IRC Logs for #crux-devel Wednesday, 2015-04-01

*** Feksclaus has quit IRC01:13
rmullon the topic of USB HDDs, if I power up my external USB HDD (bus powered), I can't mount it until I fdisk it and quit without making any changes.01:41
rmullevery time01:42
rmullpretty weird.01:42
jaegervery odd01:59
Romsterlook at dmesg when it's plugged in does it read the partition table?01:59
Romstermaybe eudev rule is messed up?01:59
rmullThere is weirdness in dmesg - I'll isolate it and post it somewhere for a review when I get a chance02:13
*** Workster has joined #crux-devel02:15
*** Workster has quit IRC02:20
*** Workster has joined #crux-devel02:20
*** Workster has quit IRC02:21
*** Workster has joined #crux-devel02:22
*** Workster has quit IRC02:28
*** Workster has joined #crux-devel02:29
*** Workster has quit IRC02:29
*** Workster has joined #crux-devel02:29
*** mavrick61 has quit IRC02:30
*** mavrick61 has joined #crux-devel02:31
*** heroux has quit IRC03:18
*** heroux has joined #crux-devel03:18
prologicjaeger, what kind of smartd config do you use on yor system?03:47
prologic$ egrep "^[^#].*" /etc/smartd.conf03:47
prologicDEVICESCAN -m -M test -M exec /root/bin/smartdnotify -s S/../.././02 -s L/../../7/0403:47
prologicI'm using this on my desktop and media03:47
prologicgood enough for my nas too?03:47
jaegerI don't, actually. I'm logged into that machine quite frequently so haven't really worried about it03:48
prologicI this this config scans all devices03:48
prologicand does short tests once per day03:49
prologicand long tests once a week03:49
prologicI think Rosmter recommended this03:49
prologicIv'e been getting errors on the 3.5" disk in my desktop and media server (both) and well the data is still there and the disks haven't died completely03:49
prologicso *shrugs*03:49
jaegeryeah, smart failures are usually early warning03:51
prologicso the question then is this03:52
prologicif you start seeing smartd errors on parts of the disk03:52
prologicwhich I've been getting for a while on my desktop and media spinning drives03:52
prologicNB: their rootfs/os is an SSD03:52
jaegerstart replacing them03:52
prologicthen when do you replace them?03:52
jaegerbefore they lose data :)03:52
prologicwhen zfs starts seeing read/write errors?03:52
prologicso first error on any disk in the array from smartd03:53
prologicswap it out for a new disk?03:53
jaegeryeah, or when smart starts reporting problems if you want to get it done early03:53
prologicsadly in that case the WD drives I have in my dekstop and media server are really shit drives then03:54
prologicthe desktop one a 3TB WD Black quite new03:54
prologicthe media server has a 1TB WD Black a few years old now03:54
prologicso great MTBF rates there :)03:54
jaegerwell, what kind of warnings are they? if it's just heat it might be a cooling issue03:54
prologicor error rates in general03:54
prologicwell could be03:55
jaegerone of the drives in is reporting heat smart errors constantly but still running, heh03:55
prologicI've kinda of been ignoring the smartd errors now03:55
prologicit's like the "boy that cried wolf"03:55
prologicsoon as I'm done putting data on the NAS03:55
prologicI think I'll rip these disks out and throw them out03:55
jaegerlook up that drive, see if anyone else has a history of errors with them03:55
prologicyeah nah03:55
prologicno heat warnings here03:55
prologicthere are in fact bad sectors from the reports03:55
prologicwell I'm just gonna throw em out tbh :)03:56
jaegerIf there are bad sectors the drive is done03:59
jaegerthat never gets better, only worse03:59
prologicyeah I agree :)04:02
prologichence why I'm going to throw them out04:02
prologicDevice: /dev/sdb [SAT], 2 Offline uncorrectable sectors04:02
prologicfixed the badly encoded filenames iwth convmv04:10
prologic$ zfs get compressratio data/media/music04:12
prologicNAME              PROPERTY       VALUE  SOURCE04:12
prologicdata/media/music  compressratio  1.00x  -04:12
prologicI guess I should have turned it off?04:12
prologicdata/media/music       21T   85G   21T   1% /mnt/data/media/music04:13
*** greduan has quit IRC04:16
*** nomius has joined #crux-devel04:22
nomiusHey folks04:22
nomiusI just want to congratulate you guys... In the last few days I've been testing the netbsd ports... And I don't know if it's because of the simplicity of Crux ports or what, but I find yours better and easier...04:24
*** mechaniputer has quit IRC04:24
nomiusThe only thing that I like about netbsd ports is the fact that it has 12000 ports :-P04:26
prologicwe don't have 12,000 ports simply because we don't port every piece of crappy software out there :)05:04
*** sudobaal has joined #crux-devel05:37
Worksteri wouldn't be doing long tests quite so often06:17
Worksteronce a month bi-weekely at the most06:17
Worksterlike 6 hours for some drives to do a long test during that time degraded performance06:17
prologichow long for 3TB drives?06:21
prologiccould I just do it at 1am on a Sunady?06:21
prologicI assume it'd do all in paralell06:22
Worksterif you look at smartctl -a /dev/sdx06:23
Worksterit'll actaulyl say estimated time in minutes06:23
Worksterunder all the stats06:23
Worksterusually it's about an hour per a TB06:23
Workster3TB ~3 hours06:23
Workstersometimes a bit more.06:23
prologickk let's see06:24
Worksterhmm wonder if there is a way to stagger the checks so they don't all land on the same week06:24
prologicDevice is:        Not in smartctl database [for details use: -P showall]06:25
prologicis that an issue?06:25
Workster-a should show stats still06:25
prologicShort self-test routine06:26
prologicrecommended polling time:  (   1) minutes.06:26
prologicExtended self-test routine06:26
prologicrecommended polling time:  ( 453) minutes.06:26
prologicthis one?06:26
Worksternot really an issue but you wont be able to get the temperature when it's not in the database06:26
Worksterthat's the one06:26
Worksterlong aka extended06:26
prologicyeah so ~7.5hrs06:27
Worksterthat time will change it's an estimation from last run or some default the manufacture set06:27
Worksteri'm heading home06:28
nomiusprologic: that makes sense (on the 12k ports :-P)06:37
prologicI hope so :)06:41
prologicgenerally most of us (I'm guilty of) maintain what we use06:42
prologicand sometimes not very well :)06:42
prologicI mean keep them up-to-date :)06:42
*** novak has joined #crux-devel07:08
pitillohey, could be possible to move two interesting ports from contrib and a private repo to opt? (ncdu and fail2ban)09:02
pitillofrinnst and Romster respectively09:02
frinnstncdu ftw09:03
frinnstwhy move from contrib?09:03
pitilloit hasn't deps... it's a terminal/ncurses tool... very usefull...09:04
frinnsti agree09:05
frinnstjust wondering what the benefit of moving from contrib to opt would be09:05
pitillodo I need to get entire contrib collection (or get the port by hand if I haven't contrib) to get it?... it's minimal and it's a nice discover for me since some time ago09:05
pitilloI have no problem to get ports by hand... just asked because I found both ports really interesting to get them from opt directly09:06
frinnstoh, Its me that maintains ncdu09:06
pitillofail2ban needs a bump too Romster09:07
pitilloyeah frinnst09:07
frinnsti'll move it later to day unless i forget :)09:07
pitilloI'll remeber you if I don't see it in opt soon :P09:07
prologicwhat does ncdu do?09:07
pitillodon't worry, there is no need to rush :)09:07
pitilloncurses du :)09:07
pitilloummm nop... it refreshes... but it's very handy tool09:09
frinnstyeah, you can navigate around09:10
Romsterpitillo, yeah i was just looking at fail2ban yesterday but not finished it yet09:19
Romsteryou want fail2ban in opt?09:19
Romsteri never really considered to do that.09:20
pitilloI'd like to see it in too (both are simple ports, without deps and very usefull), but like I said, I have no problem to get ports by hand09:21
Romsterwell fail2ban is useful.09:21
prologicwe should put zfs in opt too when 0.6.4 comes out :)09:23
Romster-_- and how many use zfs09:23
Romster1 maybe 2?09:23
prologic2 that I know of :)09:24
prologicjaeger and myself09:24
Romsterelse we would see ports in personal repositories09:24
Romsteri'd like pyhton3 in opt09:24
Romsterlooks like i'll have todo that.09:24
prologicno don't :)09:24
prologicPython 2.x is still the defaco standard09:24
Romsteri need python309:24
prologicand why Python in opt?09:24
prologicI mean I love the langauge et all09:24
Romsteri have a port that needs 2 modules of python309:25
prologicbut often I need many more things installed than just Python09:25
prologicwhich port(s)?09:25
Romstersetuptools3 and python3 along with a couple of other.09:25
prologicI think it should stay in contrib tbh09:26
Romstereither i backport 2 new modules or use pyhton3 which is it?09:26
prologicdon't I maintain all python versions?09:26
Romsteri think so... so09:26
prologicin personal ports09:26
prologicI should throw them in contrib09:26
Romsterdo we need all of them and not just python 3.409:26
prologicbackport and submit patch upstream09:26
prologicpeople really shouldn't be writing for Python3 without supporting at least 2.7 still IHMO09:27
prologicwe're not there yet :)09:27
prologicI'm going to drop 3.1/3.2 yeah09:27
prologicbut keeping/maintaining 2.6, 2.7, 3.4 and pypy09:27
prologicalso fyi pypy-stm was released proper the other day09:27
Romsterit supports 2 bah i'll just make the 2 extra packages fuck python3.09:27
prologicno more GIL09:27
Romsterso what about dropping 3.3?09:28
Romsterand how long before everyone moves to python309:28
Romsterfeels stagnant that we haven't got pyhton3.4 in opt already09:30
Romsteralso we need qt5 soon in opt for kde for alan09:30
Romsterif sepen is up to that task.09:30
Romsteralso is jue on holidays?09:32
prologicstagnent for whom?09:39
*** crash_ has quit IRC09:40
prologicis there a service that will give you better github downloads?09:50
Romsterby better you mean one that supplies names on there files?09:51
prologicand also gives you nice tarballs of git repos at particular revisions/brances09:51
Romsterisn't there a feature request open at github for that/09:51
prologiclike what I'm doing by hand09:51
prologicgit clone ...09:51
prologicgit archive ...09:51
prologicscp ...09:51
prologicit not I could write such a service09:52
prologicit would be really trivial actually09:52
Romsteroh but you want it for a particular commit to download.09:52
Romsteri'm sure it's already been done09:52
prologicfind one :)09:52
prologicif not I'll write one09:52
teK_I am concerned with folder name != $name. I think this is a bad idea09:54
Romsterdirectory name should match name=09:54
teK_see the ML :)09:55
prologicteK_, my zfs/spl git ports?09:55
prologicalready on it :)09:55
teK_the one dicussed on the ML09:55
prologicduplicate discussions in #crux too09:56
prologicalready fixing it :)09:56
teK_oh ok :)09:56
RomsterI maintain these ports :) Hi!09:56
RomsterI don't think it confuses pkgutils or prt-get09:56
RomsterThese ports are installed on my NAS.09:56
RomsterI hope they work okay for others! If not let me know!09:56
Romsterthat does not say your fixing them though <_<09:56
prologicsee above09:57
prologicduplicate discussions09:57
Romsteryeah i know...09:57
prologicI'm not going to respond to both :)09:57
Romsterno i was augmenting the same discussion09:57
prologicI'll send a email once the updated versions are pusehd :)09:57
prologicok :)09:57
prologicnothing like a live test09:57
prologicrebooting shortly09:57
prologichad to buidl nfs into the krenel anyway09:57
Romsterits like hearing half a converstiaon from on party and not the other. you don't get the whole picture. same with irc and mailing list.09:58
prologicyeah yeah09:59
prologichard to keep up everyeone at once :)09:59
Romsterand tek wasn't sure.09:59
Romsterthat is true.09:59
Romsterguess not everyone has python3 slackware don't either10:00
Romsterah frinnst got trident updated. was about to do that.10:08
*** crash_ has joined #crux-devel10:31
*** greduan has joined #crux-devel12:20
greduanan hour early12:20
greduanwrong channel, sorry12:20
jaegerprologic: no idea on quotas, I don't use them. Any decent quota system should be able to change them on the fly, though12:31
*** mechaniputer has joined #crux-devel15:17
*** greduan has quit IRC15:25
*** greduan has joined #crux-devel16:02
*** jue has joined #crux-devel16:02
jaegerhey, jue16:44
*** greduan_ has joined #crux-devel18:13
*** greduan has quit IRC18:14
frinnst20:16 - time for a 11hr powernap!18:16
frinnstsee you guys on the other side18:16
jaegerenjoy :)18:18
*** Feksclaus has joined #crux-devel21:04

Generated by 2.11.0 by Marius Gedminas - find it at!