*** hoangcx has joined #openstack-ha | 00:50 | |
*** cknight1 has joined #openstack-ha | 01:28 | |
*** cknight has quit IRC | 01:30 | |
*** hoangcx has quit IRC | 01:32 | |
*** cknight1 has quit IRC | 02:18 | |
*** sigmavirus24_awa is now known as sigmavirus24 | 02:40 | |
*** sigmavirus24 is now known as sigmavirus24_awa | 02:49 | |
*** sigmavirus24_awa is now known as sigmavirus24 | 03:09 | |
*** smoriya_ has quit IRC | 03:23 | |
*** smoriya_ has joined #openstack-ha | 03:28 | |
*** cknight has joined #openstack-ha | 03:47 | |
*** sigmavirus24 is now known as sigmavirus24_awa | 04:35 | |
*** hoangcx has joined #openstack-ha | 04:35 | |
*** cknight has quit IRC | 04:38 | |
*** mjura has joined #openstack-ha | 07:06 | |
*** mobaer has joined #openstack-ha | 07:07 | |
*** dgurtner has joined #openstack-ha | 08:20 | |
*** dgurtner has joined #openstack-ha | 08:20 | |
*** dgurtner has quit IRC | 08:29 | |
*** dgurtner has joined #openstack-ha | 08:29 | |
*** dgurtner has joined #openstack-ha | 08:29 | |
*** hoangcx has quit IRC | 09:51 | |
openstackgerrit | Bogdan Dobrelya proposed openstack/ha-guide: Updates to Galera Cluster documentation in HA Guide. https://review.openstack.org/263075 | 10:03 |
---|---|---|
openstackgerrit | Bogdan Dobrelya proposed openstack/ha-guide: Updates to Galera Cluster documentation in HA Guide. https://review.openstack.org/263075 | 10:11 |
openstackgerrit | Maria Zlatkova proposed openstack/ha-guide: Updates to Galera Cluster documentation in HA Guide. https://review.openstack.org/263075 | 10:29 |
*** ntt has joined #openstack-ha | 10:33 | |
ntt | Hi, I'm trying to use openstack-cinder-volume in an active/passive configuration with icehouse. Is this possible? Actually I have a san (active/passive) and I want to install only openstack-cinder-volume on the san | 10:34 |
ntt | At the moment I have an error: "ERROR: Cinder Volume is not connected to the AMQP server: 1" | 10:34 |
ntt | someone can help me please? | 10:34 |
*** mobaer has quit IRC | 10:36 | |
*** smoriya_ has quit IRC | 10:39 | |
ntt | dulek: I think I found the problem, but I need help to solve | 10:50 |
ntt | actually the resource does: su cinder -s /bin/sh -c "cinder-volume --config-file=/etc/cinder/cinder.conf" > /var/run/resource-agents/default.pid | 10:51 |
ntt | but the problem is that /var/run/resource-agents/default.pid is empty (because I'm using rhel7, I think) | 10:52 |
dulek | ntt: I'm not an expert in OpenStack OCF agents, but maybe _gryf, ddeja or bogdando will be able to help you? | 10:53 |
*** mobaer has joined #openstack-ha | 10:53 | |
*** jpena has joined #openstack-ha | 10:55 | |
_gryf | ntt, the error message suggest the possible culprit. did you try just to run the command that resource is trying to start (so the `su cinder -s /bin/sh -c "cinder-volume --config-file=/etc/cinder/cinder.conf"') to see what happen? | 10:58 |
ntt | dulek:ok. More details: pid file is not empty, but has the wrong pid. example: when I launch cinder-volume, I have 2 processes (31977 and 31988). In /var/run/resource-agents/default.pid I have 31977 but the ampq is connected with the pid 31988 and the check fails in cinder_volume_monitor routine in the RA. | 11:00 |
ntt | cinder_volume_monitor does: netstat -punt | grep -s "$OCF_RESKEY_amqp_server_port" and the pid is 31988 | 11:01 |
ntt | _gryf: the process works well if I try to launch from bash | 11:02 |
bogdando | ntt, dulek hm, I never tried the OCF for a cinder volume. Would be nice to have a quick set-up with vagrant for each OCF RA | 11:02 |
bogdando | like I did for the rabbit ocf | 11:02 |
bogdando | I use this repo https://github.com/bogdando/packer-atlas-example/blob/master/Vagrantfile to spawn a fast n1 n2 cluster with OCF RA running | 11:03 |
bogdando | I believe it can be easily modified to run any OCF RA one want to test/check/reproduce an issue | 11:04 |
bogdando | and here I put post-up configs for OCF RA related things https://github.com/bogdando/packer-atlas-example/tree/master/vagrant_script | 11:05 |
bogdando | so feel free to open a bug and provide details on how to reproduce, and put logs if any :) | 11:07 |
ntt | actually the only thing I did is: | 11:09 |
ntt | #volume_amqp_check=`netstat -punt | grep -s "$OCF_RESKEY_amqp_server_port" | grep -s "$pid" | grep -qs "ESTABLISHED"` | 11:09 |
ntt | modified to volume_amqp_check=`netstat -punt | grep -s "$OCF_RESKEY_amqp_server_port" | grep -qs "ESTABLISHED"` | 11:09 |
ntt | but this is really ugly fix | 11:09 |
ntt | now it "works" | 11:09 |
bogdando | well, there is 100% a bug with pid tracking, so you may want to open it as well | 11:12 |
bogdando | https://bugs.launchpad.net/openstack-resource-agents | 11:13 |
aspiers | yes please file a bug! | 11:21 |
aspiers | ntt: sounds like the OCF RA is failing to account for a cinder-volume subthread or child process being spawned? | 11:23 |
aspiers | ntt: what is the difference between 31977 and 31988? | 11:23 |
ntt | 31988 is child of 31977 | 11:23 |
aspiers | yes usre | 11:23 |
aspiers | sure | 11:23 |
ntt | nothing else | 11:23 |
aspiers | but what is that process? | 11:23 |
aspiers | what is the command line? | 11:23 |
aspiers | if it's the same command/args as 31977 then it's a subthread | 11:24 |
ntt | /usr/bin/python /usr/bin/cinder-volume --config-file /usr/share/cinder/cinder-dist.conf --config-file /etc/cinder/cinder.conf --logfile /var/log/cinder/volume.log | 11:24 |
ntt | no diffs | 11:24 |
aspiers | ok | 11:24 |
ntt | but with netstat I can see that the port 5672 is used by 31988 and not by 31977 | 11:24 |
aspiers | so the correct way would be to examine the process tree I guess | 11:25 |
aspiers | e.g. pstree -p | 11:25 |
ntt | aspiers: my approach is completely wrong. | 12:16 |
aspiers | right, I was suggesting a possible right approach :) | 12:16 |
ntt | /etc/cinder/volumes should be shared from all nodes in the cluster. I known that this is unrelated to the pid rpboelm | 12:16 |
ntt | problem* | 12:17 |
ntt | aspiers: you are welcome | 12:17 |
openstackgerrit | Merged openstack/ha-guide: Updates to Galera Cluster documentation in HA Guide. https://review.openstack.org/263075 | 12:19 |
ntt | I think that volumes_dir should be shared by the san nodes | 12:19 |
ntt | I can create an LV inside the cinder-volumes VG of small size and mount that LV before start cinder-volume | 12:19 |
ntt | I don't known if this is a good idea | 12:19 |
ntt | In this way I should have high available volumes.... even though openstack-cinder-volume is not: adding "include $volumes_dir" in /etc/tgt/targets.conf on both nodes of the san, I should have targets automagically created in case of failover :) | 12:23 |
openstackgerrit | OpenStack Proposal Bot proposed openstack/ha-guide: Updated from openstack-manuals https://review.openstack.org/268107 | 12:30 |
openstackgerrit | Merged openstack/ha-guide: Updated from openstack-manuals https://review.openstack.org/268107 | 12:39 |
*** openstackgerrit has quit IRC | 12:50 | |
*** openstackgerrit has joined #openstack-ha | 12:51 | |
*** jpena has quit IRC | 13:49 | |
*** dslev has joined #openstack-ha | 13:53 | |
*** cknight has joined #openstack-ha | 14:27 | |
*** kgaillot has joined #openstack-ha | 14:49 | |
*** sigmavirus24_awa is now known as sigmavirus24 | 15:24 | |
*** dslev has quit IRC | 16:20 | |
*** jpena has joined #openstack-ha | 17:16 | |
*** dgurtner has quit IRC | 17:17 | |
*** sigmavirus24 is now known as sigmavirus24_awa | 17:46 | |
*** cknight has quit IRC | 17:53 | |
*** sigmavirus24_awa is now known as sigmavirus24 | 17:57 | |
*** cknight has joined #openstack-ha | 17:57 | |
*** dgurtner has joined #openstack-ha | 18:11 | |
*** dgurtner has joined #openstack-ha | 18:11 | |
*** cknight has quit IRC | 18:31 | |
*** dslev has joined #openstack-ha | 18:34 | |
*** cknight has joined #openstack-ha | 18:35 | |
*** mobaer has quit IRC | 18:58 | |
*** mobaer has joined #openstack-ha | 19:04 | |
*** dgurtner has quit IRC | 19:05 | |
*** hoonetorg has quit IRC | 19:06 | |
*** hoonetorg has joined #openstack-ha | 19:19 | |
*** dslev has quit IRC | 20:24 | |
*** cknight has quit IRC | 21:40 | |
*** jpena is now known as jpena|sleep | 22:26 | |
*** sigmavirus24 is now known as sigmavirus24_awa | 22:50 | |
*** mobaer has quit IRC | 22:53 | |
*** mobaer1 has joined #openstack-ha | 22:53 | |
openstackgerrit | KATO Tomoyuki proposed openstack/ha-guide: Indicate the rationale for running cinder in active-passive mode https://review.openstack.org/266200 | 23:30 |
openstackgerrit | KATO Tomoyuki proposed openstack/ha-guide: Indicate the rationale for running cinder in active-passive mode https://review.openstack.org/266200 | 23:31 |
openstackgerrit | KATO Tomoyuki proposed openstack/ha-guide: Document why RabbitMQ should not be used via HAProxy https://review.openstack.org/266202 | 23:31 |
openstackgerrit | KATO Tomoyuki proposed openstack/ha-guide: Details for configuring cinder on RHEL based systems https://review.openstack.org/266201 | 23:34 |
openstackgerrit | KATO Tomoyuki proposed openstack/ha-guide: Details for configuring cinder on RHEL based systems https://review.openstack.org/266201 | 23:37 |
*** kgaillot has quit IRC | 23:40 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!