Perl 6 - the future is here, just unevenly distributed

IRC log for #gluster-dev, 2016-02-22

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

All times shown according to UTC.

Time Nick Message
00:01 owlbot joined #gluster-dev
00:05 EinstCra_ joined #gluster-dev
00:05 owlbot joined #gluster-dev
00:09 owlbot joined #gluster-dev
00:13 owlbot joined #gluster-dev
00:17 owlbot joined #gluster-dev
00:21 owlbot joined #gluster-dev
00:25 owlbot joined #gluster-dev
00:29 owlbot joined #gluster-dev
00:33 owlbot joined #gluster-dev
00:37 owlbot joined #gluster-dev
00:41 owlbot joined #gluster-dev
00:45 owlbot joined #gluster-dev
00:49 owlbot joined #gluster-dev
00:53 owlbot joined #gluster-dev
00:54 owlbot joined #gluster-dev
00:58 owlbot joined #gluster-dev
01:02 owlbot joined #gluster-dev
01:40 nbalacha joined #gluster-dev
01:47 nishanth joined #gluster-dev
01:48 EinstCrazy joined #gluster-dev
01:49 baojg joined #gluster-dev
02:14 pppp joined #gluster-dev
02:21 rafi joined #gluster-dev
02:48 ilbot3 joined #gluster-dev
02:48 Topic for #gluster-dev is now Gluster Development Channel - http://gluster.org | For general chat go to #gluster | Patches - http://review.gluster.org/ | Channel Logs - https://botbot.me/freenode/gluster-dev/ & http://irclog.perlgeek.de/gluster-dev/
02:52 ashiq joined #gluster-dev
03:24 sakshi joined #gluster-dev
03:28 josferna joined #gluster-dev
03:31 baojg joined #gluster-dev
03:36 overclk joined #gluster-dev
03:39 nbalacha joined #gluster-dev
03:40 josferna joined #gluster-dev
03:53 atinm joined #gluster-dev
03:58 shubhendu joined #gluster-dev
04:01 mchangir joined #gluster-dev
04:05 kanagaraj joined #gluster-dev
04:07 itisravi joined #gluster-dev
04:08 itisravi joined #gluster-dev
04:08 ppai joined #gluster-dev
04:29 karthikfff joined #gluster-dev
04:34 Manikandan joined #gluster-dev
04:35 aspandey joined #gluster-dev
04:52 hgowtham joined #gluster-dev
04:54 ndarshan joined #gluster-dev
04:56 jiffin joined #gluster-dev
05:00 PotatoGim joined #gluster-dev
05:10 Bhaskarakiran joined #gluster-dev
05:16 gem joined #gluster-dev
05:20 pppp joined #gluster-dev
05:23 EinstCra_ joined #gluster-dev
05:30 poornimag joined #gluster-dev
05:32 Saravanakmr joined #gluster-dev
05:33 inevity joined #gluster-dev
05:37 ggarg joined #gluster-dev
05:47 pranithk joined #gluster-dev
05:57 nishanth joined #gluster-dev
05:59 atalur joined #gluster-dev
06:06 skoduri joined #gluster-dev
06:06 josferna joined #gluster-dev
06:06 kdhananjay joined #gluster-dev
06:15 ppai joined #gluster-dev
06:17 rraja joined #gluster-dev
06:18 atinm joined #gluster-dev
06:21 Bhaskarakiran joined #gluster-dev
06:21 Bhaskarakiran joined #gluster-dev
06:29 kdhananjay joined #gluster-dev
06:36 aravindavk joined #gluster-dev
06:38 EinstCrazy joined #gluster-dev
06:58 atinm joined #gluster-dev
07:03 josferna rafi, ping
07:03 glusterbot josferna: Please don't naked ping. http://blogs.gnome.org/markmc/2014/02/20/naked-pings/
07:03 josferna rafi, need help ping ;)
07:04 baojg joined #gluster-dev
07:08 rafi josferna: pong
07:09 josferna rafi, what are the implication of removing fixlayout from attach tier
07:09 josferna rafi, just wanted to know what was discussed before I take this up
07:11 josferna rafi, known risks
07:13 rafi josferna: there will be performance drop for first lookup on a directory if it has anomalies with ENOENT OR ESTALE
07:14 josferna rafi, ok apart from the performance impact do we have any other implication ?
07:15 rafi josferna: we had one more race condition
07:15 rafi josferna: let me check the bug id
07:17 rafi josferna: https://bugzilla.redhat.com/show_bug.cgi?id=1286208
07:17 glusterbot Bug 1286208: high, urgent, ---, rkavunga, ASSIGNED , DHT + NFS mount : trusted.glusterfs.dht xattr was not created for directory on any sub-volume and file creation inside that directory failed with error 'No such file or directory'
07:17 josferna ok
07:19 josferna rafi, so this is the patch that will fix the layout if not done
07:20 josferna rafi, correct ?
07:20 rafi josferna: this is nothing to do with fix-layout, but it is good to fix before we remove the fix-layout
07:20 josferna rafi, "
07:21 josferna if the directory is present on all the subvolume
07:21 josferna and the layout is not proper for some reason, we
07:21 josferna will try to heal the directory."
07:21 josferna rafi, this is what the commit message says
07:21 rafi josferna: which bug ?
07:21 josferna rafi, and you have given a -1 for ur own patch
07:21 rafi josferna: pm
07:21 josferna rafi, http://review.gluster.org/#/c/12414/
07:22 josferna rafi, pm??
07:22 rafi josferna: it is not tested well, also need to clean the code,
07:23 rafi josferna: that's why I put a -1, as WIP
07:23 asengupt joined #gluster-dev
07:24 josferna rafi, so the fix for autoheal on io is not in and if I want to remove fixlayout from attach tier I need this fix, correct ?
07:25 rafi josferna: as I said earlier , this is nothing to do with fix-layout, but it is good to fix before we remove the fix-layout
07:27 josferna rafi, I understand this is nothing to do with fixyout as whole but the layout needs to be set on directory (which is fixlayout grossly) and you are trying to do it with  IO or lookup
07:27 josferna rafi, if we remove fixlayout from attach tier we will need this heal and it will cost us performance
07:27 josferna rafi, Correct ?
07:38 overclk joined #gluster-dev
07:42 Humble joined #gluster-dev
07:48 kdhananjay joined #gluster-dev
07:56 aravindavk joined #gluster-dev
08:01 post-factum hagarth: regarding 'performance/write-behind: retry "failed syncs to backend"' commit
08:02 post-factum hagarth: you've pointed me to it as to the posible reason of https://bugzilla.redhat.com/show_bug.cgi?id=1309462
08:02 glusterbot Bug 1309462: low, unspecified, ---, bugs, NEW , Upgrade from 3.7.6 to 3.7.8 causes massive drop in write performance.  Fresh install of 3.7.8 also has low write performance
08:03 post-factum hagarth: but i do not see that commit between 3.7.6 and 3.7.8
08:03 post-factum hagarth: probably, i'm missing something
08:06 post-factum hagarth: it seems i'm missing commits in my git clone. wtf
08:06 post-factum hagarth: ok, will try to revert that one and test
08:15 owlbot joined #gluster-dev
08:19 post-factum i'm definitely lost in gluster git mirror
08:19 post-factum "git log v3.7.6..v3.7.8 | grep write-behind" shows nothing
08:19 owlbot joined #gluster-dev
08:23 owlbot joined #gluster-dev
08:32 itisravi joined #gluster-dev
08:41 ira joined #gluster-dev
08:59 Bhaskarakiran joined #gluster-dev
09:05 nbalacha joined #gluster-dev
09:35 owlbot joined #gluster-dev
09:39 post-factum ok, here is update on https://bugzilla.redhat.com/show_bug.cgi?id=1309462
09:39 glusterbot Bug 1309462: low, unspecified, ---, bugs, NEW , Upgrade from 3.7.6 to 3.7.8 causes massive drop in write performance.  Fresh install of 3.7.8 also has low write performance
09:39 post-factum i see slow writes, when the following conditions are met:
09:39 post-factum 1) volume is replica (replica 2 for me)
09:39 post-factum 2) storage.linux-aio is set to on
09:39 owlbot joined #gluster-dev
09:40 post-factum 3) performance.write-behind is set to on
09:40 post-factum slow writes -- ~2 MB/s
09:41 post-factum disabling storage.linux-aio makes 25 MB/s
09:41 post-factum disabling performance.write-behind with storage.linux-aio makes 35 MB/s
09:43 post-factum enabling performance.write-behind with storage.linux-aio disabled makes 50 MB/s
09:43 post-factum so I guess it is all about storage.linux-aio and not performance.write-behind
09:43 owlbot joined #gluster-dev
09:44 ndevos post-factum: and a fresh install of 3.7.6 gives different numbers?
09:44 post-factum ndevos: going to downgrade now, give me a minute plz
09:45 ndevos post-factum: sure, I'm only asking, no idea if I can really help with it
09:47 post-factum ndevos: yes, downgrading client to 3.7.6 fixes the issue as well
09:47 ndevos post-factum: oh, interesting!
09:48 ndevos post-factum: can you make sure to add these notes to the bug report?
09:48 post-factum i will recreate volume from scratch now focusing just on those 2 options, and will definitely post an update to bugreport
09:57 ggarg joined #gluster-dev
10:01 owlbot joined #gluster-dev
10:01 post-factum updated bugreport: https://bugzilla.redhat.com/show_bug.cgi?id=1309462#c3
10:01 glusterbot Bug 1309462: low, unspecified, ---, bugs, NEW , Upgrade from 3.7.6 to 3.7.8 causes massive drop in write performance.  Fresh install of 3.7.8 also has low write performance
10:01 post-factum i blame AIO :)
10:03 ira joined #gluster-dev
10:06 poornimag joined #gluster-dev
10:06 nbalacha joined #gluster-dev
10:16 atinm sakshi, there?
10:17 sakshi atinm, hey
10:17 atinm sakshi, this is regarding http://review.gluster.org/#/c/10954
10:18 sakshi atinm, yes
10:18 atinm sakshi, I see that brick_up_status () is not included in the backport, so just cross checking that its already there in 3.7
10:18 atinm sakshi, I remember a discussion where you were supposed to take care of this discrepancy, is that patch already merged?
10:18 sakshi atinm, let me check 1 second
10:19 atinm sakshi, ok
10:19 sakshi atinm, nope, this is the patch http://review.gluster.org/#/c/13306/
10:20 sakshi atinm, it has just passed regression
10:21 sakshi atinm, this is just a backport, after this gets merged if there is any redundant function I will send a patch to remove that
10:23 atinm sakshi,http://review.gluster.org/#/c/13306/3/tests/bugs/glusterd/bug-1225716-brick-online-validation-remove-brick.t uses brick_up_status ()
10:23 atinm sakshi, my question is which patch backported the same function in 3.7
10:25 atinm sakshi, http://review.gluster.org/13276 has done that
10:25 atinm sakshi, I'm merging the patch now
10:26 sakshi atinm, oh okay..
10:28 Bhaskarakiran joined #gluster-dev
10:49 ndevos post-factum++ many thanks for the bug update!
10:49 glusterbot ndevos: post-factum's karma is now 1
10:54 josferna rafi++
10:54 glusterbot josferna: rafi's karma is now 41
10:54 itisravi joined #gluster-dev
11:00 aravindavk joined #gluster-dev
11:11 ppai joined #gluster-dev
11:13 poornimag joined #gluster-dev
11:26 rafi oing dlambrig
11:26 rafi ping dlambrig
11:45 sankarshan_away joined #gluster-dev
12:00 owlbot joined #gluster-dev
12:00 Gaurav_ joined #gluster-dev
12:02 owlbot joined #gluster-dev
12:03 hagarth joined #gluster-dev
12:16 kdhananjay joined #gluster-dev
12:18 pranithk joined #gluster-dev
12:18 ppai joined #gluster-dev
12:32 josferna joined #gluster-dev
12:42 nbalacha joined #gluster-dev
12:55 lpabon joined #gluster-dev
12:57 josferna joined #gluster-dev
13:12 poornimag joined #gluster-dev
13:17 ira joined #gluster-dev
13:35 aravindavk joined #gluster-dev
13:54 kkeithley dlambrig, josferna, nbalacha:  nightly static analysis  cppcheck is reporting several new warnings in libglusterfs/src/gfdb/gfdb_sqlite3.c.   http://download.gluster.org/pub/gluster/glusterfs/static-analysis/master/glusterfs-cppcheck/2016-02-18-a47d1bf9/cppcheck.txt
13:55 kkeithley that's on the master branch. similar warnings on the release-3.7 branch  http://download.gluster.org/pub/gluster/glusterfs/static-analysis/release-3.7/glusterfs-cppcheck/
13:57 josferna kkeithley, i will look into it
13:57 kkeithley thanks
13:59 kkeithley do you want me to file a bz?
13:59 nbalacha kkeithley, doesn't look like there were any changes to libglusterfs/src/gfdb/gfdb_sqlite3.c recently
13:59 kkeithley yeah, maybe they're not so new after all.
14:00 kkeithley might even be false positive
14:00 kkeithley lol. I saw 201x-02-18 and thought it was yesterday.  Actually it's 2015-02-18.  Missed the year. ;-)
14:02 nbalacha kkeithley, the link you sent is for 2016
14:02 nbalacha kkeithley, but the errors do show up in the earlier runs as well
14:02 kkeithley the cppcheck report is yesterday.  git blame says the last changes were 2015-02-18
14:03 kkeithley logical yesterday (Friday)
14:12 nbalacha kkeithley, ok
14:17 kdhananjay joined #gluster-dev
14:24 skoduri joined #gluster-dev
14:38 Gaurav_ joined #gluster-dev
14:44 josferna joined #gluster-dev
15:22 hchiramm joined #gluster-dev
15:31 atalur joined #gluster-dev
15:35 wushudoin joined #gluster-dev
15:56 josferna joined #gluster-dev
16:08 pranithk joined #gluster-dev
16:33 jiffin joined #gluster-dev
16:35 kanagaraj joined #gluster-dev
16:35 rafi joined #gluster-dev
16:41 shubhendu joined #gluster-dev
17:25 pranithk joined #gluster-dev
18:02 hchiramm joined #gluster-dev
18:15 Manikandan joined #gluster-dev
18:42 nishanth joined #gluster-dev
19:11 jobewan joined #gluster-dev
19:51 lpabon joined #gluster-dev
20:19 kkeithley gah, look at all the unchecked calls to snprintf. And calls to snprintf that don't check for overflow/truncation. :-/
20:20 _nixpanic joined #gluster-dev
20:20 xavih joined #gluster-dev
20:20 jtc joined #gluster-dev
20:20 csaba joined #gluster-dev
20:20 _nixpanic joined #gluster-dev
20:21 lkoranda joined #gluster-dev
20:25 samikshan joined #gluster-dev
21:20 anoopcs joined #gluster-dev

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