| *** gyee has quit IRC | 00:05 | |
| *** itlinux has joined #openstack-swift | 00:33 | |
| timburke | so i know we've established a precedent of subclassing the trivial/insecure KeyMaster implementation, first in KmsKeyMaster and now KmipKeyMaster... how committed are we to leaving that open for 3rd party libs? | 00:35 |
|---|---|---|
| timburke | 'cause i'm realizing that i *really* don't like having https://github.com/openstack/swift/blob/2.18.0/swift/common/middleware/crypto/keymaster.py#L130-L137 in _get_root_secret... it seems like some pretty horrible conflation of concerns... | 00:36 |
| *** itlinux_ has joined #openstack-swift | 00:54 | |
| *** linkmark has quit IRC | 00:54 | |
| *** itlinux_ has quit IRC | 00:55 | |
| *** itlinux_ has joined #openstack-swift | 00:55 | |
| *** itlinux_ has quit IRC | 00:55 | |
| *** itlinux__ has joined #openstack-swift | 00:56 | |
| *** zaitcev has quit IRC | 01:18 | |
| *** openstackstatus has quit IRC | 01:46 | |
| *** prasen has joined #openstack-swift | 01:48 | |
| *** openstackstatus has joined #openstack-swift | 01:49 | |
| *** ChanServ sets mode: +v openstackstatus | 01:49 | |
| *** hoonetorg has quit IRC | 01:49 | |
| *** hoonetorg has joined #openstack-swift | 02:06 | |
| *** ianychoi has quit IRC | 02:11 | |
| *** hoonetorg has quit IRC | 02:19 | |
| *** ianychoi has joined #openstack-swift | 03:22 | |
| *** nguyenhai_ has joined #openstack-swift | 04:50 | |
| *** nguyenhai has quit IRC | 04:53 | |
| *** links has joined #openstack-swift | 05:02 | |
| *** itlinu___ has joined #openstack-swift | 06:08 | |
| *** itlinu___ has quit IRC | 06:13 | |
| *** ccamacho1 has quit IRC | 06:35 | |
| *** gkadam has joined #openstack-swift | 06:43 | |
| *** silor has joined #openstack-swift | 06:48 | |
| *** silor has quit IRC | 07:16 | |
| *** prasen has quit IRC | 07:54 | |
| *** hoonetorg has joined #openstack-swift | 08:47 | |
| *** geaaru_ has joined #openstack-swift | 08:49 | |
| *** geaaru has quit IRC | 08:51 | |
| *** mikecmpbll has joined #openstack-swift | 09:03 | |
| *** prasen has joined #openstack-swift | 09:26 | |
| *** silor has joined #openstack-swift | 11:40 | |
| *** linkmark has joined #openstack-swift | 11:42 | |
| *** geaaru__ has joined #openstack-swift | 11:45 | |
| *** geaaru_ has quit IRC | 11:46 | |
| *** hoonetorg has quit IRC | 11:49 | |
| *** hoonetorg has joined #openstack-swift | 11:54 | |
| *** hoonetorg has quit IRC | 12:22 | |
| *** hoonetorg has joined #openstack-swift | 12:33 | |
| *** silor has quit IRC | 12:57 | |
| *** silor has joined #openstack-swift | 12:58 | |
| *** silor has quit IRC | 13:25 | |
| *** mvenesio has joined #openstack-swift | 14:10 | |
| *** prasen has quit IRC | 14:11 | |
| mvenesio | hi guys | 14:14 |
| mvenesio | can i decrease the replica size on a running cluster, without affect the service ? | 14:18 |
| *** silor has joined #openstack-swift | 14:19 | |
| mvenesio | we have a cluster almost at 98% | 14:19 |
| mvenesio | another question is, do you think that adding more nodes at this moment can move the data in some way can increase the disk utilization temporary that fill the cluster up | 14:32 |
| *** links has quit IRC | 14:49 | |
| *** silor has quit IRC | 14:59 | |
| *** silor1 has joined #openstack-swift | 14:59 | |
| *** silor1 is now known as silor | 15:01 | |
| DHE | for replication based storage policies you can reduce the replica factor, including fractions if you just want to buy some space... | 15:03 |
| DHE | but... you want to add more storage nodes only temporarily? | 15:05 |
| *** prasen has joined #openstack-swift | 15:23 | |
| *** itlinux has quit IRC | 15:29 | |
| *** gkadam has quit IRC | 15:58 | |
| DHE | mvenesio: ^^ | 16:42 |
| *** nicodemus_ has joined #openstack-swift | 16:46 | |
| *** silor has quit IRC | 16:54 | |
| mvenesio | DHE: no, i will add more nodes forever of course, but the question is, if adding more nodes to a zone can create temporary items in the nearly full nodes that the zone already have | 17:15 |
| *** nguyenhai_ has quit IRC | 17:20 | |
| *** nguyenhai_ has joined #openstack-swift | 17:20 | |
| *** nicodemus_ has quit IRC | 17:22 | |
| DHE | well the ring builder will move some data off current nodes to the new nodes, and then the replicators will actually do the data migrations | 17:24 |
| DHE | but for a little while space usage will be increased globally since there will be n+1 copies of objects for a little while as replication runs before the old locations delete their copies | 17:24 |
| DHE | make sense? | 17:25 |
| mvenesio | DHE: yes make sense but the problem is, this can full fill my disks completely | 17:36 |
| DHE | yes that is a distinct possibility. | 17:38 |
| mvenesio | DHE: and can i delete something manually ? | 17:40 |
| mvenesio | like the.ts | 17:40 |
| DHE | the .ts is a tombstone. it should be 0 bytes and deleting it wouldn't save any substantial space | 17:42 |
| DHE | in fact it can increase disk usage becaues the replication system depends on it to not accidentally replicate a full object back into place of a deleted object | 17:43 |
| mvenesio | I can prevent in some way, that in the rebalance the discs that are now at 98%, fill up completely | 17:45 |
| DHE | nothing comes to mind. I believe the ring builder will strictly move data onto the new hard drives, but you can't prevent, for example, if some hard drive dies during the process the handoff disk could be a full disk | 17:47 |
| *** prasen has quit IRC | 18:17 | |
| *** geaaru_ has joined #openstack-swift | 18:19 | |
| *** geaaru__ has quit IRC | 18:21 | |
| *** hoonetorg has quit IRC | 18:27 | |
| *** mvenesio has quit IRC | 18:30 | |
| *** mvenesio has joined #openstack-swift | 18:41 | |
| mvenesio | DHE: thanks for your help | 18:42 |
| mvenesio | i'm adding the new nodes and see go it goes | 18:42 |
| *** mvenesio has quit IRC | 19:12 | |
| *** itlinux has joined #openstack-swift | 19:16 | |
| *** armaan has joined #openstack-swift | 19:20 | |
| *** geaaru__ has joined #openstack-swift | 19:25 | |
| *** geaaru_ has quit IRC | 19:28 | |
| *** silor has joined #openstack-swift | 19:45 | |
| *** ^andrea^ has joined #openstack-swift | 20:07 | |
| *** itlinux has quit IRC | 20:07 | |
| *** itlinux has joined #openstack-swift | 20:10 | |
| *** silor has quit IRC | 20:16 | |
| *** nguyenhai93 has joined #openstack-swift | 20:59 | |
| *** nguyenhai_ has quit IRC | 21:02 | |
| *** ^andrea^ has quit IRC | 21:29 | |
| *** armaan has quit IRC | 21:32 | |
| *** armaan has joined #openstack-swift | 21:32 | |
| *** ejat has quit IRC | 21:36 | |
| *** armaan has quit IRC | 21:37 | |
| *** itlinux__ has quit IRC | 21:37 | |
| *** ^andrea^ has joined #openstack-swift | 21:40 | |
| *** geaaru__ has quit IRC | 22:26 | |
| *** mikecmpbll has quit IRC | 23:48 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!