Camelia, the Perl 6 bug

IRC log for #metacpan, 2013-08-01

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

All times shown according to UTC.

Time Nick Message
00:05 talexb Hmm .. bin/metacpan mapping --delete runs now but fails after 1:38, 'Timed out while waiting for socket to become ready for reading' ..
00:06 talexb However the VM is chugging away on something (/home/metacpan, apparently) so I'm going to leave it alone for a while and try again later.
00:43 shmuel joined #metacpan
01:24 klapperl joined #metacpan
02:16 mike joined #metacpan
02:25 oalders virtualbox eats up most of my available CPU even when the VM isn't really doing anything
02:26 oalders it's really annoying
02:28 talexb Agreed.
02:28 oalders i found this fix, but i don't know how to do this with debian http://www.robeesworld.com/blog/64/centos-5-​high-cpu-usage-when-running-under-virtualbox
02:28 dipsy urgh. long url. Try http://tinyurl.com/kjgmkfo
02:28 dipsy [ Robeesworld.com - Fixing CentOS high cpu usage when running as a virtual machine under VirtualBox ]
02:29 oalders talexb: how much CPU is it eating on your machine?
02:30 talexb About half of a quad core machine. :/
02:30 oalders heh
02:31 oalders it's really bad for me.  at times i can't even type
02:31 talexb Ugh.
02:32 talexb It looks like /home/metacpan/metacpan.org/app.psgi is getting ru over and over again ..
02:32 talexb Not sure why.
02:32 trs failure to spawn?
02:32 trs or rather, stay spawned.
02:32 talexb I'm watching a tail of ps axf, and it jeeps re-appearing with a new PID.
02:33 talexb May shut it down tonight and do my usual git pull tomorrow morning to see if it still occurs.
02:34 talexb All four CPUs are averaging 25% to 50% load. :/
02:35 talexb Doing a vagrant halt ..
02:36 talexb Now about 5-15% load on all four CPUs.
02:37 trs talexb: what is spawning app.psgi on the vm?  it sounds like it isn't staying spawned.
02:37 talexb I don't know enough to understand what's going on .. I assume that it was Starman.
02:38 talexb I'll have a look tomorrow. Later, all!
02:39 oalders talexb: until tomorrow!
02:43 talexb :) You bet!
02:57 oalders talexb: the solution appears to be disabling nested paging
02:57 oalders system => acceleration
02:58 oalders there's an hour of my life i'll never get back
03:00 shmuel_ joined #metacpan
03:01 shmuel joined #metacpan
03:05 * trs . o O ( it's pages all the way down... )
03:15 shmuel_ joined #metacpan
03:19 rwstauner talexb: /var/log/startserver/*.log
03:35 ether_ joined #metacpan
05:19 rwstauner common::sense 3.72 has been released and is actually installable
05:39 jayallen joined #metacpan
05:46 metacpan joined #metacpan
05:46 metacpan [cpan-api01] oalders created oalders/tidy (+2 new commits): http://git.io/HHna8g
05:46 metacpan cpan-api/oalders/tidy e950cee Olaf Alders: Adds tidyall config and pre-commit setup.
05:46 metacpan cpan-api/oalders/tidy 9a46def Olaf Alders: Tidy everything.
05:46 metacpan left #metacpan
05:46 dipsy [ Comparing e950ceeae23c^...9a46defacb96 · CPAN-API/cpan-api · GitHub ]
05:48 preflex_ joined #metacpan
06:26 Mike-PerlRecruiter_ joined #metacpan
07:53 [Sno] joined #metacpan
07:54 [Sno] riba sent me to here - you guys experiment with a ton of syntax highlighters … :)
07:54 [Sno] I'm looking for a syntax highlighter which can show perl code nicely for a talk
08:03 ranguard [Sno]: we are using http://alexgorbatchev.com/Synt​axHighlighter/manual/brushes/ - I know some stuff was tweaked and sent back up steam don't know if it got merged in
08:03 dipsy [ SyntaxHighlighter - Bundled Brushes ]
08:04 [Sno] ranguard: thank you very much
08:05 ranguard np :)
08:10 talexb joined #metacpan
08:34 daxim joined #metacpan
08:39 preflex joined #metacpan
09:21 talexb joined #metacpan
12:36 dsteinbrunner joined #metacpan
13:02 ilbot2 joined #metacpan
13:02 Topic for #metacpan is now Have you installed your MetaCPAN VM? https://github.com/CPAN-API/metacpan-developer | Chat logs available at http://irclog.perlgeek.de/metacpan/
13:34 talexb Amazing to think that common::sense needs to be at version 3.72 .. presumably, as Perl is updated, it needs to be updated as well.
13:37 * talexb gets organized, by writing a script to update all of his metacpan repos.
13:42 ranguard talexb: you looking for stuff to do :) ?
13:42 talexb ranguard, Short answer: yes. Long answer: Want to get myself (my dev environment) better organized first.
13:43 talexb I want to make it easier to do a vagrant halt, freshen all of my repos, and re-start, i the event that I find what I think is an error.
13:44 ranguard would generating some xml sitemaps be of interest (I need to add more details to the ticket if it is) ?
13:44 talexb I haven't written much Perl lately, so this gives me an opportunity.
13:44 talexb I could work on that ..
13:45 ranguard why would you need to halt to refresh the repos? - you can do /etc/init.d/<service> restart after having updated your repos
13:45 talexb Is there an issue for that? Do you have an idea for what you need?
13:45 talexb ranguard, Er .. that's a good point.
13:45 ranguard talexb: that would be amazing, https://github.com/CPAN-AP​I/metacpan-web/issues/894 (you'll see why I need to add some more details) :)
13:45 dipsy [ Create sitemap files for search engines · Issue #894 · CPAN-API/metacpan-web · GitHub ]
13:46 talexb I just worry about the system have some of its files changed - while it's running.
13:46 ranguard naa, would be fine :)
14:06 ranguard talexb: https://github.com/CPAN-AP​I/metacpan-web/issues/894 <- ticket updated
14:06 dipsy [ Create sitemap files for search engines · Issue #894 · CPAN-API/metacpan-web · GitHub ]
14:07 bowtie_ joined #metacpan
14:10 oalders ranguard: talexb had to deal with some vagrant config changes, so there was a restart needed
14:11 oalders talexb: installing https://metacpan.org/module/App::GitGot on your host may solve most of that problem
14:11 dipsy [ App::GitGot - A tool to make it easier to manage multiple git repositories. - metacpan.org ]
14:14 talexb oalders, Cool, thanks.
14:15 talexb ranguard, OK -- thanks for the background, that should be plenty for me to get started with.
14:15 * talexb just took a break and ended up helping a neighbour move new lawn furniture.
14:17 oalders heh
14:19 oalders talexb: try disabling nested paging under system => acceleration
14:19 oalders in VirtualBox
14:19 oalders that seemed to help me
14:20 talexb ool tip. Thanks, I'll try it.
14:20 oalders the other thing people do a lot is create a dummy VM and run that in parallel with the real VM they want to use
14:20 oalders apparently that keeps one VM from hogging the CPU. it's a totally bizarre hack
14:21 talexb Wow, that's odd.
14:21 oalders yep
14:22 oalders http://tech.shantanugoel.com/2009/07/07/vir​tualbox-high-cpu-usage-problem-solved.html
14:22 dipsy urgh. long url. Try http://tinyurl.com/7uljylb
14:22 dipsy [ VirtualBox High CPU Usage Problem Solved | Shantanu vs The World ]
14:24 talexb Wow. That is bizarre.
14:24 talexb ANd my machine (a desktop) seems much quieter this morning as I'm not running any VMs right now.
14:25 talexb Hmm .. I wonder if changing the number of CPUs from 2 to 1 would achieve the same goal.
14:28 talexb And one user reported that once he'd 'pinned' every virtual box to cpu_0, everything Just Worked.
14:33 oalders let me know if you find something that works.  my VM was a bit better last night, but still kind of insane
14:33 talexb Sure thing.
14:46 jayallen joined #metacpan
16:00 talexb I've forgotten how there can be additional distractions while working at home -- like door-to-door religion salesman.
16:13 oalders :)
16:35 thaljef joined #metacpan
17:27 talexb joined #metacpan
17:30 talexb Interesting -- load is down to about 102% with a single CPU.
17:32 mst I like those people
17:32 mst "what's your policy on multiple relationships?"
17:32 mst "they're evil"
17:32 mst "no sale then"
17:32 mst "... what?"
17:55 talexb Or, "Hey, would you like something to drink? I just made some fresh chicken blood .."
18:00 mst my way is especially fun with mormons, since of course they -used- to have multiple marriages and get very confused as to what to do when confronted with somebody who -dislikes- their having dropped that
18:03 talexb Heh. This guy's final pitch was about the afterlife .. did I believe in it, or did it sound like 'pie in the sky'.
18:06 talexb oalders, I have fiddled around with a few things -- nested paging setting didn't change things for me.
18:07 talexb I changed from 2 CPUs to 1, and got a reduction from 113% to 102%. But (fanfare) it's possible to change the executiion cap *while the VBox is running*, and that works pretty well.
18:07 talexb I set it to a 50% cap, and got 52% on top. Set it to 75% and got 77% on top.
18:08 talexb So one possible approach would be to let it go at 100% when starting up and running stuff, but throttle it way down when you would like to pause the system .. without actually suspending it or halting it.
18:10 oalders talexb: ah, great. i will try that next.  didn't know you could change that with the VM running
18:11 talexb Either did I .. but I'm a fan of trying stuff out.
18:11 talexb Although with reduced CPUs and/or execution caps, I notice that vagrant throws errors doing a halt .. some sort of race condition, perhaps.
18:15 rwstauner talexb: vagrant does randomly throw errors
18:16 rwstauner seems like a timeout or something waiting for VBoxManage
18:16 rwstauner happens on my machine fairly frequently
18:16 rwstauner at least frequently enough that i ignore it
18:16 talexb rwstauner, Agreed .. smells like a race condition to me.
18:16 rwstauner b/c usually the operation was successful anyway
18:16 rwstauner yeah, could be
18:16 talexb I also ignore it ..
18:16 rwstauner :-)
18:27 Mike-PerlRecruiter_ joined #metacpan
18:41 talexb I'm wondering why /usr/local/perlbrew/perls/p​erl-5.16.2/bin/start_server seems to be continuously starting new starman processes -- yet nothing in log explains why.
18:45 trs talexb: try running the app by hand to see how it fails?
18:45 oalders talexb: is this the api or www that is respawning?
18:45 talexb oalders, It's the API (port 5001).
18:46 talexb trs, Will do -- still trying VBox configurations to find one that produces the lightest load on the host system.
18:46 talexb Oh -- do I have that backwards?
18:46 oalders one thing we should do is reduce the default number of Starman workers on VM
18:47 oalders talexb: yes :)
18:47 talexb D'Oh. :/
18:47 oalders i had a similar issue yesterday and it was just because a module was missing
18:48 trs talexb: yeah, debug the out of control spawning before looking at load :)
18:48 oalders CPAN::Changes
18:50 ranguard the 'install_modules' script might be worth running from the vagrant provisioning script (e.g. so it runs at boot time) - not sure what changes would be needed - rwstauner ?
18:51 rwstauner ranguard: yeah, that's been mentioned a few times in the last couple days
18:52 rwstauner i added a few specific modules that are known to not be in the vm
18:52 rwstauner but i'm hesitant to run a full install b/c it upgrades a ton of modules needlessly and takes a long time
18:52 oalders i think carton would help there
18:52 rwstauner it would be less of an issue if we specified more versions in our prereqs
18:52 rwstauner or use carton, says oalders
18:52 ranguard oalders: good call
18:53 oalders carton + versions :)
18:53 rwstauner i haven't used or investigated it yet
18:53 rwstauner s/it/carton/
18:53 oalders it's working well for me with other stuff
18:53 rwstauner but i could add CPAN::Changes to the provisioner for the time being
18:53 oalders that would be a good stop-gap
18:53 ranguard that would be good - even if just to have it as an example
18:55 talexb This is interesting .. my last run put the VM into 'guru meditation', a mode I've never seen.
18:55 oalders heh
18:55 rwstauner VMs have modes?
18:55 talexb rwstauner, Well, normally, Stopped, Suspended or Running.
18:56 rwstauner according to what?  VirtualBox?
18:56 talexb I just asked about it on #vbox, but things are kinda quiet there right now ..
18:56 rwstauner i'd not heard of that before
18:56 talexb I think a screen shot is in order ..
18:58 * rwstauner pushes CPAN::Changes fix
18:58 Skunix joined #metacpan
18:58 talexb https://www.dropbox.com/lightbox/h​ome/Public?select=Screenshot%20fro​m%202013-08-01%2014%3A57%3A13.png
18:58 dipsy [ Dropbox - Sign in ]
18:59 talexb Hmm .. that's supposed to be public.
18:59 rwstauner :-)
18:59 rwstauner poor dipsy
18:59 oalders what purpose does dipsy actually serve? other than fetching page titles?
18:59 talexb https://dl.dropboxusercontent.com/u/6445093/Scr​eenshot%20from%202013-08-01%2014%3A57%3A13.png
19:00 oalders nice
19:03 rwstauner talexb: that's cool
19:04 rwstauner sometimes dipsy is fun
19:04 talexb I'm going to have to Google for this .. I can always power off my PC to recover, but I'd like to find a nicer way to disrupt this meditation.
19:05 talexb Go Dipsy!
19:05 rwstauner yeah, interesting
19:05 rwstauner dipsy++
19:05 dipsy rwstauner: Thanks!
19:05 rwstauner dipsy: do you do anything besides fetch page titles and try to regurgitate answers to weird questions?
19:05 dipsy rwstauner: Sorry, I don't know.
19:06 rwstauner i was afraid of that
19:06 oalders heh
19:07 oalders dipsy--
19:09 rwstauner just got my random vagrant/vboxmanage error of the day
19:09 rwstauner Stderr: VBoxManage: error: The object is not ready
19:10 rwstauner left #metacpan
19:10 rwstauner joined #metacpan
19:18 talexb OK, going off the air to reboot my machine. I have patience for meditation, but it's not going away and I need to move on.
19:20 rwstauner 😕
19:26 ranguard dipsy: metacpan?
19:26 dipsy ranguard: metacpan is the new search.cpan.org
19:26 ranguard dipsy: botsnack
19:26 dipsy ranguard: :)
19:26 ranguard dipsy: 5*50
19:26 dipsy ranguard: 250
19:40 rwstauner dipsy: oalders?
19:40 dipsy rwstauner: oalders is just my second identity ;)
19:40 * rwstauner thought dipsy might have said "oalders is the boss"
19:40 rwstauner b/c
19:41 ranguard dipsy: oalders is also the boss
19:41 rwstauner oalders is the boss
19:41 rwstauner :-)
19:41 dipsy ranguard: Okay.
19:41 ranguard oalders?
19:41 dipsy oalders is just my second identity ;) or the boss
19:41 rwstauner :-)
19:41 oalders :)
19:42 ranguard you need the 'also' to concat, or 'dipsy forget <foo>; <foo> is..'
19:42 ranguard karma oalders
19:42 dipsy oalders has karma of 68.
19:42 ranguard karma rwstauner
19:42 dipsy rwstauner has karma of 97.
19:42 ranguard karma ranguard
19:42 dipsy ranguard has karma of 190.
19:42 ranguard expain karma oalders
19:42 * rwstauner didn't know
19:43 ranguard explain rwstauner
19:43 dipsy positive: hacking on the api (oalders), thanks (mo), RemovePrereqs and skip Regex (ccushing); negative: no comments; overall: 97.
19:43 rwstauner :-)
19:43 ranguard dipsy keeps track of your ++'s :)
19:43 oalders dipsy++ # keeping track of ++
19:43 dipsy oalders: Thanks!
19:43 rwstauner yeah, i thought only purl did that
19:43 rwstauner ether: ^
19:44 ranguard oalders: so, a few purposes, not saying actually useful, but :)
19:44 oalders karma dipsy
19:44 dipsy dipsy has karma of 206.
19:44 ranguard dipsy: dipsy?
19:44 dipsy ranguard: dipsy is <req>so-named, I think, because \my antenna looks like a dipstick. or blech's bitch or confused about their gender or a bad influence or an apparently useful bot or publically accessible *ph34r* or going out with slavorg or a raving queen or one of the two male teletubbies (the
19:44 dipsy ..green one), the other being tinky winky or a generalised encyclopedia or fond of doodles or a crap bot or useless
19:45 ether aroo? 'sup?
19:45 rwstauner ether: just noticed that dipsy does indeed record karma
19:46 * ranguard goes back to trying to parse his credit card pdf file!
19:46 ether :)
19:46 rwstauner i recall you lamenting the absence of purl
19:46 * ether nodnods
19:46 ether dipsy: karma ether
19:46 dipsy ether: ether has karma of 61.
19:46 ether purl: karma ether
19:47 ether indeed
19:47 ether dipsy doesn't seem to be on as many channels
19:47 ranguard dipsy started as London.pm's bot
19:47 oalders ranguard: thanks for the bot lesson -- i stand corrected :)
19:48 ranguard http://london.pm.org/about/faq.html#dipsy
19:48 dipsy [ London.pm: FAQ for London.pm ]
19:49 oalders the link to dipsy's svn repo seems to be dead
19:53 talexb joined #metacpan
19:54 jayallen joined #metacpan
20:07 talexb Wow -- much better.
20:07 oalders no more meditation?
20:07 talexb Load is now 5-6%, and I'm not getting any constant re-spawning.
20:08 oalders what was the fix?
20:08 talexb oalders, The meditation was .. terminated, with extreme prejudice. :)
20:08 oalders :D
20:08 talexb Installing CPAN::Changes seems to have worked wonders.
20:09 talexb I'm running 1 CPU, AMD-V enabled, Nested Paging disabled ..
20:10 talexb and IO APIC disabled.
20:10 talexb That last is on the Motherboard tab; two flags are on the Acceleration tab.
20:11 oalders that sounds similar to my setup.  it's now idling at around 3%
20:11 oalders for the headless vm
20:11 talexb Yup.
20:11 talexb I'd like to add this information to the README if you don't mind ..
20:12 oalders please :)
20:13 * talexb can't imagine how much of a pain this would be with svn, or even (shudder) cvs. Yay git.
20:15 ranguard talexb: general policy is just do it (esp with documentation) - we can always revert :)
20:17 talexb ranguard, Great. :) I'm not the kind of person who casually barges in and starts messing with stuff on a new project, so my preference is to trad lightly. But I get your message.
20:17 talexb Er, 'tread lightly'.
20:21 oalders no treading lightly here
20:26 talexb EXCELLENT!
20:43 rwstauner \o/
20:44 talexb That's very cool to be running a complete MetaCPAN locally. Awesomeness.
20:44 rwstauner :-)
20:54 dsteinbrunner joined #metacpan
20:55 talexb ranguard, Having a look for WWW::Google::Sitemap that you mentioned in #894 .. but don't see it on metacpan.org.
20:57 oalders talexb: https://metacpan.org/module/Search::Sitemap might do it for you
20:57 dipsy [ Search::Sitemap - Perl extension for managing Search Engine Sitemaps - metacpan.org ]
20:57 oalders i *think* that should conform to this standard http://www.sitemaps.org/
20:57 dipsy [ sitemaps.org - Home ]
20:58 talexb oalders, Great, thanks -- just prefer not to cretae my own wheel when there are so many good ones out there.
20:58 oalders yeah, lots of wheels :)
21:30 dsteinbrunner joined #metacpan
21:35 tianon joined #metacpan
21:50 meraxes joined #metacpan
22:42 metacpan joined #metacpan
22:42 metacpan [metacpan-web01] oalders pushed 1 new commit to master: http://git.io/wjkmjg
22:42 metacpan metacpan-web/master 445a746 Olaf Alders: Bumps Encode from 2.10 to 2.51.
22:42 metacpan left #metacpan
22:42 dipsy [ Bumps Encode from 2.10 to 2.51. · 445a746 · CPAN-API/metacpan-web · GitHub ]
22:47 dolmen joined #metacpan
23:29 thaljef joined #metacpan
23:42 dsteinbrunner joined #metacpan

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