Perl 6 - the future is here, just unevenly distributed

IRC log for #gluster-dev, 2014-08-20

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

All times shown according to UTC.

Time Nick Message
00:59 topshare joined #gluster-dev
00:59 topshare_ joined #gluster-dev
01:02 vimal joined #gluster-dev
01:25 shyam joined #gluster-dev
02:00 Yuan joined #gluster-dev
02:01 Yuan_ joined #gluster-dev
03:15 _Bryan_ joined #gluster-dev
03:45 hagarth joined #gluster-dev
03:56 Yuan_ joined #gluster-dev
03:56 kdhananjay joined #gluster-dev
03:58 kanagaraj joined #gluster-dev
03:59 shubhendu joined #gluster-dev
04:03 itisravi joined #gluster-dev
04:07 Humble joined #gluster-dev
04:08 itisravi joined #gluster-dev
04:10 ppai joined #gluster-dev
04:11 hchiramm_ joined #gluster-dev
04:14 kanagaraj joined #gluster-dev
04:18 lalatenduM joined #gluster-dev
04:29 Rafi joined #gluster-dev
04:32 anoopcs joined #gluster-dev
04:32 anoopcs joined #gluster-dev
04:38 nishanth joined #gluster-dev
04:41 kshlm joined #gluster-dev
04:43 spandit joined #gluster-dev
04:43 atalur joined #gluster-dev
04:49 atinmu joined #gluster-dev
04:52 kanagaraj joined #gluster-dev
04:55 jiffin joined #gluster-dev
04:57 atinmu joined #gluster-dev
04:59 kanagaraj joined #gluster-dev
04:59 shubhendu_ joined #gluster-dev
05:13 hchiramm__ joined #gluster-dev
05:17 topshare joined #gluster-dev
05:25 hagarth joined #gluster-dev
05:29 kdhananjay joined #gluster-dev
05:32 topshare joined #gluster-dev
05:49 bala joined #gluster-dev
05:51 atinmu joined #gluster-dev
06:10 kanagaraj_ joined #gluster-dev
06:12 kanagaraj_ joined #gluster-dev
06:31 kanagaraj joined #gluster-dev
06:35 kanagaraj joined #gluster-dev
06:36 atinmu joined #gluster-dev
06:39 kanagaraj_ joined #gluster-dev
06:39 kanagaraj_ joined #gluster-dev
06:45 kanagaraj__ joined #gluster-dev
06:58 topshare joined #gluster-dev
06:59 topshare joined #gluster-dev
07:05 topshare joined #gluster-dev
07:13 kanagaraj joined #gluster-dev
07:34 itisravi_ joined #gluster-dev
07:35 hagarth joined #gluster-dev
07:43 topshare joined #gluster-dev
07:44 vimal joined #gluster-dev
07:50 raghu` joined #gluster-dev
07:53 lalatenduM joined #gluster-dev
07:54 hchiramm_ joined #gluster-dev
07:57 nishanth joined #gluster-dev
08:15 hagarth joined #gluster-dev
08:31 topshare_ joined #gluster-dev
08:40 kshlm joined #gluster-dev
08:48 ppai joined #gluster-dev
08:56 topshare joined #gluster-dev
09:08 shubhendu_ joined #gluster-dev
09:25 Humble joined #gluster-dev
09:33 Rafi_kc joined #gluster-dev
10:07 kanagaraj joined #gluster-dev
10:12 deepakcs joined #gluster-dev
10:15 kanagaraj joined #gluster-dev
10:16 spandit joined #gluster-dev
10:19 kdhananjay joined #gluster-dev
10:34 hagarth joined #gluster-dev
10:40 kdhananjay joined #gluster-dev
10:45 shyam joined #gluster-dev
10:54 shubhendu_ joined #gluster-dev
11:35 shubhendu_ joined #gluster-dev
11:42 kshlm joined #gluster-dev
11:44 ppai joined #gluster-dev
11:53 JustinClift hagarth: I'm here
11:53 JustinClift PM if you want?
11:53 * JustinClift just woke up
11:58 ira joined #gluster-dev
11:59 kkeithley gluster community meeting in 2 (two) minutes in #gluster-meeting@freenode
12:03 skoduri joined #gluster-dev
12:07 pranithk joined #gluster-dev
12:13 kshlm joined #gluster-dev
12:17 kanagaraj joined #gluster-dev
12:20 edward1 joined #gluster-dev
12:21 atinmu joined #gluster-dev
12:23 pranithk kkeithley++
12:23 glusterbot pranithk: kkeithley's karma is now 16
12:27 itisravi joined #gluster-dev
12:44 kanagaraj joined #gluster-dev
12:46 kshlm joined #gluster-dev
12:46 kshlm joined #gluster-dev
12:55 rastar joined #gluster-dev
13:00 lalatenduM jcastillo, ping, regarding http://review.gluster.org/#/c/7199
13:01 jcastillo lalatenduM: pong, tell me
13:01 shyam joined #gluster-dev
13:02 lalatenduM jcastillo, does that patch fixes the review comments by ndevos ?
13:03 jcastillo lalatenduM: no, I judged better to leave these for you. It does fix a misplaced %preu though from your patchset 5
13:03 lalatenduM jcastillo, ok, actually I am new to auto tools and hence taking time to understand these things:(
13:04 jcastillo lalatenduM: ack, take your time and let me know if I can give you a hand :)
13:05 lalatenduM jcastillo, are you are familiar with autotools?
13:08 jcastillo lalatenduM: a bit
13:08 lalatenduM jcastillo, cool:)
13:10 hchiramm_ JustinClift++
13:10 glusterbot hchiramm_: JustinClift's karma is now 9
13:25 topshare joined #gluster-dev
13:42 tdasilva joined #gluster-dev
13:44 lalatenduM ndevos, regarding bug triage , lets the bug triage mail first then the bug triage meeting plan mail
13:45 ndevos lalatenduM: thats fine with me
13:46 lalatenduM ndevos, ok, you should get the draft from me after some time, once reviewed I can send it
13:56 kanagaraj joined #gluster-dev
14:21 wushudoin joined #gluster-dev
14:32 cristov joined #gluster-dev
14:32 topshare joined #gluster-dev
14:45 JustinClift Is anyone else having trouble logging in to the Gluster wiki?
14:45 JustinClift It's telling me "GlusterDocumentation uses cookies to log in users. You have cookies disabled. Please enable them and try again."
14:45 JustinClift I really don't have cookies disabled
14:46 ndevos I had that once too, but I've managed to log after that - maybe I deleted the cookie?
14:46 JustinClift I'll give that a shot
14:47 JustinClift Nope
14:47 kshlm JustinClift, same here.
14:47 JustinClift misc: Any ideas? ^
14:49 * JustinClift wonders if our Wiki install has something wrong with it
14:49 JustinClift k, looks like the last change to the wiki was yesterday
14:50 JustinClift So this is a recent thing
14:50 JustinClift Maybe the varnish changes did something?
14:50 JustinClift semiosis: ^ ?
14:50 JustinClift kshlm: Thanks for checking. :)
14:51 kshlm YW. I checked on Chromium and Firefox. Both are failing.
14:51 JustinClift Same here on FF and Opera
14:52 JustinClift semiosis made some changes to the Varnish caching settings yesterday, to stop some script kiddie from dos-ing our site
14:52 JustinClift Maybe something for that did things (no idea)
14:52 kshlm Maybe.
14:52 JustinClift Oh well, I can put the website build instructions somewhere else for now, and transfer them to the wiki later
14:52 kshlm Gotta leave now. Catch you later.
14:52 JustinClift kshlm: :)
14:59 misc JustinClift: let me look
15:00 misc JustinClift: yeah
15:00 jobewan joined #gluster-dev
15:00 misc JustinClift: try again ?
15:03 JustinClift misc: That worked.  Logged in now. :)
15:03 JustinClift misc: Thanks. :)
15:06 misc JustinClift: indeed, there was configuration saying remove cookie"
15:06 misc I did remove it, but maybe semiosis had a reason :/
15:17 hchiramm__ joined #gluster-dev
15:22 semiosis comment in the vcl explained why
15:22 semiosis varnish keeps a different copy for each cookie value
15:23 semiosis so standard practice is to strip cookies
15:23 semiosis except of course where cookies are needed, but then those pages shouldn't be cached at all
15:23 semiosis i'll go back in & add an exclusion (just like i did for the WP blog) so that wiki bypasses the cace
15:23 semiosis cache*
15:29 semiosis done.
15:30 semiosis kkeithley: ping re deb in git
15:34 semiosis kkeithley: normally the deb source package is maintained in a separate repo from the actual project, except for debian native packages, which glusterfs is not
15:36 semiosis problem I see keeping the deb source in the project source is that it will be difficult to maintain.  changes to the deb packaging happen between glusterfs releases, so whatever is in the source tree will most often be outdated
15:37 misc semiosis: but varnish do clean the cache after a time ?
15:37 semiosis kkeithley: the debian project (through patrick) maintains their own SVN repo with the source package for glusterfs.  so i figure we ought to maintain our own for our deb releases
15:38 semiosis misc: cached objects expire after one hour (per my config).  but what are you really asking?
15:39 misc semiosis: just if my comment would have caused a cdisaster or not :)
15:39 hchiramm__ joined #gluster-dev
15:39 misc ( ie, removing the cookie striping )
15:39 semiosis kkeithley: furthermore, if people want to build debs, they'll never look in extras/foo/bar/whatever -- that's totally arcane!  they'll just go to the apt repo and download the source package, that's the normal way of things
15:39 kkeithley semiosis: yes, but the same is true for RPMs.
15:40 misc I usually do not trust upstream spec/deb, since upstream are usually not aware of the subtilities of the packaging policy
15:40 semiosis misc: no worries.  when i fixed the config just now i restarted varnish, which should purge the cache.
15:40 kkeithley well, it was a thought.
15:41 semiosis misc: i try hard to respect debian/ubuntu policy in our official debs, but you're right it's not universal
15:42 misc semiosis: I also try to avoid telling to people "take random packages from the web" as a rule
15:42 misc ( even if I realize that this is a bit antagonist to developpers who just try to help )
15:43 kkeithley when I develop I want to build, install, test, remove/clean.  The more I can do that in-tree, the happier I am.  I suppose I could use git sub-repo magic to pull in the bits from, e.g., your apt repo and build.
15:43 semiosis hmm interesting
15:43 kkeithley But since I'm not doing devel on Debian, it's a bit academic for me.
15:43 lalatenduM ndevos, JustinClift hchiramm_ kkeithley I have sent you guys a mail to review , before I send it to gluster-users/devel
15:45 kkeithley I just want to simplify and automate building new .dpkgs for Debian when we do a release
15:46 ndevos lalatenduM: yay, I've got mail!
15:46 kkeithley We have Fedora's Koji for the RPMs. Since Debian and Ubuntu work at arms-length we don't get that kind of synergy for those platforms
15:48 semiosis kkeithley: i suppose we could make a jenkins job to do deb stuff.  would take some work to get the whole process totally automated though :/
15:48 semiosis seems worth it, just hard to find the time
15:49 kkeithley semiosis: not to worry, I'm not expecting you to do it.
15:49 kkeithley unless you want to
15:50 kkeithley and when you have time
15:50 semiosis :)
15:54 lalatenduM ndevos, pm
15:54 ndevos lalatenduM: sure
15:56 kkeithley misc: going the other way, project devels often aren't  aware of the subtleties of the distro packaging policies. E.g. Fedora not using rsyslog.
15:57 hagarth joined #gluster-dev
15:58 kkeithley Speaking from experience, I can't keep track of everything Fedora is doing. Someone files a BZ very naively stating "gluster uses rsyslog".   I don't immediately grok that as a problem. Much later someone amends the BZ with a link to FESCO policy that packages should stop using rsyslog.
16:03 misc kkeithley: yep, integration is almost a full time job
16:03 misc but that's right that the current process is quite broken
16:04 misc especially since both side move at the same time
16:05 misc and packaging serve mostly as a glue between moving systems ( ie, distro ) and 1 single moving system ( upstream )
16:05 lalatenduM ndevos++
16:06 glusterbot lalatenduM: ndevos's karma is now 19
16:06 semiosis despite packaging being open & documented, seems like 99% of linux users never build a package.  so if they can't find one, they just can't use the software
16:08 misc yep
16:17 ndevos misc, semiosis: iptables has a connlimit module, that should be able to prevent many connections at the same time from one IP
16:18 semiosis sure, there's also ModSecurity
16:19 misc well, this wokr until you put all gluster people in the same conference :)
16:19 semiosis but i dont think we're at risk of DoS
16:19 semiosis and besides, how much bandwidth could we handle anyway
16:20 ndevos I have no idea how many connections were used at the same time, or if that really was the issue - just throwing an idea over the fence :)
16:20 misc semiosis: let me see if I can rent a botnet to see how far we can go :p
16:20 semiosis not many
16:21 ndevos okay
16:21 semiosis knock yourself out: https://github.com/newsapps/beeswithmachineguns
16:23 semiosis someone was brute forcing wp logins, which exhausted the apache workers (not too many of those with the default config).  the varnish config i made will prevent that specific attack.
16:24 semiosis next step would be to tune apache with more workers
16:24 semiosis however there's plenty of other attacks that could be done.  do we keep our wp & mediawiki software updated?  that's a good start
16:24 semiosis then we could add a WAF like Mod Security, which of course would also need to be kept up to date
16:25 semiosis maybe also an IDPS like snort... it never ends
16:25 misc so, bw between supercolony and my server in france
16:25 misc [  3]  0.0-10.0 sec   202 MBytes   169 Mbits/sec
16:25 semiosis sweet
16:26 misc keeping mediawiki and wp updated, yeah...
16:26 semiosis so if someone wanted to knock us offline they wouldn't need a very large botnet
16:26 misc having selinux turned one too
16:26 * misc didn't check firewall
16:27 misc I knew I shouldn't have checked
17:21 kshlm joined #gluster-dev
18:08 ira joined #gluster-dev
18:33 jobewan joined #gluster-dev
18:55 _Bryan_ joined #gluster-dev
19:03 hchiramm__ joined #gluster-dev
19:43 JustinClift semiosis misc: Looks like potentially another cookie problem.  This time with mailman though.  I auth to the admin interface ok, but when I click the submit button to approve a queued message I'm returned to the auth page.
19:43 semiosis yep
19:43 semiosis url?
19:46 semiosis JustinClift: the archives should get cached, those are in /pipermail right?  the mailman admin stuff is in /mailman?
19:47 JustinClift semiosis: Example URL: http://supercolony.gluster.org/mailman/admindb/gluster-devel
19:47 semiosis ok
19:48 semiosis should be fixed now
19:54 JustinClift semiosis: Yep, working now.  Thanks. :)
19:54 semiosis yw
19:54 semiosis anything else you can think of that should bypass the cache?  i.e. dynamic stuff?
19:57 JustinClift I'm reporting it as I hit it.
19:58 semiosis ok, fyi, i'll be traveling tomorrow through sunday so if you need something best way is to email me directly
19:58 JustinClift Cool. :)
19:58 semiosis or twitter
20:23 ndevos a little off-topic, but does anyone know how to mark a gnome3 notification as 'seen' with a keyboard shortcut, instead of clicking the (X)?
20:31 ndevos ah, with 'delete', but it also closes the notofication tray :-/
20:53 systemonkey joined #gluster-dev
22:59 shyam joined #gluster-dev
23:40 portante joined #gluster-dev
23:46 MacWinner joined #gluster-dev
23:47 ndevos joined #gluster-dev
23:47 ndevos joined #gluster-dev
23:52 wushudoin joined #gluster-dev

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