| *** yangyapeng has quit IRC | 00:05 | |
| *** fawadkhaliq has quit IRC | 00:09 | |
| *** yangyapeng has joined #openstack-freezer | 01:13 | |
| *** yangyapeng has quit IRC | 01:15 | |
| *** ddieterly has quit IRC | 01:18 | |
| *** EinstCrazy has joined #openstack-freezer | 01:27 | |
| *** fawadkhaliq has joined #openstack-freezer | 02:32 | |
| *** fawadkhaliq has quit IRC | 03:41 | |
| *** daemontool has quit IRC | 03:57 | |
| *** daemontool has joined #openstack-freezer | 03:58 | |
| *** fawadkhaliq has joined #openstack-freezer | 05:18 | |
| *** openstackgerrit has quit IRC | 06:03 | |
| *** openstackgerrit has joined #openstack-freezer | 06:03 | |
| *** yangyapeng has joined #openstack-freezer | 06:31 | |
| yangyapeng | https://review.openstack.org/#/c/312430/ | 06:31 |
|---|---|---|
| *** EinstCrazy has quit IRC | 06:46 | |
| *** EinstCrazy has joined #openstack-freezer | 06:46 | |
| *** lpetrut has joined #openstack-freezer | 07:54 | |
| daemontool | slashme, sorry I missed your message from yesterday about pep8 | 07:55 |
| daemontool | I'm not sure | 07:55 |
| daemontool | we should ask to infra | 07:55 |
| *** fawadkhaliq has quit IRC | 08:06 | |
| *** electrofelix has joined #openstack-freezer | 08:56 | |
| openstackgerrit | Saad Zaher proposed openstack/freezer: Adds checksum before backup and after restore for consistency check. https://review.openstack.org/300080 | 09:07 |
| *** lpetrut has quit IRC | 09:07 | |
| *** lpetrut has joined #openstack-freezer | 09:08 | |
| *** szaher has left #openstack-freezer | 09:10 | |
| *** szaher has joined #openstack-freezer | 09:10 | |
| *** lpetrut has quit IRC | 09:14 | |
| *** jonaspf has quit IRC | 09:22 | |
| *** etualuo_ has joined #openstack-freezer | 09:37 | |
| etualuo_ | Hi all, i am the newbie in this freezer project | 09:38 |
| etualuo_ | i would like to ask you guys if you ddo not mind | 09:39 |
| etualuo_ | the problems are the solutions of monitoring in freezer agent (in order to detect the host failed) and how to trigger the migration/fencing in this case? | 09:40 |
| etualuo_ | thanks in advanced | 09:40 |
| *** lpetrut has joined #openstack-freezer | 09:47 | |
| yangyapeng | etualuo_: Can you have a detailed instructions？ monitor ? | 10:10 |
| etualuo_ | Hi, thank you for replying | 10:11 |
| etualuo_ | actually i follow the information of freezer before | 10:12 |
| yangyapeng | etualuo_ :) | 10:12 |
| etualuo_ | but i did not see the improvement of monitoring, fencing part that tell us the status of the nodes | 10:12 |
| etualuo_ | then freezer can trigger the restore | 10:12 |
| etualuo_ | backup, etc. | 10:13 |
| etualuo_ | ddo we have the final solution of using Monasca | 10:13 |
| etualuo_ | ? | 10:13 |
| yangyapeng | etualuo_: freezer is tools for backup restore. There is no integrated monitoring，now. | 10:16 |
| yangyapeng | ping m3m0 vannif_ szaher daemontool | 10:17 |
| etualuo_ | Awesome Yang | 10:17 |
| etualuo_ | :) | 10:17 |
| etualuo_ | so it means Freezer does not automatically trigger the backup, restore | 10:17 |
| etualuo_ | we put this business to the operators to trigger it? | 10:18 |
| yangyapeng | freezer backup or restore can use freezer-scheduler | 10:18 |
| etualuo_ | i understand the architectture | 10:19 |
| yangyapeng | but monitoring fails then trigger is not powerless 。 | 10:19 |
| yangyapeng | Do you want to backup restore openstack ? | 10:20 |
| etualuo_ | yeap | 10:20 |
| yangyapeng | instance or volume ? | 10:20 |
| etualuo_ | actually, my concern is how freezer understand that a host is failed | 10:20 |
| etualuo_ | then triggers the backup session | 10:20 |
| etualuo_ | ? | 10:20 |
| yangyapeng | There is no to implement this function. :) | 10:34 |
| yangyapeng | ping m3m0 | 10:34 |
| etualuo_ | thank you Yang:) | 10:36 |
| yangyapeng | :) | 10:37 |
| etualuo_ | so I should write a spec for that, huh" | 10:37 |
| etualuo_ | ? | 10:37 |
| etualuo_ | :D | 10:37 |
| yangyapeng | hmm thank you your contribution ha-ha | 10:37 |
| etualuo_ | :D. Yeah, from last year, people trieed to discuss about it. I also tried to follow it but it seems that up to now, these functionalites are out or Freezer scope | 10:38 |
| yangyapeng | But we can add some operational functions, or other project and assist to implement | 10:39 |
| yangyapeng | I finished my work, the evening freezer and meeting . | 10:40 |
| yangyapeng | https://etherpad.openstack.org/p/freezer_meetings | 10:40 |
| *** EinstCrazy has quit IRC | 10:41 | |
| *** yangyapeng_ has joined #openstack-freezer | 10:43 | |
| *** yangyapeng has quit IRC | 10:46 | |
| *** etualuo_ has left #openstack-freezer | 10:48 | |
| *** yangyapeng_ has quit IRC | 10:54 | |
| *** yangyapeng_ has joined #openstack-freezer | 10:56 | |
| *** yangyapeng_ has quit IRC | 11:00 | |
| *** ddieterly has joined #openstack-freezer | 11:25 | |
| openstackgerrit | Eldar Nugaev proposed openstack/freezer: Remove auto_snap https://review.openstack.org/311073 | 11:43 |
| *** ddieterly is now known as ddieterly[away] | 12:02 | |
| daemontool | Hi, I'll be late at the meeting today by ~20 min, sorry | 12:26 |
| *** ddieterly[away] is now known as ddieterly | 12:26 | |
| *** ddieterly is now known as ddieterly[away] | 12:30 | |
| szaher | Hi ddieterly | 12:30 |
| szaher | ddieterly: Please, take a look and let me know what do you think about this https://review.openstack.org/#/c/300080/22/freezer/tests/freezer_tempest_plugin/tests/api/test_swift_backup.py | 12:31 |
| *** ddieterly[away] is now known as ddieterly | 12:37 | |
| *** ddieterly has quit IRC | 12:46 | |
| *** ddieterly has joined #openstack-freezer | 12:48 | |
| *** ddieterly has quit IRC | 12:49 | |
| *** yangyapeng has joined #openstack-freezer | 12:59 | |
| *** yangyapeng has quit IRC | 13:14 | |
| *** ddieterly has joined #openstack-freezer | 13:22 | |
| *** reldan has joined #openstack-freezer | 13:26 | |
| *** reldan has quit IRC | 13:33 | |
| *** reldan has joined #openstack-freezer | 13:34 | |
| *** yangyapeng has joined #openstack-freezer | 13:35 | |
| *** daemontool_ has joined #openstack-freezer | 13:35 | |
| *** reldan has quit IRC | 13:36 | |
| *** daemontool has quit IRC | 13:37 | |
| *** lpetrut has quit IRC | 13:37 | |
| *** reldan has joined #openstack-freezer | 13:45 | |
| *** ddieterly is now known as ddieterly[away] | 13:45 | |
| *** openstackgerrit has quit IRC | 13:47 | |
| *** openstackgerrit has joined #openstack-freezer | 13:48 | |
| *** EinstCrazy has joined #openstack-freezer | 13:49 | |
| *** lpetrut has joined #openstack-freezer | 13:50 | |
| *** ddieterly[away] is now known as ddieterly | 13:52 | |
| EinstCrazy | meeting？ | 14:00 |
| slashme | Yes | 14:00 |
| slashme | Let's move to openstack-meeting-alt guys | 14:00 |
| *** lpetrut has quit IRC | 14:57 | |
| daemontool_ | ddieterly, good meeting! | 15:01 |
| ddieterly | really? | 15:01 |
| ddieterly | well, thanks | 15:01 |
| slashme | yep, was great! | 15:01 |
| ddieterly | i think i just suck at these types of things | 15:01 |
| ddieterly | you guys are just saying that so i'll do it next time and you don't have to ;-) | 15:02 |
| vannif_ | :D | 15:02 |
| reldan | Please review https://review.openstack.org/#/c/311073/4 | 15:04 |
| *** sxc731 has joined #openstack-freezer | 15:04 | |
| vannif_ | are we going to handle the job concurrency ? it originates from the issue about restoring a resource while a job to backup that very resource is running | 15:05 |
| vannif_ | possible strategies: have all the freezer-invocations go through the scheduler. if it's running, it'll check if the jobs can be executed concurrently. atm it doesn't even need to check since there is no parallel execution of jobs | 15:07 |
| vannif_ | or, we could have freezer check if there's already a snapshot of the resource and exit with a warning in that case (which means we could have a --force flag to force the execution) | 15:08 |
| sxc731 | Hi guys I think I just missed the weekly mtg on #openstack-meeting-alt. The instructions say 3pm UTC but I suppose it's more like 3pm GMT/BST? (It's now 3:08 UTC but 4:08 BST) | 15:09 |
| ddieterly | yea, it's 3:00 PM British daylight time | 15:10 |
| ddieterly | i think | 15:10 |
| ddieterly | 'british summer time' or something like that | 15:11 |
| sxc731 | Ta I'll know for next week ;-) and maybe someone can update the instructions? | 15:11 |
| sxc731 | Yeah that's BST | 15:12 |
| ddieterly | vannif_ i think that the schedulers should be aware of what is being run and avoid collisions | 15:16 |
| ddieterly | vannif_ then that info could be gathered in the db and queried by the user thru the ui | 15:17 |
| ddieterly | so, the db would contain all the state of the freezer system | 15:17 |
| vannif_ | the use case is: there are backup jobs scheduler (and possibly running), the user issues a restore action using the freezer-agent (so bypassing the scheduler) | 15:18 |
| ddieterly | the agent needs to talk to the scheduler | 15:18 |
| ddieterly | to avoid collisions | 15:19 |
| vannif_ | I know it feels like "don't push both the brake and the accelerator pedals or you'll damage your car" :) | 15:19 |
| ddieterly | so, the info needs to go from the agent to the scheduler and vice versa | 15:20 |
| ddieterly | so that neither the agent nor the scheduler stomps on each other | 15:20 |
| *** sxc731 has quit IRC | 15:21 | |
| *** lpetrut has joined #openstack-freezer | 15:21 | |
| vannif_ | so the agent should search for any scheduler instance and interact with it ? it's rather have the user always interact with the scheduler even when he needs to execute immediately an action. In this case the scheduler is a kind of "wrapper" around the agent. this "decoration" would add the functionality to check any concurrency issues | 15:22 |
| vannif_ | s/it's rather/I'd rather | 15:23 |
| vannif_ | reldan , thoughts ? | 15:23 |
| reldan | vannif_: I don’t quite understand problem actually. Why we should have problems with concurrency? We can generate different mount pathes. Or I am missing something? | 15:25 |
| reldan | vannif_: If it is not related to snapshot. Let’s implement rule - one backup at the time | 15:28 |
| reldan | Let’s have lock | 15:28 |
| vannif_ | the idea is that having a backup and a restore of the very same resource might be a problem. the user might fill the space available for the snapshot | 15:30 |
| vannif_ | because the restore action actively writes to the logical volume beoing snapshotted | 15:31 |
| reldan | But it is a common problem. What if we want to do backup and restore of the same directory at the same time | 15:31 |
| reldan | not only related to snapshots | 15:31 |
| vannif_ | so, it's advisable to have a way to avoid this. the problem arises when the two actions are taken by two different players: the scheduler, and the user on the comman-line | 15:31 |
| reldan | We should have a global mark - it may be a file | 15:32 |
| reldan | with pid | 15:32 |
| reldan | of current freezer process | 15:32 |
| reldan | if we are trying to start a new process - we should check for that file | 15:32 |
| vannif_ | yes. I know. That's why it feels like an advice not to use both the brake and the accelerator. the right foot is supposed to know what the left foot is doing. but it seems it's not always the case :) | 15:32 |
| reldan | and if pid is not more available - start process with recreation of file | 15:33 |
| reldan | Only lock from my point of view | 15:33 |
| reldan | Othervise - I’m doing backup of dir test and restor of dir test | 15:33 |
| reldan | it may create inconsistent backup | 15:33 |
| *** lpetrut has quit IRC | 15:35 | |
| *** lpetrut has joined #openstack-freezer | 15:35 | |
| vannif_ | yes. in case a snapshot is not used the concurrent backup-and-restore is *bad* :) | 15:41 |
| *** jonaspf has joined #openstack-freezer | 15:42 | |
| *** jonaspf has quit IRC | 15:49 | |
| *** jonaspf has joined #openstack-freezer | 15:49 | |
| *** jonaspf has quit IRC | 15:50 | |
| *** jonaspf has joined #openstack-freezer | 15:50 | |
| *** jonaspf has quit IRC | 15:55 | |
| *** sxc731 has joined #openstack-freezer | 15:55 | |
| *** jonaspf has joined #openstack-freezer | 15:55 | |
| *** sxc731 has quit IRC | 15:58 | |
| m3m0 | yangyapeng: you here? | 15:58 |
| *** jonaspf has quit IRC | 16:00 | |
| *** jonaspf has joined #openstack-freezer | 16:01 | |
| *** lpetrut has quit IRC | 16:01 | |
| *** jonaspf has quit IRC | 16:03 | |
| *** jonaspf has joined #openstack-freezer | 16:03 | |
| *** yangyapeng has quit IRC | 16:08 | |
| *** yangyape_ has joined #openstack-freezer | 16:08 | |
| m3m0 | slashme, so, mode will be renamed as application? | 16:38 |
| slashme | I think that's what we decided | 16:50 |
| *** lpetrut has joined #openstack-freezer | 16:58 | |
| *** reldan has quit IRC | 17:00 | |
| *** ddieterly is now known as ddieterly[away] | 17:21 | |
| openstackgerrit | Saad Zaher proposed openstack/freezer: Adds checksum before backup and after restore for consistency check. https://review.openstack.org/300080 | 17:27 |
| *** electrofelix has quit IRC | 17:59 | |
| *** vannif_ has quit IRC | 18:01 | |
| *** lpetrut has quit IRC | 18:15 | |
| *** sxc731 has joined #openstack-freezer | 18:18 | |
| *** ddieterly[away] is now known as ddieterly | 18:18 | |
| *** ddieterly is now known as ddieterly[away] | 18:23 | |
| *** lpetrut has joined #openstack-freezer | 18:25 | |
| *** sxc731 has quit IRC | 18:26 | |
| *** reldan has joined #openstack-freezer | 18:36 | |
| *** ddieterly[away] is now known as ddieterly | 18:44 | |
| *** daemontool_ has quit IRC | 20:01 | |
| *** lpetrut has quit IRC | 20:32 | |
| *** daemontool_ has joined #openstack-freezer | 20:35 | |
| *** lpetrut has joined #openstack-freezer | 21:03 | |
| *** fawadkhaliq has joined #openstack-freezer | 21:04 | |
| *** lpetrut has quit IRC | 21:08 | |
| *** ddieterly is now known as ddieterly[away] | 21:17 | |
| *** fawadkhaliq has quit IRC | 21:33 | |
| *** reldan_ has joined #openstack-freezer | 21:37 | |
| *** reldan has quit IRC | 21:37 | |
| *** reldan_ is now known as reldan | 21:37 | |
| *** ddieterly[away] is now known as ddieterly | 21:41 | |
| *** reldan has quit IRC | 22:39 | |
| *** ddieterly has quit IRC | 23:08 | |
| *** reldan has joined #openstack-freezer | 23:08 | |
| *** reldan has quit IRC | 23:27 | |
| *** reldan has joined #openstack-freezer | 23:43 | |
| *** EinstCrazy has quit IRC | 23:48 | |
| *** reldan_ has joined #openstack-freezer | 23:48 | |
| *** reldan has quit IRC | 23:49 | |
| *** reldan_ is now known as reldan | 23:49 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!