**** BEGIN LOGGING AT Tue Jul 11 03:00:05 2017
Jul 11 04:25:37 <timeless> Maxdamantus: so, in redhat, the version of a package could be incredibly patched to the point that the "version" isn't remotely accurate
Jul 11 04:30:36 <Maxdamantus> timeless: you mean the minor version? Sure, so the minor version should change, which means the minor version of the OS that includes it also needs to change (since a user in that OS might compile some software that won't work correctly against the previous minor version)
Jul 11 04:31:10 <timeless> Maxdamantus: in redhat, the x.y.z tends to be fixed
Jul 11 04:31:22 <timeless> and the patch version which is x.y.z-patch is what's updated
Jul 11 04:31:29 <timeless> and it isn't really related to anything
Jul 11 04:32:21 <Maxdamantus> Is x.y.z a version for a RH distribution, a RH package or a piece of software that is packaged by RH?
Jul 11 04:32:36 <timeless> the latter
Jul 11 04:32:48 <timeless> err, the last one
Jul 11 04:36:27 * timeless frowns
Jul 11 04:47:57 <timeless> grr. one system had /proc/sys/net/ipv4/ip_forward enabled via /etc/sysctl.conf -- the other didn't
Jul 11 05:51:11 * DocScrutinizer05 doesn't get it how there's not a single script that would easily port a system to a new iron, or compare two systems (same distro/version, to start with)
Jul 11 05:52:36 <DocScrutinizer05> seems like the distro upgrade scripts should basically be capable to do the job, but they're not made for that
Jul 11 05:53:43 <DocScrutinizer05> plus fragmented across a million pre/postinstall scripts in packages
Jul 11 08:02:58 <Maxdamantus> Mm .. seems that the gas fitter damaged the lines to two of our phone jacks.
Jul 11 08:03:25 * Maxdamantus was going to say that he wasn't very familiar with RH when it cut out.
Jul 11 08:08:14 <Wizzup> DocScrutinizer05: it can't cost me anything, but I will let him know
Jul 11 08:08:30 <Wizzup> I got it 22 more times since my last msg
Jul 11 08:26:47 <xes> DocScrutinizer05: what would do you suggest? an rsync to one new VM on .maemo.org?
Jul 11 08:52:50 <rm_work> oh hey timeless ;)
Jul 11 08:52:54 * rm_work waves
Jul 11 12:34:20 <DocScrutinizer05> xes: I don't know if it needs an own VM or just an own vserver
Jul 11 12:34:57 <DocScrutinizer05> it does some heavy indexing every once in a while
Jul 11 12:35:15 <DocScrutinizer05> so maybe a container of sorts is in order
Jul 11 12:59:43 <timeless> hi rm_work 
Jul 11 13:00:14 <timeless> DocScrutinizer05: yeah...
Jul 11 13:00:52 <timeless> ok, for my next trick, i need to convince mariadb (on devuan) to `start`
Jul 11 13:00:57 <timeless> anyone know how to debug it?
Jul 11 13:07:39 <timeless> grumble
Jul 11 13:07:45 <timeless> `mysqld --verbose` works fine
Jul 11 13:07:54 * timeless wonders why `service start mysql` wouldn't work
Jul 11 13:09:25 <inz> is there anything relephant in /var/log/daemon.log?
Jul 11 13:10:08 <timeless> these files were presumably created by the `mysqld` that worked https://www.irccloud.com/pastebin/kaoWEsXQ/
Jul 11 13:10:23 <timeless> there's no sign of anything else
Jul 11 13:11:22 <timeless> oh, interesting
Jul 11 13:12:17 <timeless> when i try `service mysql status` and i'm using `mysqld --verbose` the latter says: https://www.irccloud.com/pastebin/A6EXMbAB/
Jul 11 13:12:26 <timeless> (the former for some reason says it twice...)
Jul 11 13:14:54 <timeless> i don't think debian (devuan really) managed to properly set up the mariadb system https://www.irccloud.com/pastebin/2enwAdvx/
Jul 11 13:15:12 <timeless> (which is true, the postinstall script failed, and my attempts to coax it forward failed miserably)
Jul 11 13:17:44 <timeless> Use of uninitialized value $action in string eq at /usr/share/debconf/frontend line 27.
Jul 11 13:26:31 <timeless> oh. brother
Jul 11 13:26:39 <timeless> mariadb requires systemd?
Jul 11 13:29:28 <DocScrutinizer05> DAMN! now Poettering answered (not even totally moot) https://github.com/systemd/systemd/issues/6309#issuecomment-314392132 and I can't reply, to point out that syntax check should be limited to the cases where the user= value is NOT in the password database already
Jul 11 13:30:14 <DocScrutinizer05> rationale: systemd MUST allow to use existing usernames no matter how garbled they look "syntactically"
Jul 11 13:33:49 <DocScrutinizer05> fix: swap sequence of chacks, first getpwnam() or whatever, only THEN and in case getpwnam threw error NONEXIST do syntax check
Jul 11 13:36:05 <DocScrutinizer05> or even defer syntax check to the moment when systemd actually creates a user. (and even then it's arguable if systemd should use existing policy definitions like in `man useradd|less -j6 +'/On Debian, the only constraints are.*$' ` instead of introducing/enforcing own ones
Jul 11 13:37:56 <DocScrutinizer05> a decent design would do execve("useradd"...) instead of doing own syntax check
Jul 11 13:39:43 <DocScrutinizer05> one of the few cases where execve($some-unixtool) is the only correct way to implement stuff, since the system's policy is defined in that tool and no other way to find out about the policy
Jul 11 13:40:43 <DocScrutinizer05> after all sysadmin might have replaced useradd by a proprietary version that implements however complex individual policy
Jul 11 13:41:50 <DocScrutinizer05> and systemd sysuser.d MY ASS!!!
Jul 11 13:43:20 <DocScrutinizer05> I already start puking when I read about DynamicUser=1
Jul 11 14:07:49 <sicelo> my N900 still rebooting randomly :-(
Jul 11 14:15:47 <DocScrutinizer05> damn
Jul 11 14:15:56 <sicelo> not a kernel oops, at least because sp-oops-extract only has older oopses
Jul 11 14:16:47 <DocScrutinizer05> reflash stock to rule out hw issues
Jul 11 14:18:56 <DocScrutinizer05> did you try my suggestion to run htop via WLAN ssh?
Jul 11 14:20:26 <DocScrutinizer05> and maybe even configure N900 rsyslog to log via WLAN as well
Jul 11 14:21:02 <DocScrutinizer05> alternatively run dmesg in a loop, every second
Jul 11 14:21:41 <sicelo> it happens at realy odd moments, when not in use. unfortunately no access to wifi at all times
Jul 11 14:22:02 <sicelo> i really doubt it's a case of something hogging cpu
Jul 11 14:22:24 <DocScrutinizer05> you won't find out with guessing
Jul 11 14:23:36 <DocScrutinizer05> your only chance is to get some useful diagnostic prints
Jul 11 14:24:21 <DocScrutinizer05> or strip down system to bare bones stock, then add functions until it sarts failing again
Jul 11 14:24:23 <sicelo> yes, that what i'd like to ask .. what's a conclusive way to figure this out?
Jul 11 14:24:49 <sixwheeledbeast^> have you checked boot_reason?
Jul 11 14:24:50 <DocScrutinizer05> see above
Jul 11 14:25:00 <DocScrutinizer05> wd32
Jul 11 14:25:05 <DocScrutinizer05> iirc
Jul 11 14:25:07 <sixwheeledbeast^> ah k
Jul 11 14:25:38 <sicelo> yes still wd32_to
Jul 11 14:26:02 * sixwheeledbeast^ crawls back under the rock
Jul 11 14:26:31 <sicelo> lifeguard restarts unchanged from saturday (had at least 4 restarts since)
Jul 11 14:26:57 <sicelo> reflashing not an option for a while
Jul 11 14:27:01 <KotCzarny> run rescueos and see if it reboots too?
Jul 11 14:27:05 <KotCzarny> ;)
Jul 11 14:27:06 <DocScrutinizer05> no clear idea what *exactly* is wd32 watchdog, but I am pretty sure when it times out, something is blocking the kernel
Jul 11 14:28:20 <DocScrutinizer05> indeed when reflashing is no option and you can't do any decent persistent logging like suggested above, then rescueOS is a sane test
Jul 11 14:29:12 <sicelo> it's not, because reboot doesn't happen in 'reasonable' time. so i'd not have a phone for long times
Jul 11 14:29:21 <sicelo> will log to a file
Jul 11 14:30:01 <KotCzarny> also, check swapfile for badblocks if you use it
Jul 11 14:30:05 <DocScrutinizer05> the problem with logging is: when kernel gets blocked/frozen by whatever task then any logging on system is highly prone to not getting flushed to storage, so you NEED off system logging, via WLAN, or via USB or console
Jul 11 14:30:28 <sicelo> :-)
Jul 11 14:30:40 <sicelo> looks i have to live with the reboots then 
Jul 11 14:31:16 <KotCzarny> nah, if it's dying hw it will get worse :/
Jul 11 14:31:34 <DocScrutinizer05> you may try to reduce commit time to maybe 1 second (is 30s now iirc)
Jul 11 14:32:29 <DocScrutinizer05> redoing swap is a very good suggestion
Jul 11 14:32:56 <DocScrutinizer05> or... :-D swap swap to uSD :-D
Jul 11 14:33:41 <DocScrutinizer05> any defects or oddities in swap may cause unsolicited reboots easily
Jul 11 14:34:19 <DocScrutinizer05> since kernel actually freezes as long as kswapd is busy
Jul 11 14:34:41 <sixwheeledbeast^> I had wd32 with stupid things like tracker taking an age to index stuff.
Jul 11 14:34:54 <DocScrutinizer05> yes, same here
Jul 11 14:35:14 <DocScrutinizer05> at least I guess it was wd32
Jul 11 14:35:38 <sicelo> uGod forbid that it should be dying hardware ...  i don't have replacment
Jul 11 14:35:43 <DocScrutinizer05> or browser opening a terrible website that ate all memory incl swap
Jul 11 14:36:00 <DocScrutinizer05> device getting unresponsive for an hour or two, then finally rebooting
Jul 11 14:36:38 <KotCzarny> sicelo: check both emmc and usd for badblocks, readonly test should be enough for now
Jul 11 14:36:38 <DocScrutinizer05> apropos... what's with vote?
Jul 11 14:37:21 <DocScrutinizer05> could somebody with credentials finally hit the "calculate and publish" button on https://maemo.org/vote/results.php?election_id=48
Jul 11 14:37:26 <DocScrutinizer05> PLEASE!
Jul 11 14:37:35 <sixwheeledbeast^> :nod: Modrana downloading tiles or running a tracklog on default settings a few other weird wd32's.
Jul 11 14:38:16 <DocScrutinizer05> where's juiceme?
Jul 11 14:39:02 <DocScrutinizer05> xes_: warfare: could you maybe log in to the election machine and hit the final button to calculate and publish stuff?
Jul 11 14:40:01 <DocScrutinizer05> I know there's such button and it works/worked, since I've hit it at least 3 times
Jul 11 14:42:23 <DocScrutinizer05> can't be true maemo takes longer to evaluate/count an election than the USA ;-P
Jul 11 14:43:16 <DocScrutinizer05> particularly since it's literally just a button to press
Jul 11 14:44:44 <DocScrutinizer05> http://maemo.org/vote/admin.php
Jul 11 14:46:21 <DocScrutinizer05> warfare: I don't know who except juiceme has valid credentials, mine are gone
Jul 11 14:46:42 <warfare> DocScrutinizer05: I don't. 
Jul 11 14:47:06 <DocScrutinizer05> create a set for you in mysql?
Jul 11 14:47:50 <sixwheeledbeast^> Well CC meeting tonight anyway
Jul 11 14:48:44 <DocScrutinizer05> do we need a meeting of old vanishing CC to decide to press a button to evaluate and then publish election results?
Jul 11 14:49:49 <DocScrutinizer05> back when we pondered to make that automatically when election period expires. We decided against it to have a checkpoint where manual interference is possible when something went haywire
Jul 11 14:50:34 <DocScrutinizer05> didn't think the thing to go haywire is that checkpoint itself and the human interaction needed to continue the automated process
Jul 11 14:55:05 <DocScrutinizer05> warfare: you should have credentials anyway, to recover the whole system when all credentials but yours are lost
Jul 11 14:55:20 <DocScrutinizer05> the whole voting system
Jul 11 14:55:42 <sicelo> installed crash-reporter. hope it will help narrow down the issue
Jul 11 14:55:48 <sicelo> bug 6334
Jul 11 14:55:50 <povbot> 04Bug https://bugs.maemo.org/6334 random but frequent HW watchdog reboots (/proc/bootreason contains "32wd_to") when returning from off_mode
Jul 11 14:56:08 <DocScrutinizer05> hmm, sounds to the point, yes
Jul 11 14:57:41 <DocScrutinizer05> WTF?!?? bazillion comments
Jul 11 14:59:05 <sicelo> unfortuantely
Jul 11 15:02:08 <DocScrutinizer05> >> Ext3 is journaled, but it's not as robust as UBIFS. If you have really bad luck and/or huge number of unclean reboots (32wd_to reboot or sw_rst caused by oops; sw_rst from SW watchdog are fine, file systems are then synched) and/or remove the battery while there's uncommited data (commit interval = 1s), it may get corrupted.<<  https://bugs.maemo.org/show_bug.cgi?id=6334#c415
Jul 11 15:02:10 <povbot> 04Bug 6334: random but frequent HW watchdog reboots (/proc/bootreason contains "32wd_to") when returning from off_mode
Jul 11 15:03:16 <DocScrutinizer05> it's controversial if journaling makes *any* sense on flash storage
Jul 11 15:05:21 <DocScrutinizer05> sicelo: try swapping your swap, to start with!
Jul 11 15:07:38 <ff_> sicelo: I've had the same randomly reboot issue every one, two days. This was changed: in KP53, governor to conservative, SmartReflex off vdd1=0 vdd2=0, +5 to all voltage vals, 360MHz for DSP.
Jul 11 15:08:35 <DocScrutinizer05> >> They were the same - after connecting charger and turning on the device startup took slightly longer than usually and animation (Nokia logo, sound, connecting hands) was not displayed. Then crash reporter started telling that it has report about 32wd_to reboot. Situation does not occur when I turn off the phone using power button, nor after reboot.<< sounds like fsck acting up, not unheard of
Jul 11 15:09:04 <ff_> Now it's 38days uptime and everythink is hunky-dory. Maybe in future I'll check greater DSP freq but for now it's work.
Jul 11 15:09:10 <DocScrutinizer05> smartreflex is known to be unstable
Jul 11 15:09:44 <DocScrutinizer05> that's why it's disabled in stock, until somebody proves otherwise for the reason
Jul 11 15:11:31 <DocScrutinizer05> Nokia didn't throttle the N900 on purpose, just to make it slower than it could be. All tuning at clocks and smartreflex is considered to have negative impact on system stability
Jul 11 15:12:35 <DocScrutinizer05> possibly *reducing* clock below stock settings can help on devices that suffer wear out
Jul 11 15:13:18 <DocScrutinizer05> so much for the often claimed "not a single device ever suffered harm from overclocking so far"
Jul 11 15:14:04 <DocScrutinizer05> those who claim that simply never would accept overclocking as a reason for hardware wear and resulting system instability
Jul 11 15:14:34 <ff_> dealing with performace needs some adjust
Jul 11 15:14:45 <DocScrutinizer05> though I seem to recall sicelo never OCed
Jul 11 15:15:26 <sicelo> how can i check status on mine? never needed a fast n900 and therefore never looked at those things
Jul 11 15:16:07 <DocScrutinizer05> hmm, to start with, look a zzztop for the frequencies used. and then there's that KP-tuning tool
Jul 11 15:16:46 <DocScrutinizer05> though when you never touched that stuff, you should not have any OC
Jul 11 15:16:53 <sicelo> err, isn't it accessible in /sys or something?
Jul 11 15:17:23 <DocScrutinizer05> it is, but I wouldn't know the /sys or /proc files off top of my head
Jul 11 15:17:39 <sicelo> ok
Jul 11 15:18:04 <ff_> give here output of  kernel-config show
Jul 11 15:19:41 <DocScrutinizer05> sicelo: install reswap or similar tool and transfer your swap to uSD
Jul 11 15:20:00 <DocScrutinizer05> or would you like to do it manually?
Jul 11 15:20:23 <ff_> should be sth like http://paste.debian.net/plain/975920
Jul 11 15:21:13 <DocScrutinizer05> min frequency 600? really now?
Jul 11 15:22:07 <DocScrutinizer05> aside form my system not even having kernel-config command, I doubt this is a recommended setting I see there
Jul 11 15:23:25 <ff_> I don't know but it's work flawlessly for 38days and it has nice performace
Jul 11 15:23:36 <DocScrutinizer05> ~oc
Jul 11 15:23:37 <infobot> i heard oc is an Optical Carrier, An OC1 has 672 channels or 44.736 Mbit/s
Jul 11 15:23:46 <DocScrutinizer05> ~maemo-oc
Jul 11 15:23:53 <DocScrutinizer05> whatever
Jul 11 15:25:05 <ff_> what is optical carrier in maemo?
Jul 11 15:25:34 <DocScrutinizer05> ~overclocking
Jul 11 15:25:34 <infobot> "OK, listen up.  This is your CPU."  apt drops the CPU into a hot frying pan.  "This is your CPU on overclocking.  Any questions?"
Jul 11 15:26:08 <sicelo> yes i can do it manually (relocating swap)
Jul 11 15:26:27 <sicelo> anyway for now will see what crash reporter catches first
Jul 11 15:27:12 <DocScrutinizer05> won't catch anything I'm pretty sure
Jul 11 15:27:14 <ff_> ah, ok
Jul 11 15:30:09 <ff_> I'm watching out to not beyond over 40C
Jul 11 15:30:48 <DocScrutinizer05> pointless
Jul 11 15:31:07 <ff_> but that temp is from battery sensor I think, not CPU?
Jul 11 15:33:31 <DocScrutinizer05> aaah
Jul 11 15:33:35 <DocScrutinizer05> ~omap-oc
Jul 11 15:34:01 <infobot> omap-oc is probably http://mg.pov.lt/maemo-irclog/%23maemo.2010-08-01.log.html#t2010-08-01T22:16:05  read that!, or and this http://mg.pov.lt/maemo-irclog/%23maemo.2011-03-11.log.html#t2011-03-11T03:04:11, or http://electronics.stackexchange.com/a/156130
Jul 11 15:34:05 <DocScrutinizer05> damn, poor bot suffering DDoS
Jul 11 15:38:47 <DocScrutinizer05> ff_: how could we know which temperature you're looking at. There's three tenps: battery charging chip die temperature and NTC sensor battery temperature, and CPU die temperature though CPU may have hotspots that are not detectable from the generic die temperature sensor
Jul 11 15:39:46 <ff_> temp from bq27200.sh
Jul 11 15:40:08 <DocScrutinizer05> that's basically useless, die temp of bq27200 charging chip
Jul 11 15:40:39 <DocScrutinizer05> not that any other temperature was really more useful regarding OC
Jul 11 15:41:02 <DocScrutinizer05> see third link the bot provided above
Jul 11 15:41:37 <ff_> ok
Jul 11 15:43:02 <bencoh> shit, this kernel doesn't have /sys/class/thermal
Jul 11 15:43:04 <bencoh> *meh*
Jul 11 15:43:19 <DocScrutinizer05> forget it sicelo
Jul 11 15:44:01 <DocScrutinizer05> oh bencoh, sorry
Jul 11 15:44:03 <timeless> DocScrutinizer05: so... i'm back to my main problem:  i have a `pydio` in a `devuan`.
Jul 11 15:44:14 <timeless> i have mariadb installed (in case there's a use for it)
Jul 11 15:44:23 <timeless> pydio in theory is configured to use sqlite3, and there is a database file
Jul 11 15:44:34 <ff_> previously reboots occurs even all phone was cold 27degC
Jul 11 15:44:36 <sicelo> see you all later
Jul 11 15:48:18 <DocScrutinizer05> bencoh: what for would you need thermal?
Jul 11 15:49:02 <DocScrutinizer05> btw the OMAP die temp sensor is broken, SiErr :-/
Jul 11 15:49:12 <DocScrutinizer05> sporadic
Jul 11 15:50:25 <DocScrutinizer05> eventually it freezes the value
Jul 11 15:50:38 <DocScrutinizer05> iirc
Jul 11 15:51:12 <DocScrutinizer05> recover via power cycle
Jul 11 16:10:40 <ff_> this /sys/devices/platform/omap34xx_temp/temp1_input & /sys/devices/platform/omap34xx_temp/temp1_input_raw ?
Jul 11 16:11:27 <DocScrutinizer05> yes, prolly
Jul 11 16:20:29 <ff_> strange, first N900 shows temp1_input -40 & raw 0 and sec 30 & 53
Jul 11 16:26:03 <ff_> and as you said, frozen
Jul 11 18:23:37 <Vajb> how long it should take for "flash it all" script to do it's magic?
Jul 11 18:25:01 <Vajb> it's been like 15minutes in state "kernel:[1311239.551502] Disabling IRQ #18
Jul 11 18:25:48 <Vajb> device itself shows just a steady yellow light
Jul 11 18:33:20 <freemangordon> TheKit: yes, there are semi-REed Nokia PA modules: https://notabug.org/freemangordon/pulseaudio-nokia/src/fremantle
Jul 11 18:33:46 <freemangordon> TheKit: actually, everything but voice module is REed
Jul 11 18:34:18 <TheKit> would they be needed on different device?
Jul 11 18:34:44 <freemangordon> most probably, for VOIP
Jul 11 18:35:00 <freemangordon> though...
Jul 11 18:35:31 <freemangordon> maybe not as most if not all voip clients supports their own agc, aec, etc
Jul 11 18:36:09 <freemangordon> but, it still will be perfect if we have Nokia audio processing algos on our side :)
Jul 11 18:41:53 <DocScrutinizer05> OT (mlockall() for noswap) http://paste.ubuntu.com/25069876
Jul 11 18:43:25 <DocScrutinizer05> Vajb: flash-it-all.sh takes less than 2 minutes. Never seen anything like >>kernel:[1311239.551502] Disabling IRQ #18<<
Jul 11 18:44:05 <DocScrutinizer05> ((less than 2 minutes unless you need to download stuff via an ultraslow internet link)
Jul 11 18:44:12 <TheKit> is dialer provided by rtcom-call-ui?
Jul 11 18:45:29 <DocScrutinizer05> freemangordon: ((maybe not)) that's why I suggested replacing that stuff by a feed-trhough dummy
Jul 11 18:46:51 <DocScrutinizer05> limiting the impact to reduced audio quality, worst case
Jul 11 19:34:27 <freemangordon> ok, finally I hit libhildon-plugins-notify-sv :(. Going to RE it
Jul 11 20:32:14 <DocScrutinizer05> that bugger
Jul 11 20:32:31 <DocScrutinizer05> about time it gets REed
Jul 11 20:34:20 <freemangordon> yeah. will take time though, it is not small one (60k binary)
Jul 11 21:15:46 <DocScrutinizer05> pretty sure. Does a couple of things
Jul 11 21:16:54 <DocScrutinizer05> prolly from yellow ribbons to requesters to LED patterns (via dbus to mce) to audio jungles, right?
Jul 11 21:17:45 <DocScrutinizer05> jingles even
Jul 11 21:18:39 <DocScrutinizer05> I might be wrong though
**** ENDING LOGGING AT Wed Jul 12 03:00:03 2017