IRC Logs for #crux Wednesday, 2017-08-02

*** nthwyatt has joined #crux00:02
*** darfo has quit IRC00:06
*** john_cephalopoda has joined #crux00:08
*** john_cephalopoda has left #crux ("Trees are hiding until the final day.")00:14
*** druid_droid has quit IRC00:31
*** BitPuffin has quit IRC00:47
*** edvb has joined #crux01:08
*** https_GK1wmSU has joined #crux01:50
*** https_GK1wmSU has left #crux ()01:51
*** darfo has joined #crux02:07
*** nthwyatt has quit IRC02:07
*** _________mavric6 has quit IRC02:31
*** _________mavric6 has joined #crux02:32
*** brian|lfs has joined #crux03:01
brian|lfsI got a laugh over the kernel question while I was on vacation I just got home03:02
*** brian|lfs has quit IRC05:01
*** brian|lfs has joined #crux05:19
*** tsaop has joined #crux06:26
cruxbot[opt.git/3.3]: mailx: upate to 14.9.207:15
*** edvb has quit IRC08:03
*** q6AA4FD has quit IRC08:31
*** q6AA4FD has joined #crux08:39
*** john_cephalopoda has joined #crux09:36
*** Romster has quit IRC10:01
*** tsaop has quit IRC10:10
*** Romster has joined #crux10:20
*** tsaop has joined #crux10:22
*** Romster has quit IRC10:32
*** Romster has joined #crux10:37
*** onodera has joined #crux10:39
*** abenz_ has joined #crux10:46
*** tsaop has quit IRC10:48
*** tsaop has joined #crux10:48
*** tsaop has quit IRC10:50
*** tsaop has joined #crux10:50
*** vitamin has joined #crux11:06
vitaminCould somebody tell me please what is this 'File-based locking initialisation failed.' message in dmesg?11:08
*** xor29ah has quit IRC11:30
*** xor29ah has joined #crux11:31
*** vitamin has quit IRC11:45
*** abenz_ has joined #crux12:13
*** tsaop has quit IRC12:19
*** BitPuffin|osx has joined #crux12:22
frinnstoh he left12:42
frinnstanyways, its a lvm thing during initialization iirc12:43
frinnstharmless12:43
*** Kruppt has joined #crux13:22
*** Kruppt has quit IRC13:22
*** abenz has joined #crux13:25
*** nthwyatt has joined #crux13:31
*** darfo has quit IRC13:34
*** vitamin has joined #crux13:48
cruxbot[core.git/3.3]: rc: use --sysinit with vgchange to avoid a warning at system start14:10
*** abenz_ has joined #crux14:14
ryu0am I the only one to notice ed fails to build from core?14:16
*** abenz has quit IRC14:17
frinnstbuilds for me14:17
ryu0it fails here on a barebones core install.14:17
ryu0it can't extract the package.14:17
frinnston 3 different boxes14:18
*** chinarulezzz has quit IRC14:18
ryu0some missing package? what even decompresses lzip?14:18
frinnstyeah I think we patched for .lz post release14:18
frinnstlibarchive is used14:19
ryu0frinnst: if i wanted to produce a new set of core packages to install from new time, would using the ISO scripts be enough?14:19
ryu0next*14:19
frinnsthttps://crux.nu/gitweb/?p=tools/pkgutils.git;a=commit;h=9033d1650cafd825400e836a965c7744f99b3dc014:19
frinnstwhen did we release 3.3?14:20
ryu0February 2nd, officially.14:20
frinnstbut we didnt roll a new pkgutils release for a while14:20
ryu0either way i guess it's resolved with a sysup.14:21
frinnst2017-04-08 Fredrik Rinnestambump Makefile version pkgutils-5.40.214:21
frinnstyep14:21
ryu0what about my other question?14:21
frinnstso, you want to roll new packages for core? Yes, the iso makefile would do this14:22
ryu0yea. the reason is i'd like to be able to make my own set of updated packages to routinely update LXD tarballs for crux.14:23
frinnstyou could just symlink your /usr/ports/core to <iso.git>/ports/core14:23
ryu0so fresh installs don't have to do a lot of work with sysup every time.14:23
frinnst"make core" iirc14:23
ryu0Though I also strip a few packages that are irrelevant to LXD, namely bootloaders.14:24
ryu0frinnst: i was writing a few ports and i discovered a problem. this one port i was writing generates different footprints based on some dynamic value, perhaps the time of day. so it will always produce a mismatch. will this cause trouble for actual users?14:26
frinnstother than building the port? Doesnt sould like it14:27
ryu0just seems insurmountable for some packages. best i can think to do is omit the parts that will vary.14:29
ryu0so it registers as "new"14:29
ryu0it's a shame the footprint format doesn't allow patterned file pathes to work in place.14:29
ryu0that could fix the problem.14:30
*** SiFuh has quit IRC14:32
*** SiFuh has joined #crux14:34
vitaminCould somebody tell me please why I have 'File-based locking initialization failed.' in my dmesg? Is it something to be worry about?14:38
*** tilman has joined #crux14:40
juevitamin: it's a harmless warning, should be fixed with the core/rc update I've pushed a few minutes ago14:43
cruxbot[opt.git/3.3]: [notify] samba: update to 4.6.614:43
cruxbot[opt.git/3.3]: sqlite3: update to 3.20.014:43
vitamin@jue: thank you very much for your quick reply. it's good to know I didn't botch my first kernel compile:)14:44
juenp14:44
onoderaWhat changed in rc?14:54
*** abenz_ has joined #crux15:10
jueonodera: https://crux.nu/gitweb/?p=ports/core.git;a=commitdiff;h=cc7e4277a3d53e2cce5e3197f24a4b03fb1fdb1415:10
*** BitPuffin|osx has quit IRC15:39
ryu0frinnst: i think it's time to bump automake to 1.15.1: http://git.net/ml/bug-automake-gnu/2017-07/msg00020.html15:43
frinnstwe are on 1.15.116:03
ryu0... nevermind seems my local thing was outdated16:16
onoderajue: ty :)16:24
pedjajue, would you sign samba port with opt and not your own key, please :) ?16:28
juepedja: strike :)16:33
cruxbot[opt.git/3.3]: samba: update .signature16:34
pedjathank you, good sir :)16:37
*** tsaop has joined #crux16:55
vitaminThis is probably stupid question but I don't understand it. Why is 'tmpfs on run' mounted on my system? I don't have either 'tmp' nor 'shm' listed in my /etc/fstab file.16:58
vitamin*tmpfs on /run16:58
onoderavitamin: I'm pretty sure /etc/rc does that16:59
onoderahmm maybe not16:59
onoderaI'm pretty sure shm doesn't get mounted autmotically, but yeah tmp does17:00
vitaminAnd why is it so big?17:00
vitaminhttps://cfp.vim-cn.com/cbdwP17:00
vitaminAnother question - what is it needed for?17:03
just_funvitamin, https://en.wikipedia.org/wiki/Tmpfs should answer to you questions17:09
just_funI use /tmp mounted as tmpfs to save my downloads :)17:10
just_funjue, you've been very fast with the jN/samba, but before thanking you, I wanted to triple check that :)17:14
vitaminjust_fun: not really. I still don't know what is creating that tmpfs on /run? I don't see anything in /etc/rc related to that.17:24
just_funYou've got me. Maybe it didn't answer all your questions, but some did :D17:25
just_funryu0, I think you can override some pkgmk functions in that Pkgfile, so that won't fail when the footprint is checked. Maybe make_footprint is enough.17:31
just_funvitamin, /sbin/start_udev is mounting you /run17:33
pedjavitamin, udev does. check out start_udev script from eudev port.17:33
just_funyour /run17:33
just_funpedja, I was first :)17:33
vitaminthanks guys. /sbin/start_udev is doing this indeed.17:36
pedjaRedHat is deprecating BTRFS https://news.ycombinator.com/item?id=1490777117:53
pedja'Lack of interest from paying RHEL customers' looks like the most probable reason to me17:54
tsaopis BTRFS a viable alternative to ZFS?17:55
*** nogagplz_ has joined #crux17:56
pedjadefine 'viable'17:58
pedjait depends, as always :)17:58
*** nogagplz has quit IRC17:58
tsaopviable as in "robust enough for home NAS usage"17:59
pedjaif you need raid5/6, ZFS is your only option18:00
pedjaI know that many people use something like FreeNAS as a NAS solution, and jaeger has been using ZFS for a while now without issues, iirc18:03
tsaoppedja: Yeah, I was reading about that earlier18:03
tsaopa NFS box using FreeNAS and ZFS would be an interesting project18:04
pedjabut please don't fall for 'ECC RAM or ZFS will eat your data' myth :)18:05
jaegerI use ECC RAM because my ZFS box's hardware uses it. But yeah, it's not *required* by ZFS18:08
pedjait's been debunked so many times, and still people freak out :)18:08
tsaoppedja: Eh, I was wondering about that18:09
tsaopWell, at least it means I don't have to pay more for RAM18:09
pedjawhat's the price diff, anyway?18:10
pedjaor is most of the cost mb/cpu combo that supports it?18:10
*** leah2 has quit IRC18:12
pedjaRyzens support it, but are there any m/b for it that do too?18:15
jaegermine does. http://www.biostar.com.tw/app/en/mb/introduction.php?S_ID=879#specification18:17
pedjaMini ITX? nice.18:23
pedjais ~130 euros here (here as in EU, not .rs)18:23
jaegerpedja: I could swear I'd shown it to you guys before - http://imgur.com/a/Z7LJh18:24
pedjaI must've missed that :)18:25
jaegerfair enough :)18:25
*** leah2 has joined #crux18:27
pedjaany issues with building it? being Mini ITX and all18:28
*** BitPuffin|osx has joined #crux18:28
jaegerIt's a little fiddly to build in any case that small but no issues. Just have to plan a little and scrape some fingers18:29
tsaopjust mount a mini itx board onto a standard atx case18:29
jaegermost ATX cases do have the mounting holes for ATX, uATX, and mITX, but it looks very silly when you do that. :)18:30
tsaopthat's exactly the reason18:30
tsaoplike that raspberry pi mounted onto a server rack18:30
tsaoptaking like 5% of the available space18:31
pedjamore importantly, did frinnst rant a little because you put Nvidia card in it :) ?18:31
jaegernot a strong one, heh. I did want to build it with an RX580, actually, but thanks to cryptocurrency miners I couldn't buy one18:31
vitaminwhat is prefered way to change I/O scheduler?18:32
jaegerI use a udev rule for it but I suppose the "preferred" way is whatever you prefer18:33
vitaminwell, I don't have any preferences. can I just put simple 'echo deadline > /sys/block/sda/queue/scheduler' in some file?18:36
jaegersure, like /etc/rc.local for example18:36
jaegerhttp://jaeger.morpheus.net/linux/files/99-block-scheduler.rules for comparison, if you want to see the udev route18:36
vitaminmany thanks @jaeger18:37
jaegerwelcome18:37
vitaminbtw, do you use noop for ssd? I heard that deadline is better and that's why I want to change that.18:38
jaegerI use noop. I have not benched the two to see if one's better than the other. My use of noop is based on my opinion that the SSD's controller probably knows better than I do18:38
vitaminhttps://www.phoronix.com/scan.php?page=news_item&px=Linux-4.9-IO-Scheduler-SSD18:39
vitaminyeah, that makes sense18:40
jaegeroh, nice, new version of zfs on linux recently. will have to try that18:41
pedjaI think you're gonna like resumable compressed send/receive :)18:48
jaegerIt's rare that I need to use send/receive but I do like the concept18:49
pedjaquite a big release, with native encryption postponed for 0.7.x, iirc18:50
vitaminanother question - is it safe to change cflags in /etc/pkgmk.conf file? for example '-Os' instead of '-O2'?18:51
pedjaI wouldn't recommend it18:51
pedjabut if you know what are you doing, and are ready for occasional breakage, go for it18:54
jaegerYou can always try it and see... but it's possible that will break things. Varies from software to software18:54
vitamindoes it mean that I shouldn't be messing with /etc/pkgmk.conf file? I mean except changing 'makeflags' line?18:59
pedjaset 'PKGMK_IGNORE_NEW=yes'19:03
joacimit is ok to mess around with pkgmk.conf, but certain settings are less safe than others19:05
joacimi usually leave cflags at default19:05
pedja'man pkgmk.conf'19:05
vitaminpedja: what does this mean?19:06
pedjaread the man page for it19:06
joacimnothing that cannot be fixed by a simple pkgadd from the iso tho =)19:06
just_funI've been building the last weak updates with some fortify flags, and I didn't have any problems. https://github.com/therealfun/crux-ports/blob/master/pkgmk.conf/hardening.orig19:07
just_funweek19:07
just_funBtw, I think MAKEFLAGS=-jN in pkgmk.conf could cause more porblems than -Os instead of -O2.19:08
vitaminpedja: I got no idea what 'footprint mismatch' is.19:08
just_funlol19:09
*** chinarulezzz has joined #crux19:29
jaegerNote that changing may or may not cause problems that are unrelated to CRUX. For example, some packages may not build with -Os or -march=native upstream19:45
jaegers/changing/changing CFLAGS/19:45
*** BitPuffin|osx has quit IRC19:51
onoderajust_fun: I don't think so, because (most) ports that fail with -j>1 specifify -j in the port19:53
vitamin@jaeger: I think I'll just play around a bit with those CFLAGS and see how it goes.19:53
onoderaI've beem using -O2 and -march=native without real problems for a long time now btw. not sure about -Os19:54
just_funonodera, I use CRUX from Feb/March or something, I didn't had much experience with source distros and I was excited about using MAKEFLAGS=-j$(nproc), but seeing the venerable core/ed and efi*something progs that had their ports fixed with -j1 *in 2017*, and not some cool-new-prog, it made me be a little more conservative. I do use parallel builds, but only for selected/heavy/worthy ports. If the upstream mentions parallel builds, that app has a good20:25
just_funchance to be tested, otherwise it has good chance not be tested and race conditions can be inserted without being noticed.20:25
cruxbot[contrib.git/3.3]: cppcheck: update to 1.8020:29
*** edvb has joined #crux20:42
*** tsaop has quit IRC20:49
*** vitamin has quit IRC20:53
*** https_GK1wmSU has joined #crux20:54
*** https_GK1wmSU has left #crux ()20:54
*** ryu0 has quit IRC21:15
*** ryu0 has joined #crux21:49
*** q6AA4FD has quit IRC22:01
*** q6AA4FD has joined #crux22:12
*** amonn has quit IRC22:15
*** amonn has joined #crux22:16
*** https_GK1wmSU has joined #crux22:37
*** https_GK1wmSU has left #crux ()22:40
*** john_cephalopoda has left #crux ("Trees are hiding until the final day.")22:45
*** Dell has joined #crux23:31

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!