Switch to DuckDuckGo Search
   December 1, 2018  
< | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | 31 | >
Toggle Join/Part | bottom
[00:17:59] *** k1l <k1l!~k1l@ubuntu/member/k1l> has quit IRC (Ping timeout: 246 seconds)
[00:18:42] *** APTX <APTX!~APTX@unaffiliated/aptx> has quit IRC (Quit: No Ping reply in 180 seconds.)
[00:18:53] *** k1l <k1l!~k1l@ubuntu/member/k1l> has joined #zfsonlinux
[00:23:40] <bunder> https://lists.freedesktop.org/archives/dri-devel/2018-November/198581.html
[00:23:44] <bunder> oh go hug yourself
[00:25:28] * DHE is gonna go write his own CoC based on linus behaviour in the past
[00:31:17] *** fling <fling!~user@fsf/member/fling> has joined #zfsonlinux
[00:36:33] *** flying_sausages <flying_sausages!~flying_sa@static.88-198-40-49.clients.your-server.de> has quit IRC (Quit: You just lost the game. Peace Out.)
[00:44:16] <bunder> he didn't post all of the patches, i suspect they'll quietly retreat
[00:45:56] *** flying_sausages <flying_sausages!~flying_sa@static.88-198-40-49.clients.your-server.de> has joined #zfsonlinux
[00:46:49] *** tmiw <tmiw!mooneer@voldemort.lifeafterking.org> has quit IRC (Remote host closed the connection)
[00:48:55] *** apekatten <apekatten!~apekatten@unaffiliated/apekatten> has quit IRC (Quit: no reason)
[00:49:19] *** apekatten <apekatten!~apekatten@unaffiliated/apekatten> has joined #zfsonlinux
[00:52:28] *** mavhq <mavhq!~quassel@cpc77319-basf12-2-0-cust433.12-3.cable.virginm.net> has quit IRC (Read error: Connection reset by peer)
[00:52:57] *** ntd <ntd!~ntd@gateway/tor-sasl/ntd> has joined #zfsonlinux
[00:53:44] *** mavhq <mavhq!~quassel@cpc77319-basf12-2-0-cust433.12-3.cable.virginm.net> has joined #zfsonlinux
[00:59:01] <blackflow> this CoC and PC sheisse is growing out of control.
[00:59:36] <bunder> i was hoping it would go away after the first go, but /shrug
[01:00:48] <bunder> also how are code comments against coc, it should cover person to person stuff
[01:01:03] <PMT> It's hard to draw a reasonable line people can agree on between not being hostile or insulting to others and not permitting dissent.
[01:01:29] <PMT> bunder: consider a comment of // FUCK BUNDER FOR THIS BARREN HELLSCAPE OF A FUNCTION
[01:01:43] <bunder> well that's a little more direct
[01:02:22] <bunder> but "this section of code is a fucking mess" is perfectly reasonable to me
[01:02:23] *** darkmeson <darkmeson!~darkmeson@gateway/tor-sasl/darkmeson> has quit IRC (Remote host closed the connection)
[01:02:26] <ptx0> It is a dimension as vast as space and as timeless as infinity. It is the middle ground between light and shadow, between science and superstition, and it lies between the pit of man's fears and the summit of his knowledge. This is the dimension of imagination. It is an area which we call... the Code of Conduct.
[01:03:24] *** darkmeson <darkmeson!~darkmeson@gateway/tor-sasl/darkmeson> has joined #zfsonlinux
[01:03:27] <blackflow> the latter is ok. big corps hide behind their virtual entities in order to do bad stuff. *LIMITED* Liability Company. A legal _entity_, eater of sins for all the humans on its board. So yeah, fuck nvidia.
[01:04:19] <bunder> did linus ever apologize for his "so nvidia fuck you" thing
[01:04:22] <PMT> I wasn't making a blanket statement of correct or not, just observing that reasonable people can disagree on the correct line.
[01:04:34] <blackflow> ptx0: CoCs are bait. at least I know I'm dying to call it a CoCk.
[01:04:38] <ptx0> i hope no one tries forcing linus to apologise for all his past wrongs
[01:04:49] <ptx0> we'll never get through the list
[01:04:55] <blackflow> but he already did, din't he?
[01:04:58] <PMT> I'm rather inured to most forms of direct and indirect hostile statements, but just because I don't blink from getting a diatribe in my inbox does not make that a reasonable thing to do.
[01:05:10] <blackflow> in that article/interview/what-was-it, just before he excused himself few weeks ago
[01:05:36] <PMT> People kept expecting it to come out that he was being extorted or blackmailed into it, but nothing of the sort has come out.
[01:05:38] <bunder> i love the phoronix comments... https://www.vidarholen.net/contents/wordcount/#fuck*,shit*,damn*,idiot*,fucking*,moron*,dick*
[01:07:58] <zfs> [zfsonlinux/zfs] Detect IO errors during device removal (#8161) new review comment by Brian Behlendorf <https://github.com/zfsonlinux/zfs/pull/8161#discussion_r238025042>
[01:08:45] <blackflow> the disagreement on the correct line is exactly why CoCks are scheisse. I doubt people need a written document to tell them what harassment is and that it's not okay. Because those that will harass will do so regardless of the existence of a CoC
[01:09:14] <blackflow> but its a nice tool to exert censorship and shitty PC agendas - QED
[01:09:23] <ptx0> nah
[01:09:33] <ptx0> you are overblowing it, just like the person submitting the patch
[01:09:36] <ptx0> that's the main issue here
[01:10:40] <ptx0> reminds me of the "criminals won't follow laws so shouldn't make guns illegal" mindset, except that having a law on the books that's being broken, can substantially increase the chances of obtaining a conviction
[01:10:56] <blackflow> hardly. it's not just this one instance. the whole discussion/drama over the CoC in the first place, was big circus. so much so that patches were submitted to include aliens (as in extra terestrials) into so called inclusivity.
[01:11:01] <ptx0> in the case of the CoC, something that was not officially "against the rules" now is
[01:11:15] <blackflow> emphasis on "big circus" :) that's what I think it all is.
[01:11:34] <ptx0> yes, well, many people signed off on it
[01:11:38] <zfs> [zfsonlinux/zfs] Detect IO errors during device removal (#8161) new commit by Brian Behlendorf <https://github.com/zfsonlinux/zfs>
[01:13:20] <PMT> blackflow: the two extremes here are both subjective perception judgments, and the CoC is an attempt to explicitly state where the line is so that you can avoid a class of problems where people perceive things as being of differing severity, because you have explicit rules about it.
[01:14:14] *** Essadon <Essadon!~Essadon@81-225-32-185-no249.tbcn.telia.com> has quit IRC (Quit: Qutting)
[01:14:16] <ptx0> and i think it has made the LKML a better place
[01:14:28] <ptx0> and you can suck a dick if you disagree
[01:15:04] <bunder> i've only ever used the lkml once, and they still haven't fixed my trackpad yet
[01:15:26] <blackflow> PMT: and it's failing at that since people have different interpretations
[01:15:31] <PMT> A number of my interactions with people I've known for a long time would be viewed as objectionable if they were with someone I had just met, or met in a professional setting, because we both agree where the line is and there's not much in the way of power dynamics where one of us is compelled to take the other's position for fear of X.
[01:15:50] <bunder> psmouse.synaptics_intertouch=1 lul
[01:16:32] <PMT> blackflow: yes, the subjective aspects are both the benefit and weakness of concrete rules, but are you arguing that it only solving some kinds of problems makes it worthless?
[01:18:02] <blackflow> I'm saying it's not really solving any problems. haters gonna hate regardless of the existence of a CoC.
[01:18:43] <blackflow> but *shrug*, I don't care either way. I don't need a piece of paper to tell me it's bad to harass.
[01:18:49] <PMT> blackflow: and as others said, your argument is akin to "criminals don't follow laws so we shouldn't make any"
[01:20:23] <bunder> at least the ozfs coc only applies to the dev conferences :P
[01:20:50] <bunder> i bet if ptx0 and i showed up we'd get kicked out
[01:22:08] <PMT> also i asked for comments on an idea on ozfs-dev, if you have opinions go yell at me =P
[01:22:29] <bunder> i'm not subscribed
[01:22:32] <bunder> i really should
[01:22:36] <PMT> maybe.
[01:23:59] *** ntd <ntd!~ntd@gateway/tor-sasl/ntd> has quit IRC (Quit: ntd)
[01:25:24] <bunder> the flag proposal thing? i'm all for it
[01:25:41] <PMT> Yeah, that.
[01:26:59] <bunder> even if all we get is a rewording on zpool status -v that is ;)
[01:27:33] <fling> Why is import taking 30 seconds?
[01:27:40] * cirdan yawns
[01:27:46] <PMT> fling: suffering
[01:27:46] <fling> The pool is empty mirror of two ssds
[01:28:06] <bunder> gentoo?
[01:28:11] <fling> bunder: yes
[01:28:17] <bunder> genkernel?
[01:28:20] <fling> no
[01:28:32] <bunder> using import-scan instead of import-cache?
[01:28:40] *** djdunn-n8 <djdunn-n8!~djdunn@fl-184-2-23-179.dhcp.embarqhsd.net> has joined #zfsonlinux
[01:28:42] <fling> Not sure what it is
[01:28:57] <bunder> do you have a zpool.cache file? :P
[01:29:09] <fling> /etc/zfs/zpool.cache is there
[01:29:13] <djdunn-n8> In zpool status does (repairing) = resilvering
[01:29:35] <fling> no resilvers here
[01:29:37] <PMT> repairing means "we found something wrong, we're replacing it with the correct bits from elsewhere"
[01:29:50] <fling> bunder: also happens to other pools
[01:30:03] <bunder> oh. i dunno then, i figured this was a boot thing
[01:30:20] <fling> bunder: it is slow at boot here and there.
[01:30:22] <bunder> best guess is zil replay from not exporting
[01:30:27] <PMT> fling: so trace it and see what it's doing
[01:30:33] <fling> will do
[01:30:55] <fling> I will create a testing script with an empty pool
[01:32:26] <djdunn-n8> I did a scrub yesterday on a 1.29T set of striped mirrors and it found a couple megs that failed checksum got 800k at ~50% left of a scrub today
[01:33:01] <bunder> djdunn-n8: are your drives dying? i'd be interested in seeing what smart says
[01:33:20] <djdunn-n8> Smart says nothings wrong
[01:33:29] <djdunn-n8> Even did extended tests
[01:33:43] <fling> space rays, microwaves :>
[01:33:55] <djdunn-n8> 8k-16k hours
[01:33:57] <fling> djdunn-n8: happens here all the time
[01:34:16] <djdunn-n8> I usually almost never see it
[01:34:27] <bunder> bad cables maybe?
[01:34:40] <djdunn-n8> I don't think so
[01:34:45] *** CompanionCube <CompanionCube!samis@sortix/rw-citizenship/companioncube> has quit IRC (Ping timeout: 252 seconds)
[01:34:56] <fling> bunder: will still be reported by smart then
[01:35:09] <PMT> I believe historical evidence is that SMART is only about as good as a coin flip at detecting drive failure
[01:35:10] <djdunn-n8> Its all 4 drives doing it suddenly
[01:35:34] <fling> PMT: exactly!
[01:35:54] <bunder> power supply sagging?
[01:36:24] <djdunn-n8> Sagging?
[01:36:44] <bunder> https://en.wikipedia.org/wiki/Voltage_sag
[01:37:25] <djdunn-n8> I don't think so
[01:37:56] <PMT> djdunn-n8: there are a bunch of ways that things can make multiple drives misbehave at once. Or you could have had the bit flip in memory if (it's not ECC) or kernel memory corruption, then fun when it lands on disk differently than it was.
[01:38:03] <PMT> Why don't you think it might have electrical issues?
[01:38:19] <djdunn-n8> Have ecc memory
[01:38:31] <djdunn-n8> No memory errors
[01:38:51] <bunder> ddr4 already corrects a bunch of errors without ecc, it could be masked
[01:39:04] <PMT> That just eliminates RAM bitflips, not electrical things.
[01:40:22] <djdunn-n8> Its only like 7-15 errors per scrub per drive
[01:40:36] <PMT> ...what model drive?
[01:40:43] <djdunn-n8> Wd blue
[01:40:45] <bunder> since linux sensors are kindof trash, if you have a multimeter or two i'd jam them into an unused molex cable and run a scrub
[01:40:46] * PMT is having flashbacks to Certain Firmware Bugs
[01:40:53] <PMT> Okay not at least one that I saw.
[01:41:13] <PMT> bunder: I mean, if the sensors are imprecise, that's not the Linux software stack's fault. =P
[01:41:34] <bunder> a lot of chips have no drivers, and the ones that do aren't calibrated
[01:42:11] <PMT> I hear they're working on that again.
[01:43:00] <bunder> https://pastebin.com/raw/JbwLDf7i like what
[01:44:33] <bunder> actually my bloomfield is the only board i have with working sensors, the other ones just have coretemp
[01:44:45] <bunder> if you can even call that working
[01:45:58] <PMT> bunder: I mean, ipmitool sdr list is still p. good at this.
[01:46:17] <djdunn-n8> Its much too intermittent for electrical issues
[01:46:36] <bunder> i don't have impi, by server i mean "desktop acting as a server because servers are too expensive" ;)
[01:47:04] <PMT> djdunn-n8: "occasional" is not usually an argument _against_ weird electrical problems.
[01:48:09] <djdunn-n8> Extended tests will usually say something
[01:48:47] <PMT> Believing SMART tests is a fool's errand.
[01:50:56] <djdunn-n8> I guess ill go get the multimeter
[01:51:39] <bunder> i mean i'm just theorizing, it could just as easily be a bad sata chip
[01:52:52] <bunder> or metallic dust landing on the pins
[01:53:54] <djdunn-n8> Maybe sata... Boards like a year old
[01:54:18] <PMT> Seems at least mildly unlikely.
[01:54:27] <PMT> djdunn-n8: what are the drive SMART counters, in particular?
[01:56:23] <djdunn-n8> https://paste.pound-python.org/show/1dcifkfWDSL1teWoFnVc/
[01:57:36] <djdunn-n8> I guess i have to turn it off and unpleg it to check voltage
[01:57:54] <djdunn-n8> Unplog the power from the sata drives
[01:57:57] <bunder> do you shutdown a lot? start/stop and load cycle are a bit high
[01:58:37] *** djdunn <djdunn!~djdunn@fl-184-2-23-179.dhcp.embarqhsd.net> has quit IRC (Quit: Lost terminal)
[01:59:31] <bunder> i wonder if wd blue inherited the old green power saving routines
[02:00:00] <bunder> but i think you'd be dropping drives all over the place if that was the case
[02:00:20] <PMT> nope
[02:00:24] <PMT> blue is explicitly not variable RPM
[02:00:37] <PMT> plus the variable RPM drives show up as really weird numbers, not 7200
[02:00:39] *** mz` <mz`!~emerick@gandi/mz> has quit IRC (Ping timeout: 264 seconds)
[02:00:41] <bunder> no i mean spinning down when they're not used
[02:00:59] <bunder> it was kindof like a laptop mode
[02:01:12] <PMT> I know. I'm saying that those were always exclusive to the Green drives.
[02:01:24] <PMT> As the variable RPM was.
[02:02:11] <bunder> doing the math, that drive is almost 2 years old
[02:02:32] <djdunn-n8> Why am i using a multimeter i have a power supply checker...
[02:03:24] <bunder> don't those only show pass/fail?
[02:03:31] <PMT> djdunn-n8: does it specifically check the rails going to drive power, because it's possible for it to have fine amperage on one rail and not others
[02:03:33] *** alwyn <alwyn!~alwyn@unaffiliated/alwyn> has quit IRC (Quit: WeeChat 2.3)
[02:03:49] <bunder> knowing how low the 12v rail is might be better than a boolean
[02:04:20] <bunder> if its like 10, that's kindof bad
[02:04:27] *** eightyeight <eightyeight!~eightyeig@oalug/member/eightyeight> has quit IRC (Ping timeout: 252 seconds)
[02:06:38] <bunder> the cpu can run off a low 12 but it would bake the vrm since it has to draw more current
[02:07:49] <PMT> my server's PSU has a habit of sagging if I plug more than one drive into the PSU, but it's like 250W so
[02:08:05] <ptx0> best goddamned car in the yard
[02:08:37] <bunder> since when do you drive a car
[02:08:50] <ptx0> https://www.youtube.com/watch?v=Q4HY544URjA
[02:10:44] *** djdunn-n8 <djdunn-n8!~djdunn@fl-184-2-23-179.dhcp.embarqhsd.net> has left #zfsonlinux ("Quit")
[02:12:31] <bunder> okay then
[02:13:11] <bunder> can't say i didn't try
[02:14:29] *** djdunn-n8 <djdunn-n8!~djdunn@fl-184-2-23-179.dhcp.embarqhsd.net> has joined #zfsonlinux
[02:14:45] <djdunn-n8> Power supply passed on all rails
[02:15:04] <bunder> even under load? darn
[02:15:39] <bunder> i guess that leaves sata chip or cables if smart is to believed
[02:17:09] <djdunn-n8> Its dusty
[02:17:14] <djdunn-n8> Need can of air
[02:18:45] <PMT> djdunn-n8: are there any messages in syslog when it finds errors
[02:19:18] <djdunn-n8> From zfs yeah
[02:19:48] <ptx0> all you need to do is make the area outside your home into a vacuum chamber
[02:19:52] <ptx0> then the dust gets sucked out
[02:19:59] <PMT> djdunn-n8: could you pastebin the whole syslog (from /var/log/syslog or /var/log/messages, not /var/log/kern.log or similar) from then
[02:24:17] <djdunn-n8> Yeah i just got done dusting it out
[02:24:28] <djdunn-n8> )
[02:25:13] <djdunn-n8> Ih couldnt have anything to do with the now zfs version?
[02:25:47] <ptx0> and only you have the issue? unlikely
[02:26:16] *** CompanionCube <CompanionCube!~samis@sortix/rw-citizenship/companioncube> has joined #zfsonlinux
[02:26:47] <bunder> i upgrade to git head on a semi regular basis, if it was zfs i'd know
[02:26:56] <djdunn-n8> Ok
[02:27:15] <djdunn-n8> Oh the 12v rail cawo out at 12.097v btw
[02:27:30] <bunder> seems reasonable :)
[02:28:07] <djdunn-n8> Only rail i tested with the multimeter beforu i got annoyed ond remembered the psu tester
[02:29:04] <bunder> i think if 5 was low you'd have more problems than storage
[02:29:17] <djdunn-n8> Doesnt seem like hard drive failure to me
[02:30:01] <cirdan> i had hard drives die under heavy load with an undersized power supply/spliiters
[02:30:11] <cirdan> went through a bunch of wd 250s one time
[02:30:16] <cirdan> 14 in my server
[02:30:23] <cirdan> well, in/on top of
[02:30:33] <PMT> djdunn-n8: my question about messages remains
[02:30:40] <PMT> Because your drives aren't logging errors
[02:30:55] <PMT> which they really should do unless the data is flipping elsewhere
[02:31:12] <ptx0> sounds like controller woes.
[02:31:37] *** eightyeight <eightyeight!~eightyeig@oalug/member/eightyeight> has joined #zfsonlinux
[02:31:49] <PMT> ptx0: depends whether the controller sees errors, and the UDMA_CRC SMART counter being 0 makes me think no.
[02:35:00] <djdunn-n8> Hmm kde seems to not want to work right
[02:35:22] *** djdunn <djdunn!~djdunn@fl-184-2-23-179.dhcp.embarqhsd.net> has joined #zfsonlinux
[02:35:49] <djdunn> hmm
[02:37:35] *** djdunn <djdunn!~djdunn@fl-184-2-23-179.dhcp.embarqhsd.net> has quit IRC (Disconnected by services)
[02:39:45] *** djdunn <djdunn!~djdunn@fl-184-2-23-179.dhcp.embarqhsd.net> has joined #zfsonlinux
[02:40:57] <djdunn> zed[7189]: Diagnosis Engine: error event 'ereport.fs.zfs.checksum'
[02:40:59] <djdunn> just that a handful of times
[02:41:23] <PMT> djdunn-n8: which logfile?
[02:41:33] <djdunn> thats in var/log/messages
[02:42:24] <PMT> Are there any other entries?
[02:42:47] <djdunn> Nov 30 07:57:51 Iris zed[7189]: Diagnosis Engine: error event 'ereport.fs.zfs.checksum'
[02:42:51] <djdunn> Nov 30 07:57:51 Iris zed[7189]: Invoking "all-syslog.sh" eid=68 pid=1541
[02:42:53] <djdunn> Nov 30 07:57:51 Iris zed[1543]: eid=68 class=checksum pool_guid=0x170AD7F17B903145 vdev_path=/dev/sdd1
[02:42:56] <djdunn> Nov 30 07:57:51 Iris zed[7189]: Finished "all-syslog.sh" eid=68 pid=1541 exit=0
[02:42:59] <djdunn> Nov 30 07:57:51 Iris zed[7189]: Diagnosis Engine: error event 'ereport.fs.zfs.checksum'
[02:43:02] <djdunn> Nov 30 07:57:51 Iris zed[7189]: Invoking "all-syslog.sh" eid=69 pid=1547
[02:43:05] <djdunn> Nov 30 07:57:51 Iris zed[1556]: eid=69 class=checksum pool_guid=0x170AD7F17B903145 vdev_path=/dev/sdb1
[02:43:06] <PMT> djdunn-n8: what part of "use pastebin" seemed confusing
[02:43:08] <djdunn> Nov 30 07:57:51 Iris zed[7189]: Finished "all-syslog.sh" eid=69 pid=1547 exit=0
[02:43:10] <djdunn> handfull of like messages with different eid=
[02:44:01] <djdunn> https://paste.pound-python.org/show/BTTz1FpBcMdy7WPDtuA2/
[02:45:44] <djdunn> i guess one scrub was this morning
[02:45:50] <djdunn> and one was tonight when i got home from work
[02:46:35] <djdunn> PMT: sorry theres really not much there, no i/o errors or anything
[02:47:53] <PMT> djdunn: that's fine. also, fascinatingly, somehow that log is missing individual characters sometimes
[02:48:01] <PMT> e.g. Nov 30 18:47:02 ris zed[23465]: eid=93 class=checksum pool_guid=0x170AD7F17B903145 vdev_path=/dev/sdc1
[02:48:40] <PMT> oh, confusingly, that was from my copy-paste of it somehow, nvm.
[02:48:54] <djdunn> uh
[02:48:59] <djdunn> ok
[02:48:59] * PMT makes note to figure out what the hell
[02:49:05] <djdunn> thats weird
[02:49:14] <PMT> As I said, it was a problem on my end, somehow
[02:49:21] <bunder> odd that we're not seeing any ata errors
[02:49:25] <djdunn> yeah
[02:49:30] <djdunn> thats why i got on here
[02:49:42] <PMT> djdunn: what ZoL version?
[02:49:53] <djdunn> 0.6.12?
[02:50:08] <PMT> No, that's not a real version.
[02:50:32] <djdunn> 0.7.12
[02:50:42] <PMT> Okay.
[02:51:23] <djdunn> ok i got screen split now so i can see stuf...
[02:52:08] *** djdunn-n8 <djdunn-n8!~djdunn@fl-184-2-23-179.dhcp.embarqhsd.net> has quit IRC (Quit: AndroIRC - Android IRC Client ( http://www.androirc.com ))
[02:52:34] <PMT> The times aren't particularly demonstrative of where the fires may ensue.
[02:54:14] <bunder> did you ever use 0.7.8?
[02:54:25] <djdunn> i think so
[02:55:01] <bunder> hrm, on second look, that never caused checksum errors
[02:55:13] <PMT> bunder: neither 0.7.7 nor 0.7.10 should have caused that
[02:55:17] <djdunn> its not widespread checksum errors
[02:55:47] <djdunn> i upgraded kde since i upgraded kernel to 4.19 and zol to 0.7.12
[02:56:40] <bunder> can you pastebin a zpool get all? i wonder what features you are using
[02:56:52] <PMT> djdunn: the next time it happens, before rebooting, it'd be useful if you had zpool events -v output from it. you could start zpool events -vf in a screen and it'll sit there and print additional messages to screen as they arrive. (This would be useful b/c zpool events does not persist across reboots and has a finite buffer, so e.g. regular snapshots would cause it to eventually no longer contain the
[02:56:58] <PMT> entries we want to see.)
[02:57:56] <djdunn> zpool get all https://paste.pound-python.org/show/4ZAuunrgwTXDnH4t8P8v/
[02:58:13] <PMT> djdunn: what CPU and board, specifically?
[02:59:02] <djdunn> uhh
[02:59:03] *** aither <aither!aither@im.vpsfree.cz> has quit IRC (Ping timeout: 268 seconds)
[02:59:11] <djdunn> amd fx 8350
[02:59:51] <djdunn> i caught a cksum error on zpool events -vf
[03:00:07] *** ghormoon <ghormoon!~ghormoon@ghorland.net> has quit IRC (Ping timeout: 240 seconds)
[03:00:42] <PMT> that'd be useful to pastebin, if you could
[03:00:47] <djdunn> i dont know how to zpool events -vf with wgetpaste
[03:01:34] <PMT> I mean, for example if it's running in a screen, you could use ^A [ to change to copy mode, highlight the relevant text, and then paste it into just about anything
[03:01:45] <PMT> (using, IIRC, ^A ] to paste)
[03:01:49] <djdunn> ah yeah
[03:02:55] <djdunn> it formatted the whitespace badly in vim...
[03:03:18] <ptx0> lol vim
[03:03:29] <PMT> djdunn: I mean, you could have just pasted it into cat > filename
[03:03:32] *** mikky <mikky!~mikky@grinch.ghoul.cz> has quit IRC (Ping timeout: 250 seconds)
[03:04:57] <bunder> hrm, all features, i'm not sure if that changes anything though
[03:05:09] <djdunn> or pipe it
[03:05:11] <djdunn> https://paste.pound-python.org/show/RgO20Mn5SCF9th6tWZMI/
[03:05:23] <djdunn> thats zpool events
[03:05:55] *** ghormoon <ghormoon!~ghormoon@ghorland.net> has joined #zfsonlinux
[03:06:45] <djdunn> 12 checksum errors since i booted up
[03:07:18] <djdunn> ill go back to zfs 0.7.11 and 4.18...
[03:07:22] <bunder> reading
[03:07:33] *** djdunn <djdunn!~djdunn@fl-184-2-23-179.dhcp.embarqhsd.net> has quit IRC (Quit: Lost terminal)
[03:07:38] <PMT> ...dammit
[03:07:46] <cirdan> ?
[03:07:46] <PMT> oh well he can go have fun with his broken shit
[03:08:56] *** djdunn-n8 <djdunn-n8!~djdunn@fl-184-2-23-179.dhcp.embarqhsd.net> has joined #zfsonlinux
[03:09:02] <PMT> (I was going to say moving back to 4.18 might help but I don't think moving to 0.7.11 will.)
[03:09:33] <djdunn-n8> Well thats whats in my initramfs for 4.18
[03:10:28] <PMT> I mean, I could certainly be wrong, I don't have much reason to think 0.7.11 would hurt, I just think 4.18/19 is more likely to be related, assuming my suspicions are wrong.
[03:10:36] <PMT> djdunn-n8: do you have edac-utils installed?
[03:10:43] <PMT> or mcelog?
[03:10:59] <djdunn-n8> Mcelog doesnt work with my chipset
[03:11:01] <ptx0> 4.18 and 4.19 are trash
[03:11:06] <ptx0> ime
[03:11:10] <ptx0> 4.19.5 is fixed
[03:11:19] <PMT> djdunn-n8: and what makes you think you'd see ECC errors then
[03:11:25] *** djdunn <djdunn!~djdunn@fl-184-2-23-179.dhcp.embarqhsd.net> has joined #zfsonlinux
[03:11:40] <ptx0> djdunn-n8: you are going to be auto banned by the bot for all the rejoins
[03:11:46] <ptx0> and it makes it difficult to help you
[03:11:53] <ptx0> please stop
[03:12:05] <PMT> I'm not sure he can, since I think he's typing from the system he's rebooting.
[03:12:12] *** snajpa <snajpa!snajpa@im.vpsfree.cz> has quit IRC (Ping timeout: 252 seconds)
[03:12:17] <ptx0> then he will be banned
[03:12:24] <ptx0> 21:11:52 -zfs(~zfs@unaffiliated/ptx0)- djdunn has 6 strikes
[03:12:38] <djdunn-n8> This is my phone thats my box
[03:13:01] <ptx0> ok well 4.14.84 is fixed too
[03:13:20] <ptx0> some patch made life hell for a lot of users incl zfs
[03:13:43] <bunder> 4.19.0 works fine for me
[03:13:53] <cirdan> what's broken ptx0
[03:13:55] <djdunn-n8> On 4.19.5
[03:13:56] *** ghormoon <ghormoon!~ghormoon@ghorland.net> has quit IRC (Ping timeout: 250 seconds)
[03:14:50] <fling> What is the script name for deleting older snapshots?
[03:14:51] <bunder> this might sound dumb but do you have any non ecc memory
[03:15:32] <djdunn> not for this slot
[03:16:24] <PMT> djdunn: my current hypothesis is that you have a hardware problem, since it seems like it's not writing the same checksum error to multiple places (or you'd be getting unrecoverable failures on noticing it), and you already said you have no way of knowing if the ECC has issues
[03:16:48] <djdunn> well it does report ecc errors
[03:16:54] <djdunn> just not with that utility
[03:17:06] <djdunn> its a kernel module that spits the errors out into dmesg
[03:17:18] <bunder> (the worst that can happen if you try is it goes bonkers, so have backups :) )
[03:17:54] <PMT> djdunn: are you trying to convey that you have examples of it spitting out ECC errors, or just that you believe it will log ECC errors without mcelog?
[03:19:53] <djdunn> oh yeah mcelog doesnt work but the kernel module spits errors out as the module says it does
[03:20:56] <PMT> djdunn: would you care to share any information about what module you think does this, or what evidence you have that it works
[03:21:18] <cirdan> my mcelog doesn't work, always says Cannot open `/dev/mcelog': No such file or directory even though it's there
[03:21:18] <djdunn> edac_amd64
[03:21:35] <DeHackEd> cirdan: may indicate the hardware isn't present...
[03:22:14] <cirdan> DeHackEd: i have a xeon with ecc /shrug
[03:23:13] <PMT> djdunn: i believe you should understand I'm really close to telling you "good luck bye", because you keep half-answering my questions and then correcting me with parts you didn't answer
[03:24:10] <djdunn> mcelog: ERROR: AMD Processor family 21: mcelog does not support this processor. Please use the edac_mce_amd module instead.
[03:24:29] <PMT> djdunn: sure, so mcelog doesn't work. I'm asking you what evidence you have that edac-mce-amd or whatever else does.
[03:24:39] *** snajpa <snajpa!snajpa@im.vpsfree.cz> has joined #zfsonlinux
[03:24:43] <djdunn> so i use the edac_mce_amd module that puts the logs in dmesg
[03:26:06] *** ghormoon <ghormoon!~ghormoon@ghorland.net> has joined #zfsonlinux
[03:26:57] <djdunn> edac-utils works but not mcelog
[03:27:16] *** mikky <mikky!~mikky@2a01:430:17:1::39> has joined #zfsonlinux
[03:27:19] <PMT> i give up.
[03:27:29] <djdunn> mcelog doesnt support amd cpu's
[03:27:51] <PMT> you seem to keep trying to answer question A, when I've asked you question B multiple times.
[03:28:31] <djdunn> like i said it logs edac logs in /var/log/messages
[03:28:38] <djdunn> but i havnt seen one in years
[03:28:47] <djdunn> but i know they go there
[03:29:25] <PMT> i am attempting to establish whether ECC is supported or enabled on your computer at all. You claim it is, but all the evidence I can find says that depends on the motherboard, and you ignored multiple questions about what motherboard it is.
[03:29:47] <PMT> So I asked you what evidence you have ECC was working, and you kept replying that mcelog doesn't work and edac-utils does, which doesn't answer why you think it works on this computer.
[03:29:51] <PMT> Have fun, I'm done.
[03:31:22] <djdunn> asus m5a99fx pro r2.0, ecc is enabled it works i know it does
[03:33:09] <djdunn> but no checksum errors on 4.18 with zfs 0.7.11
[03:38:08] *** delx <delx!~delx@> has quit IRC (Remote host closed the connection)
[03:40:17] *** delx <delx!~delx@> has joined #zfsonlinux
[03:54:56] *** djdunn <djdunn!~djdunn@fl-184-2-23-179.dhcp.embarqhsd.net> has quit IRC (Quit: Lost terminal)
[04:00:27] *** djdunn <djdunn!~djdunn@fl-184-2-23-179.dhcp.embarqhsd.net> has joined #zfsonlinux
[04:00:30] *** zfs sets mode: +b *!*@fl-184-2-23-179.dhcp.embarqhsd.net$#zfsonlinux-quarantine
[04:05:35] <ptx0> vOv
[04:05:56] <ptx0> djdunn-n8: use the zfs-discuss mailing list is my suggestion
[04:06:08] <ptx0> better there because you don't have to worry about disconnections
[04:07:01] *** djdunn-n8 <djdunn-n8!~djdunn@fl-184-2-23-179.dhcp.embarqhsd.net> has quit IRC (Quit: AndroIRC - Android IRC Client ( http://www.androirc.com ))
[04:21:06] *** ghormoon <ghormoon!~ghormoon@ghorland.net> has quit IRC (Ping timeout: 250 seconds)
[04:22:13] *** mikky <mikky!~mikky@2a01:430:17:1::39> has quit IRC (Ping timeout: 268 seconds)
[04:23:50] *** mikky <mikky!~mikky@grinch.ghoul.cz> has joined #zfsonlinux
[04:27:45] *** snajpa <snajpa!snajpa@im.vpsfree.cz> has quit IRC (Ping timeout: 252 seconds)
[04:31:12] *** snajpa <snajpa!snajpa@im.vpsfree.cz> has joined #zfsonlinux
[04:48:38] <bunder> huh
[04:48:45] <bunder> gentoo had nero (the cd burner)
[04:49:09] <bunder> why did i use brasero etc all these years lol
[04:52:52] *** ss23 is now known as Snowman23
[05:01:37] <DeHackEd> on linux I thought cdrecord was the gold standard
[05:02:36] <cirdan> thre were a few good ones
[05:02:40] <cirdan> cdrdao too
[05:13:12] *** mmlb <mmlb!~mmlb@76-248-148-178.lightspeed.miamfl.sbcglobal.net> has joined #zfsonlinux
[05:39:28] *** djdunn <djdunn!~djdunn@fl-184-2-23-179.dhcp.embarqhsd.net> has quit IRC (Quit: Lost terminal)
[05:40:09] *** ReimuHakurei <ReimuHakurei!~Reimu@raphi.vserver.alexingram.net> has quit IRC (Ping timeout: 252 seconds)
[05:58:57] *** neffi <neffi!~neffi@2001:19f0:5:4894:5400:1ff:feb9:c566> has quit IRC (Quit: neffi)
[05:59:54] *** neffi <neffi!~neffi@2001:19f0:5:4894:5400:1ff:feb9:c566> has joined #zfsonlinux
[06:00:09] *** ReimuHakurei <ReimuHakurei!~Reimu@raphi.vserver.alexingram.net> has joined #zfsonlinux
[06:05:57] *** gerhard7 <gerhard7!~gerhard7@ip5657ee30.direct-adsl.nl> has joined #zfsonlinux
[06:06:27] *** rosco_ <rosco_!~rosco@eslpc23.epfl.ch> has joined #zfsonlinux
[06:08:59] *** ReimuHakurei <ReimuHakurei!~Reimu@raphi.vserver.alexingram.net> has quit IRC (Ping timeout: 268 seconds)
[06:10:46] *** _rosco <_rosco!~rosco@eslpc23.epfl.ch> has quit IRC (Ping timeout: 260 seconds)
[06:27:00] *** ReimuHakurei <ReimuHakurei!~Reimu@raphi.vserver.alexingram.net> has joined #zfsonlinux
[06:56:53] <zfs> [zfsonlinux/zfs] Move strlcat, strlcpy, and strnlen (8005ca4) comment by v <https://github.com/zfsonlinux/zfs/issues/>
[07:00:01] *** mmlb <mmlb!~mmlb@76-248-148-178.lightspeed.miamfl.sbcglobal.net> has quit IRC (Read error: Connection reset by peer)
[07:00:58] *** mmlb <mmlb!~mmlb@76-248-148-178.lightspeed.miamfl.sbcglobal.net> has joined #zfsonlinux
[07:01:38] <zfs> [zfsonlinux/zfs] Fix small sysfs leak (7a23c81) comment by v <https://github.com/zfsonlinux/zfs/issues/>
[07:19:38] *** Zialus <Zialus!~RMF@> has quit IRC (Ping timeout: 245 seconds)
[07:25:08] *** Zialus <Zialus!~RMF@> has joined #zfsonlinux
[07:25:33] <PMT> DeHackEd: cdrecord had a bunch of controversy over the author's "look i swear you can license some code as GPL2 and some code as CDDL and redistribute the binaries from it" where literally everyone else said "no you can't"
[07:26:29] <PMT> https://en.wikipedia.org/wiki/Cdrtools
[07:29:13] <PMT> Debian/Red Hat/Ubuntu all think this is incorrect and stopped carrying his code. So even if you would like to argue that Debian is ideologically motivated, I doubt you could make the same argument readily for RH
[08:15:09] *** rlaager <rlaager!~rlaager@grape.coderich.net> has quit IRC (Ping timeout: 252 seconds)
[08:34:18] *** rlaager <rlaager!~rlaager@grape.coderich.net> has joined #zfsonlinux
[08:35:28] *** aither <aither!aither@im.vpsfree.cz> has joined #zfsonlinux
[08:42:12] *** prologic <prologic!~prologic@unaffiliated/prologic> has joined #zfsonlinux
[09:00:02] *** mmlb <mmlb!~mmlb@76-248-148-178.lightspeed.miamfl.sbcglobal.net> has quit IRC (Read error: Connection reset by peer)
[09:00:50] *** mmlb <mmlb!~mmlb@76-248-148-178.lightspeed.miamfl.sbcglobal.net> has joined #zfsonlinux
[09:48:51] *** stemid <stemid!~avatar@i-lands.anark.ist> has joined #zfsonlinux
[09:49:42] <stemid> zfs on fedora 27 hasn't survived a reboot yet for me. I haven't touched the systemd part of the package, I think all of it is vanilla. but everytime my server reboots I have to go through this whole process to get the zfs module loaded again. I just upgraded to 28 and same deal. what am I missing?
[09:50:04] <stemid> I did obviously enable the services installed by the package.
[09:50:29] <stemid> I've been using it since f26, with this issue. I rarely reboot. and everytime I reboot I'm in here whining :D
[09:51:33] <PMT> stemid: okay, so what does journalctl -xn say?
[09:52:06] <PMT> or systemctl --state failed perhaps would be more useful
[09:54:17] <stemid> just waiting for it to reboot again.
[09:54:55] <stemid> but from memory, from the past troubleshootings I've done with this, the zfs module fails to find kernel headers because I have dnf headless updates running.
[09:55:03] <PMT> stemid: ideally, please don't change anything from when it boots, so we can see what's broken
[09:55:06] <stemid> so the kernel is different from when it was built.
[09:55:22] <stemid> from when the zfs module was built.
[09:58:24] <PMT> stemid: I'm finding a bunch of conflicting information about what can cause that issue, so I'm gonna just wait and ask you to show me what it is you see before trying to suggest what to do about it.
[10:00:00] *** veegee <veegee!~veegee@ipagstaticip-3d3f7614-22f3-5b69-be13-7ab4b2c585d9.sdsl.bell.ca> has joined #zfsonlinux
[10:12:57] <stemid> PMT: https://bpaste.net/show/d02bf1074f3c
[10:13:05] <stemid> I think that indicates that the kernel module is built for a past kernel version.
[10:13:11] <stemid> and needs to be rebuilt for the current version.
[10:13:28] <stemid> which is what I've had to do everytime. since I run dnf upgrade periodically, each reboot results in a new kernel version for me.
[10:13:34] <stemid> so maybe I should be rebuilding zfs before I reboot.
[10:13:45] <stemid> or disable automatic updates
[10:16:40] <stemid> https://github.com/zfsonlinux/zfs/wiki/RHEL-and-CentOS#kabi-tracking-kmod I've missed this part of the docs actually.
[10:19:16] <stemid> but the kmod repos aren't listed for fedora. maybe not supported.
[10:21:20] *** nehemiah <nehemiah!~nehemiah@hs-user-138.wia.cz> has joined #zfsonlinux
[10:24:45] <stemid> trying a rebuild now.
[10:35:03] <stemid> yeah that worked. I guess I'll just disable automatic updates and make sure to rebuild at each new kernel version. finally understand what is happening at least and how to fix it.
[10:44:22] <PMT> stemid: you could just cron trying to detect it and rebuild
[11:00:01] *** mmlb <mmlb!~mmlb@76-248-148-178.lightspeed.miamfl.sbcglobal.net> has quit IRC (Read error: Connection reset by peer)
[11:00:52] *** mmlb <mmlb!~mmlb@76-248-148-178.lightspeed.miamfl.sbcglobal.net> has joined #zfsonlinux
[11:05:12] <zfs> [zfsonlinux/zfs] Move strlcat, strlcpy, and strnlen (8005ca4) comment by Sebastian Gottschall <https://github.com/zfsonlinux/zfs/issues/#commitcomment-31522729>
[11:06:17] <zfs> [zfsonlinux/zfs] commit 8005ca4 "Move strlcat, strlcpy, and strnlen" is not working as it suposed to be (#8157) comment by Sebastian Gottschall <https://github.com/zfsonlinux/zfs/issues/8157#issuecomment-443415036>
[11:08:58] *** IonTau <IonTau!~IonTau@124-171-141-231.dyn.iinet.net.au> has joined #zfsonlinux
[11:21:13] *** ntd <ntd!~ntd@gateway/tor-sasl/ntd> has joined #zfsonlinux
[11:25:39] *** ntd <ntd!~ntd@gateway/tor-sasl/ntd> has quit IRC (Client Quit)
[11:29:02] *** LeoTh3o <LeoTh3o!~LeoTh3o@phoxden.net> has quit IRC (Ping timeout: 268 seconds)
[11:38:29] *** LeoTh3o <LeoTh3o!~LeoTh3o@phoxden.net> has joined #zfsonlinux
[12:13:25] *** ntd <ntd!~ntd@gateway/tor-sasl/ntd> has joined #zfsonlinux
[12:29:50] *** apekatten <apekatten!~apekatten@unaffiliated/apekatten> has quit IRC (Ping timeout: 246 seconds)
[12:31:25] *** IonTau <IonTau!~IonTau@124-171-141-231.dyn.iinet.net.au> has quit IRC (Remote host closed the connection)
[12:38:09] *** javi404 <javi404!~quassel@unaffiliated/javi404> has joined #zfsonlinux
[12:47:38] *** LeoTh3o <LeoTh3o!~LeoTh3o@phoxden.net> has quit IRC (Quit: [BAD ERROR])
[12:51:21] *** LeoTh3o <LeoTh3o!~LeoTh3o@phoxden.net> has joined #zfsonlinux
[12:59:05] <hyper_ch> question: I created a zvol and installed a Windows VM on there using. The file system used in the vm is ntfs. Is there a way to mount it in linux?
[12:59:52] <ntd> the zvol will contain more than just the ntfs partition
[13:00:01] *** mmlb <mmlb!~mmlb@76-248-148-178.lightspeed.miamfl.sbcglobal.net> has quit IRC (Read error: Connection reset by peer)
[13:00:15] <ntd> so find the offset, mount as loopback i suppose
[13:00:44] <ntd> and for gawds sake have vm powered down when doing it
[13:00:46] <blackflow> why loopback. just mount it as if it was /dev/sda
[13:00:56] *** mmlb <mmlb!~mmlb@76-248-148-178.lightspeed.miamfl.sbcglobal.net> has joined #zfsonlinux
[13:00:56] <blackflow> or whatever other block device
[13:01:37] <blackflow> though it's been a while since I needed NTFS, I think you need the ntfs-3g utility, and can't use mount -t ntfs, not sure tho'
[13:02:14] <ntd> the zvol will also hold the MBR/GPT7etc
[13:02:56] <blackflow> yes... so treat is as any otehr block device
[13:04:53] <blackflow> oh wait, yes, silly me, kernel won't see gpt partitions off of it like a sata drive would....
[13:05:08] <pink_mist> blackflow: that only works if udev actually makes device notes for the partitions
[13:05:14] <hyper_ch> I did run fdisk -l on the "partitions" crated https://paste.simplylinux.ch/view/raw/afccae59
[13:05:39] <hyper_ch> I fail to see what there's part1 and part2 and both parts have 4 partitions
[13:05:49] <blackflow> pink_mist: yeah, forgot about that part
[13:05:58] <pink_mist> which is where the loopback comes in
[13:06:14] <pink_mist> *nodes
[13:06:14] <blackflow> there was a trick with qemu-nbd if I'm not mistaken, to mount image _files_. perhaps the same could be used here as well
[13:06:32] <hyper_ch> I think part1 is the ESP partition created while part2 is the actual nfts partition
[13:06:43] <hyper_ch> blackflow: I'll google that, thx
[13:07:07] <blackflow> hyper_ch: https://www.jamescoyle.net/tag/qemu-nbd
[13:07:27] <hyper_ch> blackflow: so zvol is also qcow2?
[13:07:57] <blackflow> hyper_ch: no I think it should be raw
[13:08:06] <hyper_ch> yeah, right :)
[13:09:01] *** chesty <chesty!~chesty@li1449-118.members.linode.com> has quit IRC (Quit: the ting go skrra)
[13:10:34] *** chesty <chesty!~chesty@li1449-118.members.linode.com> has joined #zfsonlinux
[13:12:16] <zfs> [zfsonlinux/zfs] Fix ASSERT in zfs_receive_one() (#8121) new review comment by LOLi <https://github.com/zfsonlinux/zfs/pull/8121#discussion_r238061609>
[13:12:45] *** ntd <ntd!~ntd@gateway/tor-sasl/ntd> has quit IRC (Quit: ntd)
[13:18:53] <blackflow> hyper_ch: reading that article.... I'm very curious of only running `partx -l /dev/zvol/...` would make the kernel see partitions off of it. if you treated it as raw device for qemu, it just might, so no nbd or loopback trickery might be needed
[13:19:59] <pink_mist> oh yeah, partx ... that might do it
[13:20:02] <hyper_ch> partx -l does print out two partitions if I run in on zvol and not zvol-part1/2
[13:21:27] <hyper_ch> https://paste.simplylinux.ch/view/raw/19f88485
[13:21:34] <blackflow> not sure, but partitions should be come devices under /dev/disk/by-...
[13:24:04] <hyper_ch> partx -a WinTest-part2 --> partx: /dev/zd16: error adding partition 2
[13:26:30] <hyper_ch> things can be so simple :)
[13:26:35] <blackflow> hyper_ch: looking at the manpage, I suppose all you need is -l or as the first example shows, partx --show - /dev/zvol/... (or what's the path to the zvol device you use)
[13:26:52] *** mquin is now known as ISO3103
[13:26:55] <hyper_ch> losetup /dev/loop0 /dev/tank/path/to/VMs/TestWin
[13:27:02] <blackflow> hyper_ch: and then I'd look at /dev/disk/by-.... for individual partitions
[13:27:11] <hyper_ch> kpartx -a /dev/loop0
[13:27:53] <hyper_ch> mount /dev/mapper/loop0pX /path/to/mount
[13:27:59] <blackflow> did that work?
[13:28:11] <hyper_ch> root@subi:/tmp/2# ls
[13:28:13] <hyper_ch> bootmgr 'Documents and Settings' pagefile.sys ProgramData 'Program Files (x86)' '$Recycle.Bin' 'System Volume Information' Windows
[13:28:14] <hyper_ch> BOOTNXT 'Dokumente und Einstellungen' PerfLogs 'Program Files' Programme swapfile.sys Users
[13:28:37] <blackflow> excellent! :)
[13:28:51] <hyper_ch> it's so easy - once you find out how
[13:29:07] <hyper_ch> so, you can just run loosetup on the whole zvol
[13:29:13] <blackflow> imho that loopback should be redundant (should be able to read partitions off of the zvol block device directly)
[13:29:14] <hyper_ch> use kpartx to create mappers
[13:29:27] <hyper_ch> and then just mount the mappers
[13:29:47] <hyper_ch> you mean like /dev/zvol16?
[13:29:52] <hyper_ch> or zd16 or whatever?
[13:29:54] <blackflow> yes
[13:30:02] <hyper_ch> it probably can... I just haven't tried it
[13:30:14] <blackflow> (which I mentioned first thing ;) )
[13:30:51] <blackflow> I mean when you give zvol as a (raw) device to qemu, it's treating it as if it would any other block device, there's no extra magic layer
[13:31:50] <hyper_ch> I'll try it now
[13:32:59] <blackflow> that's how I shared a GPT partition between qemu on linux, and that same root fs under a VirtualBox on a windows host. wasn't zvol, but it was a block device used as raw disk for virt, and two different hypervisors to bood (vbox and qemu-kvm)
[13:33:10] <blackflow> *to boot
[13:34:23] <hyper_ch> also seems to work
[13:34:34] <hyper_ch> losetup /dev/loop6 /dev/zd16
[13:34:40] <hyper_ch> kpartx -a /dev/loop6
[13:34:51] <hyper_ch> mount /dev/mapper/loop6p2 /tmp/xxx
[13:35:02] <blackflow> how about no losetup at all..... can you use kpartx on /dev/zvol/.... directly?
[13:36:09] <hyper_ch> let's try
[13:36:45] <bunder> sorry just waking up, what are you trying to do? use a zvol for a qemu vm?
[13:36:58] <hyper_ch> bunder: I alreardy did that
[13:37:06] <hyper_ch> I then wondered how to access that zvol from the host
[13:37:33] <blackflow> bunder: tl;dr windows VM on a zvol, wants to mount the NTFS partition off of it, on the host side
[13:37:45] <bunder> ah
[13:38:16] <hyper_ch> blackflow: you can run kpartx directly on the zvol
[13:38:34] <blackflow> very nice, then. :)
[13:38:41] <hyper_ch> https://paste.simplylinux.ch/view/raw/b71bae76
[13:38:53] <hyper_ch> the only thing I keep stumbling over is that kpartx wants full path
[13:38:59] <hyper_ch> otherwise it segfaults
[13:39:08] <hyper_ch> kpartx -a /dev/zd16
[13:39:18] <blackflow> tehy always do, it's not geom that can take dev names only ;)
[13:39:21] <blackflow> *they
[13:39:27] *** gerhard7 <gerhard7!~gerhard7@ip5657ee30.direct-adsl.nl> has quit IRC (Quit: Leaving)
[13:39:36] <bunder> with volmode set to default you should have a /dev/ node for the partition inside the zvol, you should just be able to mount that
[13:39:40] <blackflow> (they = linux tools)
[13:39:48] <hyper_ch> I thought when I'm in the /dev folder and just say xxx that it understands /dev/xxxx
[13:40:11] <blackflow> bunder: I suggested that at first but forgot that windows will partition it, qemu treating the zvol as a raw disk
[13:40:23] *** Albori <Albori!~Albori@216-229-78-154.fidnet.com> has quit IRC (Ping timeout: 268 seconds)
[13:40:36] <blackflow> so the kernel needs to re-scan for (gpt) partitions off of the zvol, hence kpartx
[13:41:13] <hyper_ch> well, in /dev I do hvae zd16, zd16p1, zd16p2
[13:42:02] <DeHackEd> if you just want the kernel to rescan partition tables, blockdev --rereadpt
[13:42:07] <bunder> https://pastebin.com/raw/Hua33izb
[13:42:08] <DeHackEd> assuming the device is idle
[13:43:19] <bunder> yes the p1/p2 are the inner partitions
[13:43:30] <bunder> zd16 is the outer zvol
[13:43:41] <blackflow> bunder: ah very nice, zfs does it automatically.
[13:44:27] <bunder> i usually set volmode to dev so you don't see them, but yeah
[13:44:52] <bunder> you just need to know which partition is which
[13:48:46] <DeHackEd> it should speed up importing a bit since the block devices don't need to be read for their partition tables either.
[13:49:58] <bunder> also linux's ntfs support is read only iirc
[13:50:40] <bunder> i also recall something about fat32 not liking huge files
[13:55:18] <blackflow> bunder: ntfs3g can read-write, but it's FUSE, so...
[13:56:08] <blackflow> thought the in-kernel driver _can_ write as well, not sure about features
[14:00:24] <bunder> don't believe so, lemme check again
[14:00:33] <DeHackEd> it's safe to overwrite a large file, so some implementations installed linux into a huge binary blob on NTFS to eliminate the need to repartition
[14:00:47] <DeHackEd> it does have general write support, but it's not considered nearly as good and safe as ntfs-3g
[14:01:34] <bunder> yeah it has write support but its overwrite only as long as the length changes, no new files, no deletes
[14:02:04] <bunder> so its 90% useless heh
[14:02:49] <bunder> s/length changes/doesn't change
[14:02:51] *** Turl <Turl!~Turl@cyanogenmod/maintainer/Turl> has quit IRC (Ping timeout: 252 seconds)
[14:02:56] <bunder> gah more coffee please
[14:02:56] *** cyteen <cyteen!~cyteen@> has joined #zfsonlinux
[14:03:35] <blackflow> bunder: the in-kernel driver? maybe but the FUSE thing works on the fs normally, as in you can add new files, delete files, modify files, ...
[14:04:09] <blackflow> pretty sure of it, at least it was the case last time I used it a year or so ago. shared partition with windoze on a multi-boot system
[14:05:23] *** Albori <Albori!~Albori@216-229-78-154.fidnet.com> has joined #zfsonlinux
[14:06:54] <bunder> i've never tried the fuse one
[14:07:19] <bunder> tbh i only have fuse on my laptop for that appimage binary thingy
[14:08:06] <blackflow> ntfs3g lib is FUSE. default on pretty much all linux distros for ntfs ops
[14:08:38] <ptx0> ntfs3g has insane cpu overhead
[14:08:41] <ptx0> it is slow as hell
[14:10:03] <bunder> doesn't all fuse
[14:22:08] <ptx0> not necessarily
[14:22:45] <bunder> well, maybe not "slow as hell" but "slow" because userland
[14:30:21] *** Turl <Turl!~Turl@cyanogenmod/maintainer/Turl> has joined #zfsonlinux
[15:00:01] *** mmlb <mmlb!~mmlb@76-248-148-178.lightspeed.miamfl.sbcglobal.net> has quit IRC (Read error: Connection reset by peer)
[15:00:57] <ptx0> this is cool
[15:01:02] *** mmlb <mmlb!~mmlb@76-248-148-178.lightspeed.miamfl.sbcglobal.net> has joined #zfsonlinux
[15:01:07] <ptx0> when my zfs recv finishes it sits there slowly climbing in cpu use
[15:01:17] <ptx0> sent a 500M dataset, now zfs recv is just doing *something*
[15:01:29] <ptx0> root 31125 40.5 0.0 251456 4088 ? R 08:48 5:51 zfs recv -u -F -o canmount=off newrpool/backup/millennium/olympic
[15:01:38] <ptx0> root 31125 41.2 0.0 251456 4088 ? R 08:48 6:01 zfs recv -u -F -o canmount=off newrpool/backup/millennium/olympic
[15:01:45] <ptx0> it's climbing at .1% CPU per second
[15:01:47] <bunder> its not finished if its still running :P
[15:01:49] <ptx0> what the hell could that be
[15:02:00] <ptx0> bunder: the sending side finished writing all data to the recv side
[15:02:23] <ptx0> this has been sitting here for almost 5 minutes doing, what exactly
[15:02:34] <ptx0> and finally moved on
[15:03:18] <bunder> maybe the data wound up in arc or something and its just purging to disk
[15:06:30] <bunder> then again 5 minutes is a little long
[15:07:18] <bunder> does recv re read the stream when its done to make sure its actually done?
[15:09:02] <bunder> (not that it can ask for send to resend missing blocks)
[15:10:49] <ptx0> i doubt it
[15:11:20] <bunder> time to bust out strace i guess :P
[15:11:32] <ptx0> it is a kernel thread
[15:11:49] <ptx0> the userspace tracing is not helpful
[15:12:00] <hyper_ch> gazillions of snapshots with no actual change in data?
[15:12:11] <ptx0> no
[15:12:51] <bunder> shouldn't the userspace still be doing something since its consuming cpu time?
[15:13:04] <ptx0> yeah it's reading from HPET
[15:13:08] <ptx0> not very helpful
[15:13:19] <hyper_ch> what's hpet?
[15:13:29] <ptx0> google it
[15:13:38] <hyper_ch> google hates me :(
[15:13:44] <ptx0> use bing
[15:13:55] <hyper_ch> bing doesn't return anything useful
[15:14:06] <hyper_ch> (no duckduckgo?)
[15:14:17] <ptx0> then you are out of luck
[15:14:37] <hyper_ch> but why would you suggest bing instead of duckduckgo?
[15:15:24] <ptx0> all that time wasted when you could have just looked up what it is
[15:16:02] <hyper_ch> I don't consider social interaction as a waste of time :)
[15:16:33] <ptx0> don't know if you haven't figured it out yet but this is a support channel
[15:16:56] <hyper_ch> one doesn't exclude the other
[15:17:09] *** ChanServ sets mode: +o ptx0
[15:17:15] *** ptx0 sets mode: +q hyper_ch!*@*
[15:17:20] <ptx0> please stfu
[15:17:43] *** ptx0 sets mode: -o ptx0
[15:19:20] <ptx0> this system has a shitty broken TSC so that's why it uses HPET
[15:19:36] <ptx0> but i'm not sure that's the reason
[15:22:14] <bunder> spl does seem to make use of timers
[15:22:54] <bunder> and txg's
[15:26:10] <ptx0> root 32334 0.2 0.0 251588 4316 ? S 09:03 0:03 zfs recv -u -F -o canmount=off newrpool/backup/millennium/olympic/Documents
[15:26:16] <ptx0> low cpu use while it's actually running
[15:26:52] <bunder> if you're up at 5am, ask me to check my server
[15:27:11] <ptx0> i've finally got my sleep schedule almost back to normal but i wake around 7am now
[15:27:20] <bunder> heh
[15:27:45] <bunder> i don't get that luxury, rotating shifts
[15:34:56] <fling> I have a permanent error in a snapshot. How do I identify in which one and how to recover?
[15:35:27] <fling> I have all the snapshot backups.
[15:36:00] *** gerhard7 <gerhard7!~gerhard7@ip5657ee30.direct-adsl.nl> has joined #zfsonlinux
[15:36:24] <bunder> doesn't it say, or is it giving you one of those 0xjibberish numbers?
[15:37:18] <fling> Yeah numbers.
[15:37:42] <fling> Should I clone an earlier snapshot and send the backups back?
[15:37:47] <fling> Still need to identify which one
[15:38:58] <ptx0> it's probably in multiple snapshots
[15:39:13] <ptx0> just create bookmarks and destroy all of the snapshots?
[15:40:13] <bunder> iirc finding which snap is hard, i think the last time it came up it was easier to toss them all
[15:40:46] <ptx0> yep
[15:40:47] <bunder> because you can toss the first time the block is in a snap, but then the snap after that is holding the bad block
[15:41:11] <bunder> oh you just said that lel
[15:42:18] <ptx0> Microsoft: "Open source developers are communists" also Microsoft: "we love open source, we changed the paradigm"
[15:43:00] <ptx0> as if being a commie is a bad thing <che.png>
[15:43:14] <bunder> i blame bill
[15:43:21] <ptx0> bill ballmer?
[15:43:24] <bunder> satya seems to see the writing on the wall
[15:45:58] *** ghormoon <ghormoon!~ghormoon@ghorland.net> has joined #zfsonlinux
[16:04:16] <zfs> [zfsonlinux/zfs] External Fragmentation leading to Out of memory Condition (#4953) comment by Igor Podlesny <https://github.com/zfsonlinux/zfs/issues/4953#issuecomment-443432484>
[16:05:36] <bunder> lol nope
[16:12:14] *** elxa <elxa!~elxa@2a01:5c0:e09b:5891:883a:8d79:2511:e3c1> has joined #zfsonlinux
[16:18:43] <fling> So what to do exactly to fix the situation? :P
[16:19:19] <ptx0> just told you
[16:19:22] <ptx0> 09:39:37 < ptx0> just create bookmarks and destroy all of the snapshots?
[16:23:06] <fling> ptx0: what to do with bookmarks after that?
[16:23:45] <bunder> i'm not even sure why you would want bookmarks in this situations since you can restore from backups but /shrug
[16:24:02] <ptx0> convenience
[16:24:09] <ptx0> not throwing the baby with the bathwater etc
[16:28:25] <fling> so I clone from an older snapshot, then I send everything back from the backup, then I send everything new to the clone, then I promote, then destroy the filesystem and rename the clone?
[16:28:50] *** ntd <ntd!~ntd@gateway/tor-sasl/ntd> has joined #zfsonlinux
[16:29:27] <ptx0> ..............
[16:29:35] <ptx0> wtf
[16:29:41] <ptx0> why
[16:30:00] <ptx0> dear god, just destroy the snapshots
[16:30:23] <ptx0> cloning and promoting and all that does absolutely nothing
[16:31:29] <fling> But it will restore my snapshots
[16:32:01] <ptx0> the corrupt snapshots
[16:32:36] <fling> umm
[16:32:43] <fling> I have the intact snapshots in backup
[16:43:56] *** cyteen <cyteen!~cyteen@> has quit IRC (Ping timeout: 246 seconds)
[17:00:01] *** mmlb <mmlb!~mmlb@76-248-148-178.lightspeed.miamfl.sbcglobal.net> has quit IRC (Read error: Connection reset by peer)
[17:00:58] *** mmlb <mmlb!~mmlb@76-248-148-178.lightspeed.miamfl.sbcglobal.net> has joined #zfsonlinux
[17:16:35] *** miRR <miRR!~miRR@> has joined #zfsonlinux
[17:18:00] *** cirdan <cirdan!~cirdan@2601:85:4400:6fd3:f64d:30ff:fe60:210d> has quit IRC (Ping timeout: 252 seconds)
[17:56:11] *** cirdan <cirdan!~cirdan@c-69-141-186-131.hsd1.nj.comcast.net> has joined #zfsonlinux
[17:58:02] *** ntd <ntd!~ntd@gateway/tor-sasl/ntd> has quit IRC (Quit: ntd)
[18:05:29] *** f_g <f_g!~f_g@213-47-131-124.cable.dynamic.surfer.at> has quit IRC (Ping timeout: 246 seconds)
[18:18:45] *** f_g <f_g!~f_g@213-47-131-124.cable.dynamic.surfer.at> has joined #zfsonlinux
[18:51:21] <ptx0> whew, 80fps in ashes of the singularity now
[18:51:30] <ptx0> with those qemu patches
[18:52:21] <zfs> [openzfs/openzfs] 6037 zfs(1M) needs to handle unknown uid/gid in context of allow/unallow more gracefully (#690) comment by Yuri Pankov <https://github.com/openzfs/openzfs/issues/690#issuecomment-443444683>
[18:56:17] <fling> How to deal with this error when send/recv migrating a pool? -> local origin for clone pool/fs@snap does not exist
[18:56:39] <fling> Got the error in the middle of a migration and mbuffer: error: outputThread: error writing to <stdout> at offset 0xa0ece3800: Broken pipe
[19:00:01] *** mmlb <mmlb!~mmlb@76-248-148-178.lightspeed.miamfl.sbcglobal.net> has quit IRC (Read error: Connection reset by peer)
[19:00:48] *** mmlb <mmlb!~mmlb@76-248-148-178.lightspeed.miamfl.sbcglobal.net> has joined #zfsonlinux
[19:00:52] *** zfs sets mode: +b *!*@76-248-148-178.lightspeed.miamfl.sbcglobal.net$#zfsonlinux-quarantine
[19:08:22] <Shinigami-Sama> bunder: the writing on the wall? hes the one that Sacked the rest of their QA and moved to nearly test-free rolling release system :\, and shoved Bing into desktop searches
[19:12:16] <zfs> [zfsonlinux/zfs] Cannot mount received raw encrypted filesystem (#8103) comment by herrernst <https://github.com/zfsonlinux/zfs/issues/8103#issuecomment-443446186>
[19:18:54] *** k1l <k1l!~k1l@ubuntu/member/k1l> has left #zfsonlinux ("kkthxbb")
[19:24:16] <zfs> [zfsonlinux/zfs] zfs seems to use more memory than it should (#5035) comment by "Gabriel A. Devenyi" <https://github.com/zfsonlinux/zfs/issues/5035#issuecomment-443447039>
[19:30:11] *** ksa <ksa!~m@> has quit IRC (Ping timeout: 246 seconds)
[19:30:14] <fling> How do I replicate the whole pool properly? I don't understand why are clones preventhing this.
[19:34:58] *** DeHackEd <DeHackEd!~dehacked@atmaweapon.dehacked.net> has quit IRC (Quit: Reboot for PCI device reassignment)
[19:36:38] <bunder> Shinigami-Sama: i mean opensourcing more of their stuff, and allowing WSL
[19:37:07] <Shinigami-Sama> oh, thats just him outsourcing yet more QA, it has nothing to do with open source
[19:38:50] <bunder> just because they deleted your mydocs folder doesn't mean they don't test things
[19:39:09] *** darkmeson <darkmeson!~darkmeson@gateway/tor-sasl/darkmeson> has quit IRC (Ping timeout: 256 seconds)
[19:40:20] <Shinigami-Sama> you've not been supporting ~350 companys with 2016 servers and win10 desktops have you?
[19:40:55] <Shinigami-Sama> I bricked a fresh install of 2016 just last week by doing the standard, install from iso, click update
[19:41:02] <zfs> [zfsonlinux/zfs] Cannot mount received raw encrypted filesystem (#8103) comment by Tom Caputi <https://github.com/zfsonlinux/zfs/issues/8103#issuecomment-443448253>
[19:44:34] <bunder> we're still using windows 7 and god knows how many versions of winserver
[19:44:53] *** darkmeson <darkmeson!~darkmeson@gateway/tor-sasl/darkmeson> has joined #zfsonlinux
[19:48:26] *** cirdan <cirdan!~cirdan@c-69-141-186-131.hsd1.nj.comcast.net> has quit IRC (Ping timeout: 250 seconds)
[19:49:20] <fling> Shinigami-Sama: what is update?
[19:51:32] <Shinigami-Sama> fling: who knows, one of the random CUs that take 4 hours to install on afresh system and consumes a singular core for no appearnt reason
[19:51:45] <Shinigami-Sama> I just flattened it and started again and it just worked
[19:52:01] <Shinigami-Sama> even though the process was exactly the same...
[19:53:35] <fling> I promoted a clone and now I'm able to send/recv the part of the pool yay!
[19:53:41] <fling> Will now try to send the whole pool.
[19:53:53] *** cirdan <cirdan!~cirdan@c-69-141-186-131.hsd1.nj.comcast.net> has joined #zfsonlinux
[20:09:28] <fling> Need to promote more clones…
[20:14:47] <bunder> oh intel... their new video card is called arctic sound
[20:17:41] <daedalus^> poor volt-
[20:25:08] *** mmlb <mmlb!~mmlb@76-248-148-178.lightspeed.miamfl.sbcglobal.net> has quit IRC (Ping timeout: 246 seconds)
[20:31:44] *** Essadon <Essadon!~Essadon@81-225-32-185-no249.tbcn.telia.com> has joined #zfsonlinux
[20:32:07] *** Essadon <Essadon!~Essadon@81-225-32-185-no249.tbcn.telia.com> has quit IRC (Max SendQ exceeded)
[20:32:31] *** Essadon <Essadon!~Essadon@81-225-32-185-no249.tbcn.telia.com> has joined #zfsonlinux
[20:37:08] <zfs> [zfsonlinux/zfs] zfs seems to use more memory than it should (#5035) comment by petermaloney <https://github.com/zfsonlinux/zfs/issues/5035#issuecomment-443452580>
[20:46:32] <xedniv> the new sysfs code in the zfs module seems problematic with constify
[20:47:18] <bunder> hrm
[20:47:23] <xedniv> so far ive written a patch that moves some of the obviously static sysfs_ops structs into const static declarations and then reference them by pointer, keeping the sysfs_ops const and the compiler happy
[20:47:44] <bunder> what setting in the kernel does constify
[20:48:12] <xedniv> basically you should not pass under any circumstances a fptr to adjust a sysfs_ops structure on runtime. ever. regardless of constify.
[20:48:21] <xedniv> it's in the general settings under gcc plugins
[20:48:34] <bunder> ah, i haven't looked into those, seen them in there for a while though
[20:48:42] <xedniv> the proper way to do this is to pass a whole sysfs truct, already instantiated
[20:48:46] <xedniv> struct*
[20:49:00] <bunder> if you have a patch, you should file a pr ;)
[20:49:15] <xedniv> ideally defined as const in the original module/location
[20:49:25] <xedniv> i do, but i need someone who knows that particular code to clarify something
[20:49:54] <bunder> eh, throw it up there anyways, i'm sure they can give you feedback
[20:51:00] *** ntd <ntd!~ntd@gateway/tor-sasl/ntd> has joined #zfsonlinux
[20:51:02] <xedniv> in zprop_iter_common we have a call from zfs_sysfs where the properties are processed. inside the kobj there is the showfunc fptr which gets passed around. it seems there are no other sysfs ops defined anywhere but those in zfs_sysfs
[20:51:38] <xedniv> so my current patch should cover everything. i need to confirm. then all that is left is figuring out which one of the sysfs_ops needs to be set
[20:51:56] <xedniv> for the zprop_iter_common case
[20:52:43] <bunder> ah that's a little more complex than i thought
[20:52:48] <xedniv> the others are obvious as they essentially pass the static funcs to the init function, i broke the behavior into two, one that respects proper kernel "rules" and passes the sysfs_ops, referencing the predefined ones, whichever is provided by caller
[20:54:29] <xedniv> not tested: http://paste.ubuntu.com/p/7R4m2tschk/
[20:54:32] <bunder> not sure how many people are here who have that intimate of a knowledge of the code, the mailing list might also help there
[20:54:40] <PMT> fling: I don't know what flags you're using, but it sounds like it tried to send a cloned dataset without sending the parent first
[20:55:01] <xedniv> basically no legitimate users of zfs_kobj_init_dynamic_showfunc should exist
[20:55:24] <zfs> [zfsonlinux/zfs] zfs seems to use more memory than it should (#5035) comment by "Gabriel A. Devenyi" <https://github.com/zfsonlinux/zfs/issues/5035#issuecomment-443453780>
[20:55:26] <xedniv> so i just need to confirm which one of the static sysfs_ops apply to the current caller of zfs_kobj_init_dynamic_showfunc (which is the zprop iter function)
[20:57:18] <xedniv> in any case, whoever wrote the sysfs stuff doesnt seem to understand the implications. even if this compiled, say, without constify, and with an old toolchain, it might cause runtime problems. let alone being a security risk as writable fptrs are a prime candidate for abuse.
[20:57:42] <bunder> if its the recent sysfs stuff, lol brady bunch
[20:58:18] *** TheBloke <TheBloke!~TomJ@unaffiliated/tomj> has quit IRC (Read error: No route to host)
[20:58:59] <bunder> https://github.com/zfsonlinux/zfs/commit/e8bcb693d6a19962291e4429f2dc4ce60693bef7 ?
[20:59:12] <xedniv> some of these bugs are likely obscured by the fact that zfs runs as LKM
[20:59:19] <xedniv> yeah the recent sysfs stuff
[20:59:30] <bunder> oh god it is
[21:00:00] <xedniv> yeah but that particular commit does not contain the kernel side of things
[21:00:11] <xedniv> check module/zfs/zfs_sysfs.c
[21:01:13] *** TheBloke <TheBloke!~TomJ@unaffiliated/tomj> has joined #zfsonlinux
[21:06:04] <bunder> https://github.com/zfsonlinux/zfs/commit/7a23c81342df05ace730bd303b4a73854dba43dd#commitcomment-31522146
[21:06:41] <bunder> looks like someone else found a problem there to
[21:06:44] <bunder> too
[21:06:52] *** miRR <miRR!~miRR@> has quit IRC (Ping timeout: 250 seconds)
[21:08:31] <zfs> [zfsonlinux/zfs] Cannot mount received raw encrypted filesystem (#8103) comment by herrernst <https://github.com/zfsonlinux/zfs/issues/8103#issuecomment-443454656>
[21:09:11] <xedniv> bunder, seems like it's from context, easy fix
[21:09:15] <xedniv> im writing the rest of the patch
[21:09:23] <xedniv> that means the other crap can go
[21:09:58] <xedniv> bunder, that would be yours truly
[21:10:01] <xedniv> haha
[21:10:08] <bunder> oh okay lol
[21:10:08] <xedniv> im going to rely on you for future issue scouting.
[21:10:27] <xedniv> lets see if this fixes it
[21:16:46] <zfs> [zfsonlinux/zfs] Cannot mount received raw encrypted filesystem (#8103) comment by Tom Caputi <https://github.com/zfsonlinux/zfs/issues/8103#issuecomment-443455234>
[21:17:00] <xedniv> bunder, http://paste.ubuntu.com/p/prrDP6Kv4z/
[21:17:00] *** Stern <Stern!~Sternisae@a83-162-246-136.adsl.xs4all.nl> has quit IRC (Read error: Connection reset by peer)
[21:17:01] <xedniv> fixed
[21:17:06] <xedniv> compiles cleanly
[21:17:15] *** DeHackEd <DeHackEd!~dehacked@> has joined #zfsonlinux
[21:17:22] <xedniv> if someone can test on sacrificial data/pools, i'd appreciate that
[21:17:27] *** Stern <Stern!~Sternisae@a83-162-246-136.adsl.xs4all.nl> has joined #zfsonlinux
[21:18:46] <bunder> when you open a pr it gets run against a bunch of test bots, but i guess they don't use gcc plugins
[21:20:20] <zfs> [zfsonlinux/zfs] Fix small sysfs leak (7a23c81) comment by v <https://github.com/zfsonlinux/zfs/issues/>
[21:20:51] <xedniv> bunder, atm im swamped, cant go thru the hipster devops XXI century cycle
[21:20:55] <xedniv> might do that tomorrow or such
[21:21:03] <xedniv> i left a comment in the original issue
[21:21:26] <bunder> no worries its the weekend anyways
[21:21:31] <xedniv> exactly
[21:22:00] <xedniv> i miss the days when people used some cranky patchutils-based shell script kit and mail lists to do stuff. instead of ticketing our way to death/
[21:22:02] *** stemid <stemid!~avatar@i-lands.anark.ist> has left #zfsonlinux
[21:22:37] <xedniv> so far it hasnt kept people from writing shit code and screwing up, it seems, so clearly we dont need some more kubernetes/docker/insert-fancy-devops-marketing-mumbo-jumbo/travis/etc.
[21:22:44] <xedniv> alas, too late!
[21:23:08] <xedniv> module builds fine
[21:24:14] <bunder> actually its all custom code that uses aws vm's along with the test suite
[21:24:17] <xedniv> insmodding, count one for valhalla.
[21:24:43] <xedniv> that sort of automation is useful, im not against that. just think we dont need more abstractions for a sign off/review process.
[21:24:54] <bunder> afaik no travis here, codecov but that's more informational
[21:25:06] <bunder> and 99% wrong lol
[21:25:36] <CompanionCube> xedniv: needs more javascript
[21:26:03] <xedniv> lolol
[21:26:32] <xedniv> the answer to all tech problems since 2017 (or so) is "needs more javascript", definitely.
[21:26:58] <CompanionCube> if it isn't enough
[21:27:03] <CompanionCube> sprinkle in some blockchain and AI/ML
[21:27:10] <xedniv> you forgot Go
[21:27:17] <xedniv> and if you need some more edge, rust.
[21:27:25] <CompanionCube> eh, go isn't to the same degree
[21:27:44] <bunder> how do you get positive code coverage from a man page https://github.com/zfsonlinux/zfs/pull/8134#issuecomment-439518978
[21:27:49] <xedniv> go is actually decent, i just dislike the build framework/approach but you have to understand the issue it tackles
[21:27:53] <xedniv> (dependency hell)
[21:28:15] <xedniv> rofl
[21:28:22] <xedniv> that made my day, almost
[21:28:30] <CompanionCube> bunder: nondeterminism in the tests?
[21:28:35] <xedniv> there goes proof the world truly is going down the crapper.
[21:28:54] * xedniv waits for the ZoL code of conduct so he has to go back to using reiserfs
[21:29:35] <bunder> only if you go to the conferences
[21:30:04] <xedniv> you will get hunted down for saying you cant grope people online
[21:30:27] <xedniv> i mean, dont you feel all that "sending love" in social media? well, there's also "sending gropes"...
[21:30:40] <pink_mist> bunder: well, if the man page is now shorter there'd be less bytes in the entire distribution, so coverage could potentially go up
[21:30:59] <bunder> that's not code though :P
[21:31:21] <xedniv> this is hilarious but offtopic. let's move to -social before we get someone upset
[21:31:48] <xedniv> like, "ill make a fluid website with your name saying you are a rapist" level of upset. /end of the bill burr jokes.
[21:32:07] <pink_mist> eh, man-pages can be construed as code since it's not pure text
[21:32:14] <pink_mist> it's a stretch though
[21:32:31] <xedniv> if it was tex you could actually have a real point
[21:32:49] <xedniv> any outstanding issues i should know of with the 8 rc2 for native crypto?
[21:33:24] <xedniv> also, why are we using ancient code without optimizations for the hashing and such? unless im looking at the wrong tree, it's lacking sse/avx & co optimizations
[21:34:57] <DeHackEd> there's SIMD for a bunch of algorithms
[21:36:08] <DeHackEd> check the module/zcommon directory for checksum stuff
[21:38:44] <xedniv> ack
[21:38:47] <xedniv> gotta reboot
[21:38:56] <xedniv> tested two trees, building without issues
[21:42:11] <storrgie> I'm doing my first install of ZFS on debian stretch, I'm wondering if there is a particular reason that out of the box these services wouldn't be able to start (https://gist.githubusercontent.com/storrgie/eaea159026f575a5c5a46fd77230528d/raw/c3ed69c181808acb1673dce32b1c9efae2f9d336/gistfile1.txt)
[21:42:26] *** ntd <ntd!~ntd@gateway/tor-sasl/ntd> has quit IRC (Quit: ntd)
[21:48:55] *** eckomute <eckomute!~eckomute@d66-183-83-212.bchsia.telus.net> has joined #zfsonlinux
[21:49:57] *** _rosco <_rosco!~rosco@eslpc23.epfl.ch> has joined #zfsonlinux
[21:51:56] *** mf <mf!~mindfly@reactos/developer/mf> has quit IRC (Ping timeout: 246 seconds)
[21:52:17] *** rosco_ <rosco_!~rosco@eslpc23.epfl.ch> has quit IRC (Ping timeout: 246 seconds)
[22:02:30] *** miRR <miRR!~miRR@> has joined #zfsonlinux
[22:03:44] *** devster31 <devster31!~devster@ns348877.ip-91-121-108.eu> has quit IRC (Read error: Connection reset by peer)
[22:07:25] *** devster31 <devster31!~devster@ns348877.ip-91-121-108.eu> has joined #zfsonlinux
[22:12:09] *** oGG <oGG!~ogg@h-34-198.A182.priv.bahnhof.se> has joined #zfsonlinux
[22:14:37] <zfs> [zfsonlinux/zfs] zfs seems to use more memory than it should (#5035) comment by kpande <https://github.com/zfsonlinux/zfs/issues/5035#issuecomment-443459022>
[22:20:30] <storrgie> apparently this is going on --> https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=864348
[22:22:21] <ngharo> storrgie: is the kernel module loaded?
[22:23:14] <xedniv> back
[22:23:17] <xedniv> bunder, works
[22:24:09] *** gerhard7 <gerhard7!~gerhard7@ip5657ee30.direct-adsl.nl> has quit IRC (Quit: Leaving)
[22:24:36] *** JanC_ <JanC_!~janc@lugwv/member/JanC> has joined #zfsonlinux
[22:24:56] <ngharo> storrgie: add zfs to /etc/modules.conf easy fix
[22:25:00] *** JanC <JanC!~janc@lugwv/member/JanC> has quit IRC (Ping timeout: 252 seconds)
[22:25:45] <ngharo> s/.conf//
[22:25:51] *** JanC_ <JanC_!~janc@lugwv/member/JanC> has quit IRC (Remote host closed the connection)
[22:26:09] *** JanC_ <JanC_!~janc@lugwv/member/JanC> has joined #zfsonlinux
[22:27:49] *** JanC_ is now known as JanC
[22:31:31] *** IonTau <IonTau!~IonTau@124-171-141-231.dyn.iinet.net.au> has joined #zfsonlinux
[22:34:38] *** thuttu77 <thuttu77!~thuttu77@82-181-37-124.bb.dnainternet.fi> has quit IRC (Ping timeout: 246 seconds)
[22:35:56] *** thuttu77 <thuttu77!~thuttu77@82-181-37-124.bb.dnainternet.fi> has joined #zfsonlinux
[22:36:51] *** thuttu77 <thuttu77!~thuttu77@82-181-37-124.bb.dnainternet.fi> has quit IRC (Remote host closed the connection)
[22:40:51] <zfs> [openzfs/openzfs] 6037 zfs(1M) needs to handle unknown uid/gid in context of allow/unallow more gracefully (#690) comment by Allan Jude <https://github.com/openzfs/openzfs/issues/690>
[23:07:36] *** asjackson <asjackson!~user@xavier.cs.nott.ac.uk> has joined #zfsonlinux
[23:29:58] *** nehemiah <nehemiah!~nehemiah@hs-user-138.wia.cz> has quit IRC (Remote host closed the connection)
[23:38:46] *** Greyztar <Greyztar!~irc@gateway/tor-sasl/greyztar> has joined #zfsonlinux
[23:43:45] *** Greyztar <Greyztar!~irc@gateway/tor-sasl/greyztar> has quit IRC (Quit: "")
[23:58:11] *** Greyztar <Greyztar!~irc@gateway/tor-sasl/greyztar> has joined #zfsonlinux
   December 1, 2018  
< | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | 31 | >