*** openstackgerrit has quit IRC | 00:01 | |
*** openstackgerrit has joined #openstack-cue | 00:02 | |
*** openstack has joined #openstack-cue | 00:03 | |
*** abitha has quit IRC | 00:11 | |
dkalleg | hey harlowja , looking into some zookeeper stuff, trying to use tooz to get a lock. It asks for a connection string and a member id. I'm not sure what the member id should be. In my devstack deployment (with zookeeper) my /var/lib/zookeeper/myid is actually uninitialized. | 00:32 |
---|---|---|
harlowja | dkalleg hey | 00:33 |
dkalleg | vipul: tells me you might be able to help with zk stuff :) | 00:33 |
harlowja | possibly | 00:33 |
harlowja | haha | 00:33 |
harlowja | so myid in the var/lib dir is more of the zookeeper identifier for the zookeepr server (used when its in a cluster) | 00:33 |
dkalleg | ok, so likely not related to the memberid | 00:33 |
harlowja | the memberid in tooz is the member 'name' of the client | 00:33 |
dkalleg | of what client? | 00:34 |
harlowja | sorry, the tooz coordinator object u are fetching | 00:34 |
harlowja | its not really needed for locks though i think, but its required for group membership apis | 00:34 |
harlowja | but typically a uuid u make is fine | 00:34 |
harlowja | we (tooz folks) probably could make the member_id optional | 00:35 |
harlowja | because for locks i don't think its used | 00:35 |
dkalleg | oh ok cool | 00:36 |
dkalleg | I did notice in the tooz tests that a random uuid is generated & passed in for the member id | 00:36 |
dkalleg | Great, let me give it a shot with some random value. | 00:37 |
harlowja | k | 00:37 |
dkalleg | Thanks harlowja! | 00:37 |
harlowja | can u open a bug on https://bugs.launchpad.net/python-tooz 'something something don't need member_id for locks, don't require it'.... | 00:37 |
harlowja | think it can be investigated and or made optional | 00:38 |
dkalleg | sure no problem | 00:40 |
dkalleg | also, the get_lock() doesn't seem to be complaining, so I'll assume it worked | 00:40 |
harlowja | woot | 00:42 |
dkalleg | hm.. i even gave a bogus backend url in get_coordinator() then started & get_lock() and lock.acquired is True | 00:44 |
dkalleg | Is that expected? | 00:45 |
harlowja | hmmmm, how bogus was the backend url, lol | 00:45 |
harlowja | medium bogus? | 00:45 |
harlowja | or very bogus? | 00:46 |
dkalleg | zake://444.444.444.444:2181 | 00:46 |
dkalleg | tried changing port too, same result | 00:46 |
harlowja | ya, so zake pretty much ignores all the hostname, port | 00:46 |
harlowja | since zake is pretty much all local 'fake/mock' zookeeper | 00:46 |
harlowja | pretty sure in that case only zake:// is actually used (to figure out which driver to load) | 00:47 |
dkalleg | gave 'foobar' and it failed | 00:47 |
harlowja | the rest is just unsued | 00:47 |
harlowja | *unused | 00:47 |
dkalleg | so zake://:2181 is really all it reads? Whatever in the middle is stripped out? | 00:47 |
harlowja | just zake:// actually in this case | 00:47 |
harlowja | the rest is read, but not used | 00:48 |
dkalleg | ok | 00:48 |
harlowja | certain drivers use different things from that url | 00:48 |
harlowja | probably should have better docs on what is used from the url(s) | 00:48 |
dkalleg | So that means there's a requirement that wherever you're running tooz must be on the same host as a zk node? | 00:48 |
harlowja | negative | 00:49 |
harlowja | zake is a different one | 00:49 |
dkalleg | oh. I'm still learning zake/zk/tooz :) | 00:49 |
harlowja | zake basically mimics the kazoo (zookeeper client) api with a implmentation that doesn't actually talk to zookeeper | 00:49 |
harlowja | zake version is mostly for testing (where no zookeeper can be ensured exists) | 00:49 |
dkalleg | oh! so if I want to get a lock from my running zookeeper service, I should not be using zake. Is that so? | 00:50 |
harlowja | u probably want zookeeper://localhost:2181 as the driver | 00:50 |
harlowja | right | 00:50 |
dkalleg | :) | 00:50 |
dkalleg | feel a bit silly now | 00:50 |
harlowja | np | 00:50 |
dkalleg | progress! Thanks, will go try that out | 00:50 |
harlowja | cools | 00:50 |
harlowja | https://github.com/openstack/tooz/blob/master/setup.cfg#L26 (Various drivers) | 00:51 |
harlowja | zake is the special one in that group | 00:51 |
harlowja | *more special* they are all special obviously, lol | 00:51 |
dkalleg | aha, this certainly seems to be working. My oslo_service is also spitting out kazoo.client zookeeper connection CONNECTED / CLOSED | 00:52 |
harlowja | thats good (hopefully, ha) | 00:53 |
harlowja | probabaly don't make to many coordinators dynamically, since they each contain a client | 00:53 |
harlowja | which consumes resources | 00:53 |
dkalleg | Ok | 00:54 |
dkalleg | For what I'm doing, I'll be trying to acquire a lock every 60s from 1 to 3 nodes, but I can keep the coordinator around so that I only ever create 1 to 3 coordinators at init time | 00:55 |
dkalleg | Thanks a bunch! Big help | 00:55 |
harlowja | np :) | 00:57 |
harlowja | and tell your friends, tooz is cool | 00:58 |
harlowja | lol | 00:58 |
dkalleg | You got it. You guys have stickers? haha, I'm all about the sticker | 01:01 |
dkalleg | s | 01:01 |
harlowja | lol | 01:06 |
harlowja | not yet, perhaps we should get some, ha | 01:06 |
*** dkalleg has quit IRC | 01:15 | |
*** HanSolace has joined #openstack-cue | 03:20 | |
*** esmute has joined #openstack-cue | 04:27 | |
*** dkalleg has joined #openstack-cue | 04:53 | |
*** HanSolace has quit IRC | 04:54 | |
*** dkalleg has quit IRC | 04:58 | |
*** HanSolace has joined #openstack-cue | 05:01 | |
*** HanSolace has quit IRC | 05:44 | |
*** esmute has quit IRC | 05:46 | |
*** jhesketh has quit IRC | 06:34 | |
*** jhesketh has joined #openstack-cue | 06:35 | |
*** dkalleg has joined #openstack-cue | 06:42 | |
*** dkalleg has quit IRC | 06:46 | |
*** dkalleg has joined #openstack-cue | 08:30 | |
*** dkalleg has quit IRC | 08:34 | |
*** dkalleg has joined #openstack-cue | 10:19 | |
*** dkalleg has quit IRC | 10:24 | |
*** dkalleg has joined #openstack-cue | 12:08 | |
*** dkalleg has quit IRC | 12:12 | |
openstackgerrit | Victor Stinner proposed openstack/cue: py3: Get urllib modules from six.moves https://review.openstack.org/229131 | 12:43 |
openstackgerrit | Victor Stinner proposed openstack/cue: py3: Replace unicode with six.text_type https://review.openstack.org/229130 | 12:55 |
*** dkalleg has joined #openstack-cue | 13:57 | |
*** dkalleg has quit IRC | 14:01 | |
*** sputnik13 has joined #openstack-cue | 14:47 | |
*** HanSolace has joined #openstack-cue | 14:56 | |
*** sputnik13 has quit IRC | 15:14 | |
*** sputnik13 has joined #openstack-cue | 15:16 | |
*** HanSolace has quit IRC | 15:16 | |
*** dkalleg has joined #openstack-cue | 15:46 | |
*** dkalleg has quit IRC | 15:50 | |
*** ViswaV has joined #openstack-cue | 15:58 | |
*** ViswaV has quit IRC | 16:02 | |
*** ViswaV has joined #openstack-cue | 16:03 | |
*** ViswaV has quit IRC | 16:35 | |
openstackgerrit | Victor Stinner proposed openstack/cue: py3: Get urllib modules from six.moves https://review.openstack.org/229131 | 16:37 |
*** harlowja has quit IRC | 16:52 | |
*** dkalleg has joined #openstack-cue | 17:06 | |
*** abitha has joined #openstack-cue | 17:07 | |
*** abitha has quit IRC | 17:19 | |
*** abitha has joined #openstack-cue | 17:20 | |
*** abitha has quit IRC | 17:24 | |
*** abitha has joined #openstack-cue | 17:25 | |
*** abitha has quit IRC | 17:27 | |
*** abitha has joined #openstack-cue | 17:28 | |
*** ViswaV has joined #openstack-cue | 18:05 | |
*** ViswaV has quit IRC | 18:09 | |
*** ViswaV has joined #openstack-cue | 18:10 | |
*** harlowja has joined #openstack-cue | 18:13 | |
*** openstackgerrit has quit IRC | 18:16 | |
*** openstackgerrit has joined #openstack-cue | 18:17 | |
*** openstackgerrit has quit IRC | 18:46 | |
*** openstackgerrit has joined #openstack-cue | 18:47 | |
*** ViswaV has quit IRC | 19:12 | |
*** HanSolace has joined #openstack-cue | 19:46 | |
*** sputnik13 has quit IRC | 20:04 | |
*** ViswaV has joined #openstack-cue | 20:20 | |
*** ViswaV has quit IRC | 20:24 | |
*** ViswaV has joined #openstack-cue | 20:25 | |
*** sputnik13 has joined #openstack-cue | 20:32 | |
*** dkalleg has quit IRC | 20:45 | |
*** ViswaV has quit IRC | 21:52 | |
*** ViswaV has joined #openstack-cue | 21:56 | |
*** ViswaV_ has joined #openstack-cue | 21:58 | |
*** ViswaV has quit IRC | 22:00 | |
*** dkalleg has joined #openstack-cue | 22:33 | |
*** dkalleg has quit IRC | 23:12 | |
*** dkalleg has joined #openstack-cue | 23:13 | |
*** dkalleg has quit IRC | 23:15 | |
*** HanSolace has quit IRC | 23:48 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!