
14:35:36  <juliangruber>oh right
14:35:54  <juliangruber>there needs to be a message bus between the tabs with auto discovery
14:36:19  <dominictarr>I wrote a thing for that
14:36:42  <dominictarr>http://npm.im/tab-stream
14:38:48  <dominictarr>might be able to do it with webrtc too
14:43:54  * timoxleyquit (Ping timeout: 240 seconds)
14:45:04  * no9joined
14:48:52  <dominictarr>juliangruber: another approach is to have use replication
14:49:15  <juliangruber>nice
14:49:16  <juliangruber>oh yeah
14:49:20  <dominictarr>so that your application is designed to remain usable while disconnected
14:49:25  <juliangruber>just always have 2 databases
14:49:36  <juliangruber>the 2nd one will be used during redeploy only
14:49:42  <dominictarr>yes, or that too
14:49:58  <juliangruber>with my initial approach the application is always usable too
14:50:06  <dominictarr>or have a cluster, and roll them over gradually
14:50:09  <juliangruber>could just be slower because it has to use multilevel
14:50:24  <juliangruber>depends on whether you want to cluster or not
14:50:26  <juliangruber>but true
14:58:44  * st_lukequit (Remote host closed the connection)
15:01:54  * werlejoined
15:23:32  <dominictarr>it's good to assume that there can be disconnections, because there are connectivity problems (network partitions all the time)
15:24:02  <dominictarr>if your app can handle that, you only need a static server to send index.html
15:25:07  <dominictarr>and all the smart stuff can you can just stop and then start again
15:25:09  <dominictarr>hmm
15:25:18  <dominictarr>maybe you could use cluster
15:25:29  <dominictarr>and have a static server in master
15:25:39  <dominictarr>and then the app as a child process...
15:46:15  * thl0quit (Read error: Connection reset by peer)
15:47:16  * thl0joined
15:49:17  * thl0quit (Remote host closed the connection)
16:21:31  * no9quit (Ping timeout: 264 seconds)
16:34:05  * no9joined
16:43:31  * dominictarrquit (Quit: dominictarr)
16:46:38  * babof1tosjoined
16:48:16  * b4bofitosquit (Ping timeout: 245 seconds)
16:55:18  * niftylettucequit (Ping timeout: 264 seconds)
16:55:51  <brycebaril>http://hyperdex.org/performance/leveldb/
17:22:41  * no9quit (Ping timeout: 248 seconds)
17:25:15  <chapel>juliangruber: I was thinking about this a few days ago, as far as being able to use levelup with cluster
17:25:48  <chapel>have a plugin or something that when clustered, hosts the levelup on the master, and the children talk to the master for db commands
17:26:28  <chapel>though I could see it getting complicated fast with hooks and other things unrelated to just get/put/delete
17:30:04  <juliangruber>chapel: that's actually pretty trivial
17:30:12  <juliangruber>the master uses multilevel to host it on e.g. a unix socket
17:30:23  <juliangruber>and the children just use the multilevel.client to speak to it
17:34:13  <chapel>juliangruber: well, the problem is if you send a hook to everyone, and lets say a hook is making a change, if it happens that change could be different between processes, that might not be good
17:34:23  <chapel>not to mention adding extra workload
17:35:04  <chapel>in that situation, Id almost rather have the master handle hook related workloads
17:35:09  <juliangruber>oh right
17:35:10  <chapel>and the children only do get/put/delete
17:35:13  <juliangruber>it needs to round robin
17:35:32  <juliangruber>there needs to be a version of multilevel.server that distributes things
17:35:50  * no9joined
17:35:59  <chapel>with multilevel you basically are syncing the same data across?
17:36:19  <juliangruber>yes
17:37:01  <chapel>and it supports master <-> master?
17:37:12  <juliangruber>no
17:37:15  <chapel>hmm
17:37:16  <juliangruber>multilevel is master client
17:37:30  <chapel>kk
17:38:44  <werle>juliangruber: I am loving multilevel dude
17:43:00  * eugenewarequit (Quit: Leaving.)
17:45:45  * thl0joined
18:13:33  * eugenewarejoined
18:18:04  * Acconutjoined
18:23:37  * eugenewarequit (Ping timeout: 276 seconds)
18:50:56  * eugenewarejoined
18:51:34  * Acconutquit (Ping timeout: 276 seconds)
18:53:32  * niftylettucejoined
18:55:27  * eugenewarequit (Ping timeout: 260 seconds)
19:08:52  * thl0_joined
19:09:21  * thl0quit (Ping timeout: 248 seconds)
19:21:41  * eugenewarejoined
19:26:15  * eugenewarequit (Ping timeout: 260 seconds)
19:35:21  * Acconutjoined
19:37:42  * Acconutquit (Client Quit)
19:52:24  * eugenewarejoined
19:56:53  * eugenewarequit (Ping timeout: 248 seconds)
20:23:11  * eugenewarejoined
20:24:29  * Acconutjoined
20:25:14  * chilts_changed nick to chilts
20:26:59  * Acconutquit (Client Quit)
20:27:28  * eugenewarequit (Ping timeout: 246 seconds)
20:50:27  * thl0_quit (Remote host closed the connection)
20:53:52  * eugenewarejoined
20:58:58  * eugenewarequit (Ping timeout: 276 seconds)
21:13:39  * thl0joined
21:24:38  * eugenewarejoined
21:29:31  * eugenewarequit (Ping timeout: 276 seconds)
21:48:41  * chirino_mquit (Ping timeout: 245 seconds)
21:50:21  * chirinojoined
21:55:23  * eugenewarejoined
21:59:31  * eugenewarequit (Ping timeout: 246 seconds)
22:07:22  * thl0quit (Remote host closed the connection)
22:19:59  * CPartridgejoined
22:26:05  * eugenewarejoined
22:30:36  * eugenewarequit (Ping timeout: 261 seconds)
22:36:35  * weetabeexquit (*.net *.split)
22:36:37  * hij1nxquit (*.net *.split)
22:48:14  * weetabeexjoined
22:48:14  * hij1nxjoined
22:53:00  * thl0joined
22:54:23  * thl0quit (Remote host closed the connection)
22:56:51  * eugenewarejoined
23:01:20  * st_lukejoined
23:01:31  * weetabeexquit (*.net *.split)
23:01:32  * hij1nxquit (*.net *.split)
23:02:11  * eugenewarequit (Ping timeout: 298 seconds)
23:03:01  * dominictarrjoined
23:04:49  * weetabeexjoined
23:04:49  * hij1nxjoined
23:07:29  * ChrisPartridgequit (Disconnected by services)
23:07:34  * CPartridgechanged nick to ChrisPartridge
23:27:38  * eugenewarejoined
23:32:11  * eugenewarequit (Ping timeout: 260 seconds)
23:34:37  * chirinoquit (Quit: Leaving...)
23:58:18  * eugenewarejoined