IRC Logs for #crux-devel Monday, 2015-01-19

*** Roosevelt has quit IRC02:37
*** Roosevelt has joined #crux-devel02:40
*** mavrick61 has quit IRC03:46
*** mavrick61 has joined #crux-devel03:47
*** Feksclaus has quit IRC04:45
diversejaeger: anything new lately?06:30
Romsteris it fair to say Moritz Wilhelmy isn't doing anything in contrib anymore?;a=history;f=usb_modeswitch;hb=HEAD08:03
Romsterbeen in there since 2010 no changes08:03
frinnstsend him an email?09:00
Romsteri haven't bothered...12:02
*** jue has quit IRC13:16
jaegerdiverse: not since the hashmap chat a few days back13:26
*** jue has joined #crux-devel13:33
*** Feksclaus has joined #crux-devel17:10
*** leo-unglaub has joined #crux-devel20:44
*** leo-unglaub has quit IRC21:18
frinnstteK_: can you unlist my ports from the portdb? or perhaps tell me where to do it myself22:16
frinnstall ports i have listed are of shitty quality and 99.9% just for my personal use22:17
frinnstFINE, LEAVE THEN22:17
jaegeryou scared him22:19
frinnstfeels like I should get my head around how everything works on I might logon once per year or so22:21
frinnstalso, anybody know why sh seems to segfault a lot? check dmesg22:23
jaegerhrmm, odd. no idea off the top of my head22:25
frinnstGNU bash, version 3.2.48(1)-release (i686-pc-linux-gnu)22:25
frinnstthat version is vulnerable to shellshock, no?22:26
frinnstnah, seems patched22:29
frinnstactually no22:29
frinnstno, its vulnerable22:30
frinnst"oops"22:30 is running a pretty old install currently22:30
diversejaeger: how old?22:36
*** teK_ has joined #crux-devel22:36
prologicoops :)22:37
prologicdo we have for hosting in any way?22:37
prologicor is it dondated?22:37
prologicI shoudln't talk, several of my systems are/were 2.722:37
diverseprologic: just be sure to upgrade :)22:38
diverseare you planning to give additional hosting?22:38
prologicyou mean keep my systems up-to-date this time aroudn? :)22:38
teK_frinnst: done; what's up?22:38
prologicno reason I can't dondate hosting for crux22:39
frinnstproblem is (as always) lack of control/kvm22:39
jaegerit's a 32-bit machine donated by mavrick61 running 2.8 I believe22:40
frinnstteK_: I was wondering how to remove a repo from the portdb22:40
jaegerhosting and hardware donated by mavrick6122:40
frinnstnah dude, 2.422:40
diverseah, not as old as I thought22:40
jaegeris it? I thought teK_ had updated a bit more :D22:40
diverseoh really?22:40
jaegerheh, ok22:40
frinnstanyways, thats what /usr/bin/crux says :)22:41
frinnstits been patched a bit since22:41
prologicjeager: specs?22:41
prologicjust wondering if it can be matched by or vultr.com22:41
jaegerI asked teK_ about possibly upgrading to the 3.1 unofficial x86 ISO but I think that'll have to wait a while22:41
jaegerI don't recall, we set it up in 200522:41
prologicwhat does free -m and df -h say?22:42
prologicthey're the most important :)22:42
frinnstdisk is not a problem22:42
frinnst1 gigabytezzz of ram22:42
diversejaeger: 10 years ago :o22:42
jaegerIt's worked quite well for as old as it is, and the price is right22:43
frinnsttek: pkgadd -u /home/rehabdoll/bash/bash#3.2.57-1.pkg.tar.gz ?22:44
frinnstbash has segfaulted quite a bit22:44
prologicprice is right :)22:45
prologic$0/month :)22:45
prologiccan't really beat that really22:45
frinnstI can probably get physical access to the server if we want to do something in the future22:46
frinnstonly live 140km away22:46
prologiconly? :) lol22:47
prologicI guess it's only a few hours drive22:47
prologicnice thing about and is that they give you full console access to the VM22:47
prologicbut the price isn't $0 :)22:48
frinnstits not that bad.. ~90 minutes. Summer vacation project :)22:48
teK_frinnst: see ~crux/bin/pdbcacher.php for details22:49
teK_the only thing I'm a bit upset about is the server's kernel22:49
teK_the rest of it is pretty patched up22:50
teK_and don't get me started with PHP :>22:50
frinnstbash is pretty outdated. vulnerable to shellshock22:50
frinnstI have a patched port in my homedir if you're feeling up for it22:51
teK_we use bash to process web requests?22:51
jaegernever hurts to be careful22:51
teK_in both ways22:52
frinnstI just noticed a shitload of sh segfaults when checking dmesg22:52
teK_saw them, too22:52
frinnstbtw you commented out my ports just now?22:53
teK_yes. It will take until tomorrow to make an effect22:53
frinnsta dl320 gen8 isnt that expensive22:54
teK_what for22:54
frinnst2 disks only22:55
frinnstIf we want to upgrade :)22:55
jaegerI've got a much newer server I'd be happy to donate... the problem is just shipping to kalmar :D22:55
teK_we are fine hardware-wise22:55
prologicthat that's the funny thing about Bash and shellshock22:55
prologicwho uses Bash for web requests?22:55
teK_charly is fixing disks really quickly22:55
jaegerYeah, he's been very good to us for the free hosting and hardware22:56
frinnstyeah charlie is a hero :)22:56
prologicI'll say22:56
prologicfree hosting all these years22:56
teK_we fixed things a day before christmas once :)22:56
jaegermavrick61: are you around? we're talking nice about you! :D22:56
teK_had a nice chat :]22:56
frinnsthaha yeah I remember, poor guy22:56
teK_he had good company (and me)22:57
frinnstGNU bash, version 3.2.57(1)-release (i686-pc-linux-gnu)22:58
teK_and I still don't think we use any bash in relation with apache.22:59
jaegerseems unlikely22:59
teK_i.e. the segfaults should originate from git hooks or something22:59
jaegerwhat do you guys think about upgrading to 3.1? I know it'd be some work but might be worth it22:59
frinnstI think its long overdue23:00
jaegerIf they do originate from git hooks we should fix those :)23:00
prologicjeager: remotely?23:00
frinnstI can be on location23:00
frinnst.. depending on when23:00
teK_My take is that I took care of most things for the last year or two so you will either have to wait a few more months or find out things yourself23:01
frinnstspring / summer project?23:01
jaegeron location would be ideal but it depends more on you guys than on me. It's not remotely convenient for me to go to kalmar :)23:01
prologicbeen a while since I've done a remote upgrade :)23:01
prologictoolchain upgrades can be tricky ;023:01
teK_we are not doing a real remote upgrade23:01
teK_come on guys, you are better than that..23:01
prologicso frinnst will have to do it then :)23:01
prologicon location :)23:01
teK_initially no23:02
prologicwe don't have kvm access to the machine though?23:02
jaegerwell, another option would be to ask mavrick to put a USB or CD into it and start SSH :P23:02
frinnsthehe relaaaaax tek :)23:02
prologicthat could work :)23:02
prologicin fact that would be ideal23:02
prologicno one has to go anywhere then :)23:02
teK_this is a linux box, right?23:02
prologicyeah and?23:03
frinnsttek is really nervous right now23:03
teK_frinnst: I'm relaxed as shit. We just had bump budget for the current project by 10% and there's more to come :-)23:03
teK_(we == at work)23:03
teK_so the thing is23:03
jaegerObviously we're not all on the same page. Personally I don't see a big problem with upgrading. Yes, it'll take some manual intervention. Also, I don't see a big rush on it, I just asked to get us talking about it23:04
prologicI'm on the same page :)23:04
teK_it's dead simple to pull the server files from our working backup for a local test of either bumping things manually (probably not) or doing a clean install and migration of services23:04
jaegerFor what it's worth the only "upgrade" I'd suggest would be an upgrade from a CD/USB, not a manual compiling upgrade. I'm not against a clean install, though, if necessary.23:05
prologicI'd do the CD/USB thing23:06
jaegerEven a CD upgrade would likely have some unexpected quirks since it's so old now23:06
prologicany quirks would likely be in configuration of newer software23:06
prologicand possibly outdated software23:06
jaegerlots of packages have been obsoleted or renamed, like util-linux-ng -> util-linux, etc.23:07
jaegerstuff that we could of course track by looking at setup-helper from version 2.4 up to now23:07
prologicwe could also upgrade it and point to a mirror for rsync (core/opt/contrib) and website23:08
prologicuntil the upgrade/migration is done23:08
teK_you can go on that kamikaze trip without me23:08
Worksterget charlie to do the cd upgrade of each version to 3.123:08
jaegerIs this a touchy subject for some reason? I honestly don't get what the deal is23:08
teK_no not touchy at all23:09
prologicme neither :)23:09
teK_I just don't like doing things the hard way23:09
jaegerI'm not trolling, I'm genuinely curious. you seem agitated about this23:09
prologicit's just an upgrade23:09
jaegerOk, so what would you suggest, then?23:09
teK_if you are going by CD it's not AN upgrade it's six upgrades23:09
jaegerI was talking about a single CD upgrade, not all of them, for what that's worth23:10
jaegeras in "put in the 3.1 CD" and go23:10
Worksterthat'll break most likely... *shrugs*23:10
teK_done on the live object without a local test?  hell no :)23:10
jaegerI never suggested doing any of it without testing23:10
frinnstI upgrading our old server to 3.1 really an option? we need to keep maintaining i68623:11
frinnstI'd rather get a new server somehow and install & migrate23:11
jaegeri686 isn't as much work to maintain as we thought it might be, fortunately23:11
frinnstyeah but the hardware is oold. fans might to start to burn out and whatnot23:12
prologicDo we actually have CRUX users on i686?23:12
frinnstoh fuck, late23:12
teK_jaeger: no you didn't, just as much as I did not suggest you did :P23:12
jaegerteK_: would a clean install be your preference?23:12
teK_considering this would be worthwhile, imho23:13
jaegerI've no objection to it23:13
teK_as I said, my take is that I know the installation best23:13
jaegerI'm honestly not sure which would be less work23:13
teK_and I have a feeling that this would be the best thing to do23:13
jaegerYou undoubtedly do23:13
jaegerEven if setting up a new set of services were a lot of work (which it really shouldn't be since we have all the current stuff and backups) there's something to be said for a clean slate23:15
teK_users, old cruft and hopefully not: no backdoors etc23:15
jaegermaybe we should consider redoing some things, too, like having local users for git access23:16
jaegeror maybe nobody cares about that either way, heh23:16
teK_there is a ton of possible improvements, yes23:17
teK_I made some good experience with gitolite23:18
jaegerI use gitosis, myself, which I think is similar to gitolite23:18
jaegerI think gitolite is newer23:18
prologicI've used gitolite too23:20
prologicit's not too bad23:20
teK_lost the admin repo once :P23:20
prologicI'd +1 the local users23:20
jaegeroops, heh23:20
prologicbe nice to have local accounts for devs, contributors, etc :)23:20
jaegerIt's a shame that gitlab and similar ones are such heavy stacks, they look nice23:21
teK_I don't see a point in that23:21
teK_i.e. nobody ever stepped up to ask for shell access for the last X years23:21
jaegerI've used mine many times to check on some config or occasionally add/remove a repo, or to fix git hooks23:22
jaegerbut not constantly23:22
teK_that's an admin, not a dev/contributor task :)23:23
jaegersorry, I'm mixing up the conversation, you're right23:23
jaegerteK_: unrelated to this stuff, any questions/comments about that mirror cleanup script?23:23
jaegerIf you've had time to look at it23:23
teK_I checked the idea behind it (a sqlite db)23:24
teK_I did not comment yet because, imho, our mirroring approach is not perfect in the sense that everybody does the synching themselves23:25
teK_I think we should have a master mirror and rsync from that to the other mirrors because this makes add new mirrors a one-liner23:25
jaegerMy VPS has a bandwidth limit so I considered that maybe the best way. Though it is a generous limit23:25
teK_we have yet to hit that :P23:26
teK_I have had ~3-4 requests during the last three days23:26
teK_ we have to go official once the cleaning thing is sorted out23:26
jaegerI'm not worried about it soon, was just thinking ahead :)23:26
teK_once we surpassed arch :)))23:26
jaegeranyway, the cleanup script isn't tightly tied to the new file download, at least23:26
prologic+1 for master mirror23:27
teK_I still got no feedback wrt download speeds from to australia (i.e. the worst case :D)23:28
teK_I can offer to run the rsync thingy but we can go another route, too23:28
jaegermaybe even host the master collection on but only allow downloads (via your rewrites and script) from the leaf mirrors?23:29
teK_the thing with the master mirror is that one could add a trigger to our commit hooks so the distfiles will be fairly up to date23:29
jaegerthat would also faciliate trigger hooks23:29
teK_well bandwidth has been an issue in the past for crux.nu23:29
teK_that's true, local hooking would be far more elegant.. but we could also limit things with SSH keys on the master server23:30
jaegerplenty of options, I guess23:30
teK_(I am thinking of the command="" directive of auth_keys)23:30
teK_yes indeed23:30
teK_gotta go. The next telephone call is lurking just in the morning D:23:32
teK_licensing talks <323:32
jaegerok, take care23:33
teK_see you soon, and frinnst / jaeger / Workster / prologic please make up your mind about the master mirror :)23:35
prologicmaster mirror23:35
prologicI opt for that :)23:35
prologicand we (slaves) rsync form it23:35
teK_one mirror to enslave them all23:35
teK_afk for now.23:35
teK_good night, folks23:35

Generated by 2.11.0 by Marius Gedminas - find it at!