| *** vnaboychenko has quit IRC | 01:05 | |
| *** vnaboychenko has joined #magnetodb | 01:59 | |
| *** openstackgerrit has quit IRC | 02:02 | |
| *** openstackgerrit has joined #magnetodb | 02:03 | |
| *** k4n0 has joined #magnetodb | 06:03 | |
| openstackgerrit | Illia Khudoshyn proposed a change to stackforge/magnetodb: Add recognition of 'INSERT IF NOT EXISTS' https://review.openstack.org/116937 | 07:06 |
|---|---|---|
| *** vnaboychenko has quit IRC | 07:08 | |
| *** rushiagr_away is now known as rushiagr | 08:08 | |
| *** ajayaa has joined #magnetodb | 08:09 | |
| *** rushiagr is now known as rushiagr_away | 08:12 | |
| *** boris-42 has quit IRC | 08:21 | |
| isviridov | Hello everybody | 08:24 |
| *** boris-42 has joined #magnetodb | 08:33 | |
| *** ikhudoshyn has joined #magnetodb | 09:28 | |
| dukhlov | Hi | 09:34 |
| ajayaa | hi all. | 09:35 |
| ajayaa | I am trying to reproduce https://bugs.launchpad.net/magnetodb/+bug/1351068 | 09:35 |
| ikhudoshyn | 0/ | 09:35 |
| ajayaa | Any help would be appreciated. | 09:36 |
| ikhudoshyn | https://review.openstack.org/#/c/116937/ guys, could you take a look once gerrit verifies that? | 09:36 |
| ajayaa | As far as I understand, in the bug we are trying to get an item by secondary index and secondary index does not exist. | 09:37 |
| ajayaa | That should not be 500. | 09:37 |
| dukhlov | ajayaa: hi! | 09:41 |
| ajayaa | dukhlov, hi | 09:41 |
| dukhlov | yes, It seems you are right | 09:41 |
| dukhlov | do you have problem to reproduce it? | 09:41 |
| ajayaa | dukhlov, I don't know how to query with a secondary index? I will paste my table schema somewhere. | 09:42 |
| ajayaa | http://paste.openstack.org/show/101604/ | 09:43 |
| ajayaa | I could probably look through the test files. | 09:46 |
| isviridov | Hello ajayaa | 09:49 |
| ajayaa | isviridov, hi | 09:49 |
| isviridov | Are you familiar with postman Chrome plugin? https://chrome.google.com/webstore/detail/postman-rest-client/fdmmgilgnpjigdojojpjoooidkmcomcm | 09:50 |
| isviridov | I have sample requests for it and can share with you | 09:50 |
| isviridov | Here it is https://www.getpostman.com/collections/27123d8110f2ff9216f0 | 09:50 |
| ajayaa | isviridov, I am familiar with sending REST api calls. | 09:50 |
| ajayaa | I use "Advanced REST Client" btw. I am not sure whether I will be able to import this. | 09:51 |
| ajayaa | let me install POSTMAN | 09:51 |
| dukhlov | ok, first of all you data model is not valid because local secondary indices are not allowed on tables with HASH key only (you could specify 'last_name' attribute as RANGE key and then you primary index). | 09:59 |
| dukhlov | hm, if you created such table successfully, I guess it is a bug and we need to add validation for this case | 10:00 |
| *** isviridov is now known as isviridov_awa | 10:00 | |
| *** isviridov_awa is now known as isviridov_awat | 10:00 | |
| *** isviridov_awat is now known as isviridov_away | 10:00 | |
| ajayaa | dukhlov, It is a bug then. | 10:03 |
| openstackgerrit | Dmitriy Ukhlov proposed a change to stackforge/magnetodb: Fix small bugs with cql quating and unicode encoding https://review.openstack.org/116034 | 10:11 |
| ajayaa | dukhlov, btw I don't think there is a bug. I have created local secondary index with both HASH and RANGE keys. | 10:14 |
| dukhlov | ajayaa, I meant table with primary key which consists of HASH key only | 10:19 |
| dukhlov | in your example you have created table with HASH key only primary key | 10:20 |
| ajayaa | I didn't know that. Why is that? Why wouldn't you allow such a query? | 10:20 |
| ajayaa | Yes. Primary key is hash only. | 10:20 |
| dukhlov | because it do nothing | 10:24 |
| dukhlov | if you query by index you should specify hash key and then add some condition for for LSI RANGE key | 10:26 |
| dukhlov | but if your table has HASH key only it means that each hash key stores only single item | 10:26 |
| dukhlov | and there nothing to query | 10:26 |
| *** isviridov_away is now known as isviridov | 10:37 | |
| ajayaa | isviridov, I sent the query "query email_data by Author and Date" with a non-existent index name and it returns 400. | 10:47 |
| isviridov | What the mdb log output? | 10:55 |
| ikhudoshyn | https://bugs.launchpad.net/magnetodb/+bug/1362549 | 10:58 |
| isviridov | ikhudoshyn, i see. Any other information? Stacktrace, how many requests? | 11:05 |
| ikhudoshyn | I'll update once I have the full info | 11:05 |
| isviridov | Great. Till then it is marked as 'Incomplete' | 11:07 |
| ajayaa | isviridov, http://paste.openstack.org/show/101675/ | 11:45 |
| openstackgerrit | A change was merged to stackforge/magnetodb: Add recognition of 'INSERT IF NOT EXISTS' https://review.openstack.org/116937 | 12:04 |
| isviridov | ajayaa, do you have index with name foo in your table? | 12:12 |
| ajayaa | No. | 12:13 |
| ajayaa | isridov, | 12:13 |
| ajayaa | I believe that is the bug. I should get 500 instead of correct validation and return. | 12:13 |
| ajayaa | https://bugs.launchpad.net/magnetodb/+bug/1351068 | 12:14 |
| isviridov | ajayaa, actually 400 is correct response code. Means Bad Request | 12:20 |
| *** ominakov has joined #magnetodb | 12:20 | |
| ajayaa | Yes. I am trying to reproduce the above bug. | 12:20 |
| ajayaa | I thought If I send an get_item with non-existing secondary index, I will be able to reproduce the bug. | 12:21 |
| isviridov | Seems it has been fixed with refactoring of validation | 12:21 |
| ajayaa | isviridov, okay. cool! | 12:22 |
| isviridov | Good catch! Just confirm and we will resolve it | 12:22 |
| ajayaa | The bug should be marked fix-commited or something equivalent to it. | 12:23 |
| isviridov | Let me do it. Whats is your launchpad id? | 12:23 |
| ajayaa | https://launchpad.net/~ajayaa | 12:23 |
| ajayaa | isviridov, What do you want me to do? It is already fixed. :) | 12:25 |
| isviridov | ajayaa, you have finished with it. Bug verification is also job, and I think you have done it. Apparently, I've assigned it to you. | 12:26 |
| isviridov | Marked as Fix Commited | 12:27 |
| ajayaa | isviridov, cool! I will try to find some other bug to kill. | 12:29 |
| isviridov | ajayaa, great. The prioriti for now is juno-3 milestone https://launchpad.net/magnetodb/+milestone/juno-3 and we are fighting for user expierence, so 'ux' tagged bugs are in the focus now https://bugs.launchpad.net/magnetodb/+bugs?field.tag=ux | 12:31 |
| isviridov | #offtopic Russian troops in Ukraine https://mapsengine.google.com/map/u/0/embed?mid=zjOaq88N3N_g.knEcWTZi7F_M | 12:35 |
| *** ominakov has quit IRC | 12:42 | |
| ajayaa | isviridov, the bug https://bugs.launchpad.net/magnetodb/+bug/1348336 is also not reproducible. | 12:47 |
| ajayaa | I tested with the third attribute both present in secondary index and not present in secondary index. | 12:48 |
| isviridov | This bag is assigned to aostapenko, please comment your test results. I believe he will come to IRC today | 12:52 |
| ajayaa | isviridov, yes there is a bug. I misunderstood it. | 12:54 |
| ajayaa | isviridov, Shall I go ahead and fix it? | 12:54 |
| isviridov | Better to talk to aostapenko first. | 12:57 |
| isviridov | IIRC he not working on it, but better to check | 12:57 |
| *** ominakov has joined #magnetodb | 12:57 | |
| isviridov | Hey ominakov, is aostapenko around? | 12:57 |
| ominakov | hey isviridov, not yet | 12:58 |
| *** achuprin has joined #magnetodb | 13:04 | |
| achuprin | hi team! | 13:04 |
| isviridov | Hi achuprin. Any news about gate? | 13:05 |
| achuprin | isviridov, I created new patch set https://review.openstack.org/#/c/112304/ | 13:06 |
| achuprin | isviridov, review in progress | 13:07 |
| isviridov | Great! | 13:07 |
| *** k4n0 has quit IRC | 13:10 | |
| *** aostapenko has joined #magnetodb | 13:10 | |
| *** ajayaa has quit IRC | 13:11 | |
| isviridov | !m achuprin | 13:17 |
| [o__o] | You're doing good work, achuprin! | 13:17 |
| openstack | isviridov: Error: "m" is not a valid command. | 13:17 |
| isviridov | [o__o], help | 13:17 |
| openstackgerrit | Dmitriy Ukhlov proposed a change to stackforge/magnetodb: Fix small bugs with cql quating and unicode encoding https://review.openstack.org/116034 | 13:21 |
| *** charlesw has joined #magnetodb | 13:48 | |
| aostapenko | Hi, ajayaa! | 13:54 |
| *** ominakov has quit IRC | 14:03 | |
| *** ominakov has joined #magnetodb | 14:03 | |
| *** denis_makogon has joined #magnetodb | 14:08 | |
| denis_makogon | Hi, guys | 14:09 |
| denis_makogon | Can we discuss how are we going to accomplish backup procedure? | 14:09 |
| ominakov | denis_makogon, hello | 14:15 |
| denis_makogon | ominakov, 'sup =) | 14:15 |
| charlesw | Good topic, denis_makogon, we can schedule a time so more people can attend | 14:16 |
| achudnovets | denis_makogon: hi! | 14:16 |
| ikhudoshyn | denis_makogon, hi | 14:16 |
| denis_makogon | charlesw, hello, i just wanted to share sore knowledge about how does cassandra does backups | 14:17 |
| dukhlov | hello, denis_makogon | 14:17 |
| denis_makogon | charlesw, other things we can discuss within wide community | 14:17 |
| ominakov | denis_makogon, sounds good | 14:17 |
| denis_makogon | achudnovets, dukhlov, hi, guys | 14:17 |
| ikhudoshyn | denis_makogon, I guess I got more experience in C* backups | 14:17 |
| ikhudoshyn | denis_makogon, you could start | 14:18 |
| denis_makogon | cassandra uses for backups nodetool CLI util (see http://www.datastax.com/documentation/cassandra/2.0/cassandra/tools/toolsSnapShot.html) | 14:19 |
| ikhudoshyn | but if you think it would be a lot of typing u could start with a letter to maillist | 14:19 |
| denis_makogon | it allows to do backups of all keyspaces, certain keyspace, and per-CF backup | 14:19 |
| denis_makogon | ikhudoshyn, sure i can do that | 14:20 |
| ikhudoshyn | denis_makogon, I dot't actually want to stop you)) it's just a hint | 14:21 |
| denis_makogon | ikhudoshyn, sure | 14:21 |
| ikhudoshyn | either way for us will be good. But if u start with an email we could schedule a discussion in advance, as charlesw suggested | 14:22 |
| denis_makogon | the only one impact i see that we need to have direct access via SSH onto each node and snapshotting/collecting files/sending them to remote storage | 14:22 |
| denis_makogon | ikhudoshyn, sure, if it works for all of you, then i'm ok | 14:22 |
| dukhlov | denis_makogon: is Cassandra node stopped for backup time? | 14:22 |
| denis_makogon | dukhlov, no downtime | 14:23 |
| denis_makogon | dukhlov, to accomplish consistency nodetool provides an ability to flush un-commited changes into file system | 14:23 |
| ikhudoshyn | denis_makogon, so, r we to proceed now or you'll start with the letter? | 14:26 |
| dukhlov | cassandra node uses sstables and commit logs for persistance, memtables, and in memory caches for fast request processing | 14:26 |
| dukhlov | could you provide deeper information what cassandra actually backups? | 14:27 |
| dukhlov | what is procedure? | 14:27 |
| ikhudoshyn | what i'm really interested in is process description, inputs and outputs | 14:27 |
| dukhlov | so when cassandra receives backup request what is performed step by step? | 14:28 |
| denis_makogon | dukhlov, snapshotting doesn't involve cassandra at all | 14:28 |
| ikhudoshyn | dukhlov, I thought this is a side tool, not a part of C* process, denis_makogon am I wrong? | 14:28 |
| denis_makogon | ikhudoshyn, correct | 14:29 |
| ikhudoshyn | if so, I suggest first to concentrate on tool invocation, prerequisites and artifacts produced | 14:29 |
| denis_makogon | ok | 14:30 |
| denis_makogon | let me describe whole procedure of backup in ML | 14:30 |
| ikhudoshyn | denis_makogon, sounds good | 14:31 |
| ominakov | denis_makogon, and one more question about SSH onto each node. What we need is a SSH or nodetool will suffice? | 14:32 |
| denis_makogon | ominakov, we need to do ssh'ing | 14:33 |
| denis_makogon | at least once, to call snapshot command from any cluster node for each node in the rin | 14:33 |
| denis_makogon | but when we need to collect data we would need to ssh onto each node | 14:34 |
| dukhlov | ok, main point of my interest is how we nodetool achieve snapshotting without downtime (for example it is not clear to me why compaction or memtable flushing doesn't broke our backup) | 14:37 |
| dukhlov | also it is very important to know about potential problems known for you. | 14:38 |
| *** bogeyon18 has joined #magnetodb | 14:41 | |
| *** tnurlygayanov has quit IRC | 14:42 | |
| *** ominakov has quit IRC | 15:28 | |
| *** ominakov has joined #magnetodb | 15:29 | |
| aostapenko | isviridov: i've decided not to add new filter for time measuring and just to probe existing filters. Actually I think that to measure connection_handler filter will be enough to understand the situation. Do you have any thoughts about than? | 15:51 |
| openstackgerrit | A change was merged to stackforge/magnetodb: Integration test scenarios https://review.openstack.org/114877 | 15:54 |
| isviridov | aostapenko, what the problem with time measure filter? | 15:59 |
| isviridov | aostapenko, I don't like an approach to add code to folters when can avoid it. I would siggest to implement 2 filters: one for measuring whole request and one for measuring specific filter, working as wrapper on target and delegating the call. | 16:03 |
| isviridov | aostapenko, whta do you think? | 16:03 |
| *** isviridov is now known as isviridov_away | 16:18 | |
| aostapenko | isviridov: I need some time for investigation about 2nd filter. I like your approach | 16:27 |
| *** bogeyon1_ has joined #magnetodb | 16:38 | |
| *** ominakov has quit IRC | 16:39 | |
| *** bogeyon18 has quit IRC | 16:40 | |
| *** ominakov has joined #magnetodb | 16:40 | |
| *** vnaboychenko has joined #magnetodb | 16:44 | |
| *** ominakov has quit IRC | 16:58 | |
| *** achuprin has quit IRC | 17:01 | |
| *** aostapenko has quit IRC | 17:08 | |
| *** denis_makogon has quit IRC | 17:10 | |
| *** vnaboychenko has quit IRC | 17:24 | |
| *** vnaboychenko has joined #magnetodb | 17:24 | |
| *** bogeyon1_ has quit IRC | 17:28 | |
| *** vnaboychenko has quit IRC | 17:29 | |
| *** vnaboychenko has joined #magnetodb | 18:08 | |
| *** vnaboychenko has quit IRC | 18:09 | |
| *** vnaboychenko has joined #magnetodb | 18:18 | |
| *** achuprin has joined #magnetodb | 18:45 | |
| *** ominakov has joined #magnetodb | 18:59 | |
| *** bogeyon18 has joined #magnetodb | 19:01 | |
| *** openstackgerrit has quit IRC | 19:31 | |
| *** openstackgerrit has joined #magnetodb | 19:33 | |
| *** bogeyon18 has quit IRC | 19:41 | |
| *** bogeyon18 has joined #magnetodb | 20:10 | |
| *** bogeyon1_ has joined #magnetodb | 20:14 | |
| *** bogeyon18 has quit IRC | 20:14 | |
| *** bogeyon1_ has quit IRC | 20:25 | |
| *** ominakov has quit IRC | 20:30 | |
| *** ominakov has joined #magnetodb | 20:39 | |
| *** bogeyon18 has joined #magnetodb | 21:06 | |
| *** bogeyon18 has quit IRC | 21:21 | |
| *** bogeyon18 has joined #magnetodb | 21:23 | |
| *** bogeyon18 has quit IRC | 21:32 | |
| *** achuprin has quit IRC | 21:39 | |
| *** ominakov has quit IRC | 21:42 | |
| *** aostapenko has joined #magnetodb | 21:56 | |
| *** ominakov has joined #magnetodb | 22:11 | |
| *** igormarnat has quit IRC | 22:13 | |
| *** ominakov has quit IRC | 22:16 | |
| *** igormarnat has joined #magnetodb | 22:16 | |
| *** charlesw has quit IRC | 22:51 | |
| *** ominakov has joined #magnetodb | 22:52 | |
| *** ominakov has quit IRC | 23:41 | |
| *** ominakov has joined #magnetodb | 23:43 | |
| *** openstackgerrit has quit IRC | 23:46 | |
| *** openstackgerrit has joined #magnetodb | 23:48 | |
| *** bogeyon18 has joined #magnetodb | 23:57 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!