Perl 6 - the future is here, just unevenly distributed

IRC log for #gluster, 2016-10-28

| Channels | #gluster index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:37 ahino joined #gluster
00:47 shyam joined #gluster
00:53 shdeng joined #gluster
00:57 RameshN joined #gluster
01:06 kramdoss_ joined #gluster
01:30 hagarth joined #gluster
01:36 derjohn_mobi joined #gluster
01:38 Pupeno joined #gluster
01:46 kramdoss_ joined #gluster
02:06 d0nn1e joined #gluster
02:38 sadbox joined #gluster
02:39 jeremyh joined #gluster
02:40 Lee1092 joined #gluster
02:44 haomaiwang joined #gluster
02:45 haomaiwa_ joined #gluster
02:47 blu__ joined #gluster
02:52 legreffier joined #gluster
03:03 magrawal joined #gluster
03:39 Pupeno joined #gluster
03:41 nbalacha joined #gluster
03:51 sanoj_ joined #gluster
03:58 hgowtham joined #gluster
04:01 Saravanakmr joined #gluster
04:10 shubhendu joined #gluster
04:20 sankarshan_ joined #gluster
04:23 Muthu joined #gluster
04:38 plarsen joined #gluster
04:46 [diablo] joined #gluster
05:00 poornima_ joined #gluster
05:05 ndarshan joined #gluster
05:12 jiffin joined #gluster
05:17 ashiq joined #gluster
05:18 [diablo] joined #gluster
05:23 Debloper joined #gluster
05:30 gem joined #gluster
05:30 ppai joined #gluster
05:31 karthik_us joined #gluster
05:39 aravindavk joined #gluster
05:50 kdhananjay joined #gluster
05:51 itisravi joined #gluster
05:54 karnan joined #gluster
05:57 gem joined #gluster
06:02 mhulsman joined #gluster
06:03 Bhaskarakiran joined #gluster
06:11 ankitraj joined #gluster
06:19 hchiramm joined #gluster
06:22 arc0 joined #gluster
06:28 riyas joined #gluster
06:29 rafi joined #gluster
06:30 Philambdo joined #gluster
06:35 atinm joined #gluster
06:41 nigelb joined #gluster
06:44 Muthu joined #gluster
06:56 nigelb so, if I'm getting this error when trying to peer probe, what am I doing wrong?
06:56 nigelb peer probe: failed: Probe returned with Transport endpoint is not connected
06:56 ivan_rossi joined #gluster
06:57 nigelb oh, I might just be running into a firewall.
06:58 misc the error message could indeed be improved
07:00 nigelb misc: yep
07:00 msvbhat joined #gluster
07:03 nishanth joined #gluster
07:04 mhulsman joined #gluster
07:10 DV joined #gluster
07:10 Muthu joined #gluster
07:17 prth joined #gluster
07:30 fsimonce joined #gluster
07:33 hackman joined #gluster
07:34 nishanth_afk joined #gluster
07:40 kramdoss_ joined #gluster
07:45 devyani7_ joined #gluster
07:47 Muthu joined #gluster
07:52 mhulsman joined #gluster
07:59 riyas joined #gluster
08:08 devyani7_ joined #gluster
08:22 panina joined #gluster
08:28 Slashman joined #gluster
08:33 mhulsman joined #gluster
08:45 Philambdo joined #gluster
08:45 rastar joined #gluster
08:45 Javezim joined #gluster
08:51 derjohn_mobi joined #gluster
08:52 flying joined #gluster
09:00 arc0 joined #gluster
09:04 cloph JoeJulian: you wrote "I would also recommend turning off client-side heals and just relying on the self-heal daemon for healing: cluster.{data,metadata,entry}-self-heal off" - would that increase performance? (concurrent access is not an issue, volume used for qemu disk images, so only accessed by one process/client at a time)
09:07 Debloper joined #gluster
09:07 armyriad joined #gluster
09:10 Pupeno joined #gluster
09:12 prth joined #gluster
09:15 jkroon joined #gluster
09:15 rafi1 joined #gluster
09:23 devyani7_ joined #gluster
09:28 kramdoss_ joined #gluster
09:37 rastar joined #gluster
09:43 satya4ever joined #gluster
09:49 rafi1 joined #gluster
09:58 kramdoss_ joined #gluster
10:01 rafi joined #gluster
10:14 rafi joined #gluster
10:27 panina joined #gluster
10:42 prasanth joined #gluster
10:47 Bhaskarakiran joined #gluster
10:47 ankitraj joined #gluster
10:57 david_ joined #gluster
10:58 david_ hi all, i'm running gluster 3.5.4. Does anyone know if I have different sizes brick in Distributed Volume, will it a problem? i.e: 2 bricks with 10TB, Brick C with 20TB
10:59 david_ will gluster distribute more files to Brick C to make the most of the space ?
10:59 david_ thanks
11:00 nbalacha david_, the current version will. But I don't think that behaviour was available in 3.5.x
11:10 nishanth joined #gluster
11:11 partner yeah it was feature for 3.6 to support "heterogeneous bricks"
11:12 nbalacha joined #gluster
11:15 david_ oh no.
11:15 david_ we actually have 10 x 10TB bricks
11:15 david_ that means we have to either upgrade all bricks, or upgrade the software
11:17 gem joined #gluster
11:44 Norky_ joined #gluster
11:45 Norky joined #gluster
11:49 karolyi joined #gluster
11:50 karolyi hey, I have a question: can use any object storage (aws s3/google bucket) for bricks in a glusterfs pool?
11:50 jiffin ppai: ^^
11:50 karolyi so basically, I don't want to use persistent disks to back up glusterfs, but object storages
11:52 ppai karolyi, glusterfs does not provide that "backup to object storage" internally. You could use external tools for that.
11:53 karolyi ppai: not "backup to", but instead of using persistent disk storage provided by cloud providers, I'd like to use object storages as bricks
11:53 karolyi but I guess that's a no then, just like I thought
11:54 ppai karolyi, no you cannot use object storage as bricks. glusterfs needs a filesystem in the backend
11:54 karolyi basically what I'm pondering on is to create a docker image that mounts a google bucket with gcsfuse and exports it on nfs
11:55 karolyi this way I could manage that countainer with k8s
11:55 ppai karolyi, even if was possible, wouldn't it be terribly slow ? check this out https://github.com/s3fs-fuse/s3fs-fuse
11:55 glusterbot Title: GitHub - s3fs-fuse/s3fs-fuse: FUSE-based file system backed by Amazon S3 (at github.com)
11:56 ppai why'd you need gluster to export the gcsfuse mount over nfs ?
11:57 karolyi ppai: not necessarily with gluster, I was considering glusterfs as an option. I need to export it so I can get k8s storage out of it
11:57 karolyi ppai: so the general problem is, kubernetes lacks a good storage provider
11:57 karolyi (readwritemany I meant)
11:57 ppai glusterfs is a good fit to give persistent storage in k8s env
11:58 ppai glusterfs also has a object interface to it
11:58 karolyi yes... with persistent disks, fo which you'll have to create your own managing instance
11:58 ppai right
11:58 karolyi that's what I'd like to avoid
11:59 karolyi ppai: don't get me wrong, I like the idea of glusterfs, but I don't like the idea of having an instance just to manage it, which I couldn't do with kubernetes
12:00 karolyi btw the fuse mounts involve caching and in theory, could be managed via k8s
12:00 karolyi (I say in theory because I'm just trying to put together a container for it to see if it works)
12:01 ppai karolyi, I get your point. Your best bet would be s3fs or gcsfs mount and export it over NFS. But expect it to be really slow. At best, it'll suite sequential workloads
12:01 ppai suit*
12:02 rafi joined #gluster
12:03 karolyi it's worth a try I'd think
12:04 karolyi so you have dedicated instances as glusterfs servers?
12:05 ppai karolyi, correct, nodes with bricks (disks) are designated to run gluster containers
12:06 ppai karnan, oh and btw, if nfs ganesha supported s3/gcs as backend FSAL, it would probably be a good option. Let me check if it does...
12:08 panina joined #gluster
12:09 karolyi well there are certain constraints, no posix compatibility for example
12:09 Gugge_ joined #gluster
12:09 karolyi I don't know about ganesha btw
12:09 ppai it looks like doesn't have a cloud backend, yet. https://github.com/nfs-ganesha/​nfs-ganesha/tree/next/src/FSAL
12:09 glusterbot Title: nfs-ganesha/src/FSAL at next · nfs-ganesha/nfs-ganesha · GitHub (at github.com)
12:10 squeakyneb_ joined #gluster
12:11 arc0life_ joined #gluster
12:12 ppai one could write a s3/gcs FSAL for ganesha by looking at ceph rgw implementation as reference
12:13 karolyi I'm not a C professional unfortunately
12:14 karolyi but ganesha looks good anyway, I could do everything in usermode
12:15 johnmilton joined #gluster
12:18 pfactum joined #gluster
12:22 jiffin joined #gluster
12:23 marlinc joined #gluster
12:23 unclemarc joined #gluster
12:25 flying joined #gluster
12:32 RameshN joined #gluster
12:34 magrawal_ joined #gluster
12:35 eryc joined #gluster
12:35 eryc joined #gluster
12:35 mrEriksson joined #gluster
12:35 partner joined #gluster
12:35 malevolent joined #gluster
12:35 yalu joined #gluster
12:36 juo joined #gluster
12:36 Marbug joined #gluster
12:36 Sebbo2 joined #gluster
12:36 scuttle|afk joined #gluster
12:36 renout joined #gluster
12:36 d-fence_ joined #gluster
12:36 rossdm joined #gluster
12:36 ahino joined #gluster
12:37 ic0n_ joined #gluster
12:37 plarsen joined #gluster
12:37 ic0n_ joined #gluster
12:37 ankitraj joined #gluster
12:39 luizcpg joined #gluster
12:41 plarsen joined #gluster
12:47 rafi1 joined #gluster
12:50 luizcpg left #gluster
12:55 mchangir joined #gluster
12:56 mchangir amye, can I publish my blogs about glusterfind on blog.gluster.org or publish a link to my blog at blog.gluster.org ?
12:57 poornima_ joined #gluster
12:58 nishanth joined #gluster
12:58 jerrcs_ joined #gluster
13:00 RameshN joined #gluster
13:04 nthomas_ joined #gluster
13:07 eMBee joined #gluster
13:08 shyam joined #gluster
13:16 nishanth joined #gluster
13:18 Lee1092 joined #gluster
13:23 nishanth joined #gluster
13:28 skylar joined #gluster
13:38 fsimonce joined #gluster
13:44 nishanth joined #gluster
13:47 nbalacha joined #gluster
14:01 nbalacha joined #gluster
14:08 armyriad joined #gluster
14:10 nishanth joined #gluster
14:15 karolyi left #gluster
14:16 nishanth joined #gluster
14:20 ankitraj joined #gluster
14:28 squizzi joined #gluster
14:28 unclemarc joined #gluster
14:29 ira joined #gluster
14:34 shaunm joined #gluster
14:40 farhorizon joined #gluster
15:02 Gnomethrower joined #gluster
15:07 hackman joined #gluster
15:09 wushudoin joined #gluster
15:22 derjohn_mobi joined #gluster
15:35 om2_ joined #gluster
15:37 JoeJulian mchangir: blog.gluster.org can federate your blog. Amye will need an rss link that filters for gluster related articles.
15:39 JoeJulian nigelb, misc: I agree. The whole peering process could be improved with regard to error messages.
15:39 gem joined #gluster
15:40 luizcpg joined #gluster
15:41 ic0n_ joined #gluster
15:50 ankitraj joined #gluster
15:55 kxseven joined #gluster
15:55 kpease_ joined #gluster
15:59 ivan_rossi seems that there are SSL problems on download.gluster.org
15:59 ivan_rossi curl https://download.gluster.org/pub/g​luster/glusterfs/3.8/3.8.5/rsa.pub
15:59 ivan_rossi curl: (35) error:140770FC:SSL routines:SSL23_GET_SERVER_HELLO:unknown protocol
15:59 ivan_rossi from firefox: An error occurred during a connection to download.gluster.org. SSL received a record that exceeded the maximum permissible length. Error code: SSL_ERROR_RX_RECORD_TOO_LONG
16:00 ivan_rossi confirmed both from italy and germany
16:02 cloph I guess listen 443 but forgot the "ssl" (listen 443 ssl)
16:08 ivan_rossi well someone reconfigured ssl today, since yesterday it was fine...
16:12 JoeJulian misc: ^
16:14 JoeJulian emailed -infra
16:14 ivan_rossi thx, joe.
16:15 BitByteNybble110 joined #gluster
16:15 prth joined #gluster
16:16 ivan_rossi guess that our plan to mirror the repo will get a boost. :-D
16:16 rafi joined #gluster
16:17 klaas joined #gluster
16:20 magrawal joined #gluster
16:21 rafi joined #gluster
16:39 ivan_rossi left #gluster
16:40 kkeithley unless someone has just changed it, the ssl.conf already has "Listen 443 https"
16:42 fyxim joined #gluster
16:42 devyani7 joined #gluster
16:42 Lee1092 joined #gluster
16:42 kkeithley do you think it should be "Listen 443 ssl" instead of "Listen 443 https" ?
16:43 telius joined #gluster
16:43 yawkat joined #gluster
16:43 satya4ever joined #gluster
16:44 Peppard joined #gluster
16:44 rwheeler joined #gluster
16:45 mmckeen joined #gluster
16:45 shyam joined #gluster
16:49 JoeJulian imho, it should be "listen 443 http2"
16:49 kkeithley it's apache, not nginx.
16:49 JoeJulian Well that's the other problem. ;)
16:49 kkeithley all the examples I see have https,
16:51 kkeithley but ssl.conf (where listen 443 https is) hasn't been touched in over a week
16:52 kkeithley download.gluster.org_ssl.conf was changed earlier today
16:53 kkeithley http://stackoverflow.com/questions/119336/s​sl-error-rx-record-too-long-and-apache-ssl says people getting this error have an incorrectly configured local proxy
16:53 glusterbot Title: configuration - ssl_error_rx_record_too_long and Apache SSL - Stack Overflow (at stackoverflow.com)
16:54 squizzi joined #gluster
16:55 JoeJulian $ curl http://download.gluster.org:443/
16:55 JoeJulian <html>
16:55 JoeJulian <title>Gluster.org Download Server</title>
16:55 glusterbot Title: Gluster.org Download Server (at download.gluster.org:443)
16:58 kkeithley ???
17:00 haietza joined #gluster
17:01 prth joined #gluster
17:01 haietza hi everyone. i'm trying to set up gluster geo-replication and i'm getting the gsync peer_gsec_create command not found error when i try to run gluster system:: execute gsec_create. can anyone help?
17:01 kkeithley try again. I used `curl https://download.gluster.org:443/`
17:01 kkeithley and got the whole page
17:03 kkeithley ivan_rossi: try now please.
17:07 kkeithley oh, he's gone
17:07 * cloph was talking nginx config btw, but that server is running apache, so sorry if that confused people
17:08 JoeJulian kkeithley: lgtm
17:09 hackman joined #gluster
17:10 kkeithley okay. I changed <VirtualHost *:443> to <VirtualHost _default_:443>  on the advice of http://stackoverflow.com/questions/119336/s​sl-error-rx-record-too-long-and-apache-ssl
17:10 glusterbot Title: configuration - ssl_error_rx_record_too_long and Apache SSL - Stack Overflow (at stackoverflow.com)
17:10 cloph same here, seems fixed.
17:10 kkeithley I'll let nigelb and misc know so they can fix their ansible/salt/whatever it is they're using
17:12 natarej_ joined #gluster
17:13 victori joined #gluster
17:15 hchiramm joined #gluster
17:17 * kkeithley goes off to figure out where to send the invoice
17:18 kkeithley lessee, off hours, overtime, travel premium, 15 minutes rounded up to the nearest four hour multiple
17:18 JoeJulian +1
17:18 JoeJulian Is it a holiday somewhere?
17:18 kkeithley It's Diwali in India
17:18 JoeJulian Well there you go. Holiday pay too.
17:19 kkeithley Well, Sunday is Diwali. Close enough
17:19 haietza hi everyone. i'm trying to set up gluster geo-replication and i'm getting the gsync peer_gsec_create command not found error when i try to run gluster system:: execute gsec_create. can anyone help?
17:20 JoeJulian "command not found" suggests that gsync isn't installed.
17:20 raghu joined #gluster
17:21 mchangir kkeithley, well the Diwali festivities lasts almost three weeks ... they start five days before new moon and go up to the full moon day
17:21 haietza @JoeJulian I'm following the instructions on https://gluster.readthedocs.io/en/latest/​Administrator%20Guide/Geo%20Replication/ and I didn't see anything about installing gsync...should that be intalled with gluster, or do i need to install it separately?
17:22 JoeJulian Which distro?
17:22 haietza centos 7
17:23 JoeJulian Looks like the storage sig breaks that out into a glusterfs-geo-replication package.
17:23 haietza ah...i'll check into that. thank you so much!
17:23 JoeJulian You're welcome.
17:27 misc kkeithley: that's quite weird, because reverting your change do make make the problem come back, I wonder if there was a issue with apache needing a restart and we didn't
17:28 misc (and I am at openstack summit and out for the evening, and in PTO next week, so unlikely to be able to fix anything soon, unless from the airport)
17:28 kkeithley ?? reverting made it come back? Or reverting it didn't come back?
17:28 misc also, wifi from the hotel is craptastic
17:28 misc kkeithley: didn't made the issue come back
17:28 misc kkeithley: like, curl now work here
17:28 misc (because ansible would revert your change anyway)
17:29 kkeithley yeah, okay, maybe it just needed a restart
17:29 kkeithley yup
17:30 luizcpg_ joined #gluster
17:34 dgandhi joined #gluster
17:47 johnmilton joined #gluster
18:03 prth joined #gluster
18:04 haomaiwang joined #gluster
18:30 panina joined #gluster
18:33 farhorizon joined #gluster
18:36 snehring joined #gluster
18:48 JoeJulian Ansible doesn't know to restart apache if you change the configuration? Sounds about as useful as chef.
18:58 d0nn1e joined #gluster
19:08 Slashman joined #gluster
19:12 shaunm joined #gluster
19:24 [diablo] joined #gluster
19:28 farhoriz_ joined #gluster
19:31 haomaiwang joined #gluster
19:54 DV joined #gluster
19:57 haomaiwang joined #gluster
20:27 irated JoeJulian: typically thats left up to the user
20:27 irated salt and chef have watch and reload flags iirc
20:30 unclemarc joined #gluster
20:41 farhorizon joined #gluster
20:48 haomaiwang joined #gluster
21:11 JoeJulian Salt does for sure. During the short time I used chef, it didn't seem to know how to do anything and every little thing had to be programmed in order for it to manage state. I hated it.
21:18 misc JoeJulian: I did look and it do restart apache, so I am puzzled
21:19 * JoeJulian shrugs
21:20 JoeJulian The reason I blamed the tool is that I have more faith in your ability than I have in ansible - mostly because I've never used ansible.
21:21 misc well, i am also upstream coder on ansible, so I can take blame anyway :)
21:21 JoeJulian hehe
21:23 misc Oct 25 05:25:43 download01 python: detected unhandled Python exception in '/tmp/ansible_7VptUW/ansible_module_cron.py'
21:23 misc ok so I think I really need to setup the monitoring ASAP
21:30 JoeJulian +1
21:31 ira joined #gluster
21:36 misc ok so that bug is fixed in 2.2, and crash ansible before it could restart httpd
21:36 misc and the bug is a specific case with either running on travis, or the salt bus plugin lack of tty
21:36 misc at least, the RCA is easy
21:37 misc given it is half of RCA of another issue
21:39 haomaiwang joined #gluster
21:45 JoeJulian Nice.
22:00 misc I might write a complete bug report in the airport tomorrow
22:03 farhoriz_ joined #gluster
22:33 arpu joined #gluster
22:42 haomaiwang joined #gluster
23:25 Klas joined #gluster

| Channels | #gluster index | Today | | Search | Google Search | Plain-Text | summary