| openstackgerrit | Merged openstack/swift master: Fail short reads in SegmentedIterable https://review.opendev.org/754814 | 02:17 |
|---|---|---|
| openstackgerrit | Merged openstack/swift master: Log storage policy with s3api requests https://review.opendev.org/755587 | 02:19 |
| *** evrardjp has quit IRC | 04:33 | |
| *** evrardjp has joined #openstack-swift | 04:33 | |
| openstackgerrit | Merged openstack/swift master: wsgi: Ensure _response_headers is a list https://review.opendev.org/755903 | 05:27 |
| openstackgerrit | Tim Burke proposed openstack/swift master: squash: More logging refactoring; stop needing an extra catch_errors https://review.opendev.org/755906 | 05:35 |
| openstackgerrit | Tim Burke proposed openstack/swift master: squash: More logging refactoring; stop needing an extra catch_errors https://review.opendev.org/755906 | 06:21 |
| openstackgerrit | Merged openstack/swift master: Use bigger GreenPool for concurrent EC https://review.opendev.org/754412 | 06:54 |
| *** fyx has quit IRC | 07:24 | |
| *** fyx has joined #openstack-swift | 07:30 | |
| *** mikecmpbll has quit IRC | 08:51 | |
| *** mikecmpbll has joined #openstack-swift | 08:55 | |
| *** recyclehero has left #openstack-swift | 10:03 | |
| *** mikecmpbll has quit IRC | 10:27 | |
| *** Hamidreza has joined #openstack-swift | 14:27 | |
| Hamidreza | Hi | 14:27 |
| Hamidreza | I've mount my disks with bellow options on fstab: | 14:28 |
| Hamidreza | noatime,nodiratime,nobarrier,logbufs=8 0 2 | 14:28 |
| Hamidreza | but when I commit this command , it shows the changed and modified time !!! | 14:29 |
| Hamidreza | stat /srv/node/sdb/objects/1600: | 14:30 |
| Hamidreza | Modify: 2020-10-02 09:43:18 | 14:30 |
| Hamidreza | Change: 2020-10-02 09:43:18 | 14:30 |
| Hamidreza | Birth: - | 14:30 |
| DHE | yeah... because atime is access time, not modified time (mtime) or changed time (ctime) | 14:31 |
| Hamidreza | ok | 14:32 |
| Hamidreza | how can i disble them? | 14:32 |
| DHE | you can't | 14:32 |
| Hamidreza | but I've access time ! | 14:32 |
| Hamidreza | ]Access: 2020-10-02 09:43:18 | 14:32 |
| DHE | it doesn't stop the field from existing | 14:33 |
| Hamidreza | why i can't? | 14:33 |
| DHE | it stops it from being updated on a file/dir read | 14:33 |
| Hamidreza | you mean I can not update the object any more if I disable (mtime) and (ctime) | 14:34 |
| Hamidreza | ? | 14:34 |
| Hamidreza | I read a text that said we have atime enabled in kernel by default | 14:36 |
| Hamidreza | relatime* | 14:37 |
| DHE | relatime cheats atime. it only updates atime if either 1) it's been 12 hours since the last update or 2) current atime is before mtime | 14:39 |
| DHE | since most applications are mainly interested on whether atime < mtime or atime > mtime it works out pretty well | 14:39 |
| DHE | but swift itself doesn't care about these fields. at least I'm pretty sure it doesn't. timestamps for objects are stored differently | 14:40 |
| Hamidreza | I want to disable mtime and ctime for my disks] | 14:44 |
| Hamidreza | Because of increasing performance | 14:44 |
| DHE | you can't | 14:45 |
| DHE | but I don't see how this would make it faster | 14:45 |
| Hamidreza | does swift need this ctime and mtime? | 14:47 |
| DHE | I just answered that question | 14:47 |
| Hamidreza | You said timestamps for objects are stored differently | 14:48 |
| Hamidreza | you didn't say why swift need this info | 14:49 |
| DHE | because swift resolves races of uploading objects based on upload time | 14:49 |
| DHE | which is not necessarily the same as a file's mtime/ctime because files can be copied around between hosts or there could be network latency, etc | 14:49 |
| Hamidreza | I want to delete old objects based on this info | 14:51 |
| Hamidreza | I mean the objects with 2years ago access time | 14:51 |
| Hamidreza | but im not sure this is true or not | 14:51 |
| DHE | well if you have noatime then you can't trust the atime value | 14:51 |
| Hamidreza | but I said that by default relatime is enabled on new kernels | 14:53 |
| Hamidreza | but I'm concern about relatime because maybe swift proxy always routed request to one specific disk | 14:54 |
| Hamidreza | and it can confuse my statistics | 14:55 |
| *** Hamidreza has quit IRC | 15:36 | |
| *** donnyd has quit IRC | 15:42 | |
| *** gmann has quit IRC | 15:43 | |
| *** gmann has joined #openstack-swift | 15:46 | |
| *** donnyd has joined #openstack-swift | 15:53 | |
| *** rcernin has joined #openstack-swift | 18:16 | |
| *** rcernin has quit IRC | 18:24 | |
| *** mikecmpbll has joined #openstack-swift | 18:40 | |
| *** mikecmpbll has quit IRC | 19:13 | |
| *** rcernin has joined #openstack-swift | 22:14 | |
| *** rcernin has quit IRC | 22:27 | |
| openstackgerrit | Merged openstack/swift stable/ussuri: py3: Fix swift-dispersion-populate https://review.opendev.org/754852 | 22:50 |
| *** rcernin has joined #openstack-swift | 22:53 | |
| *** rcernin has quit IRC | 23:07 | |
| *** zamba has quit IRC | 23:17 | |
| *** zamba has joined #openstack-swift | 23:21 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!