| *** sapd1 has joined #senlin | 00:14 | |
| *** sapd1 has quit IRC | 00:18 | |
| *** thainn2 has joined #senlin | 02:04 | |
| *** sapd1 has joined #senlin | 05:04 | |
| sapd1 | dtruong, are you here? | 05:04 |
|---|---|---|
| openstackgerrit | sapd proposed openstack/senlin master: Fix memory leak when create connection with openstacksdk https://review.openstack.org/627215 | 05:08 |
| dtruong | sapd1 yes. I’m here | 06:38 |
| dtruong | Regarding your patch set, I have never seen this problem happen so I’m hesitant to merge this. | 06:42 |
| dtruong | We are also running senlin master and have not encountered this issue. | 06:42 |
| dtruong | Also, we are at feature freeze before final Stein release so we should not be merging any changes unless absolutely necessary. | 06:52 |
| *** thainn2 has quit IRC | 10:52 | |
| *** sapd1_x has joined #senlin | 12:17 | |
| sapd1_x | dtruong: I just have a question. Do you have any health policy is attached to your cluster. Because my co-worker has same issue to me. | 12:32 |
| dtruong | sapd1_x not atm. But I can test that. | 13:56 |
| dtruong | What type of detection mode are you using in your health policy? | 13:56 |
| sapd1_x | dtruong: yes. if we do not attach any health policy. we will not get this issue. | 13:56 |
| dtruong | Also, do you have this change in your version of senlin: https://review.openstack.org/#/c/634811/ | 13:57 |
| dtruong | This was merged 10 days ago to master and has fixes to how the health manager works. | 13:58 |
| sapd1_x | dtruong: I have tried this. | 14:02 |
| sapd1_x | I use NODE_STATUS_POLLING type for health policy | 14:02 |
| dtruong | So your version of senlin is same as master upstream? | 14:10 |
| sapd1_x | yes. just installed master branch. | 14:10 |
| *** thainn2 has joined #senlin | 14:16 | |
| *** thainn2 has quit IRC | 14:17 | |
| dtruong | What is your interval setting in your heath policy? | 14:26 |
| dtruong | Actually can you copy your health policy to paste.openstack.org and link it here? | 14:27 |
| sapd1_x | dtruong: about 1 minute. | 16:22 |
| sapd1_x | dtruong: because I would like to detect fail node asap. | 16:23 |
| *** sapd1_x has quit IRC | 16:35 | |
| dtruong | sapd1_x how many clusters / nodes are you running? | 17:11 |
| dtruong | Running the health checks at 1 min interval might be exhausting your connection pool with a lot of nodes. | 17:13 |
| *** openstack has joined #senlin | 19:45 | |
| *** ChanServ sets mode: +o openstack | 19:45 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!