| onefang | Anyone got any clues why using "rsync --partial" doesn't actually resume interrupted transfers? You may have noticed my power dropped out a little while ago. The really big transfer that was half complete, started again from the beginning after I restarted it. It left the half complete temporary file, but created a new one to start again. | 10:10 |
|---|---|---|
| onefang | sync -a --stats --progress --partial --safe-links --one-file-system /source /destination | 10:11 |
| onefang | Oops, with an r at the start. lol | 10:11 |
| onefang | Hmmm, maybe that only works if it's rsync at both ends? I'm rsyncing a network mount, so only rsync at the receiving end. | 10:54 |
| CueXXIII | ah, if you have only one rsync then it must rescan the whole source file again, since it does not store continuation information | 10:59 |
| onefang | I'll remember that for next time. This multiday file transfer is down to the last two files. | 11:00 |
| CueXXIII | also rsync might keep the old partial files around till the file is transfered completely, not sure how it is implemented | 11:01 |
| CueXXIII | do you have ssh access to the network mount? so you could run 2 rsync processes locally to each file system | 11:02 |
| onefang | The network mount IS via ssh. B-) | 11:02 |
| onefang | Meh, not worth it for now. This is a one time thing that is almost done. I'll remember to do it that way next time. | 11:03 |
| CueXXIII | rsync --some-options server:/path/to/fileordir /path/to/fileordir ## automatically defaults to ssh to login to server and start rsync there | 11:04 |
| onefang | Cool. Though in this case I was giving sshfs a go. | 11:05 |
| CueXXIII | sshfs is nice but for rsync :) | 11:07 |
| onefang | Well I could come up with various ways it could work fine without rsync at both ends, guess the rsync devs didn't. | 11:09 |
| onefang | But you basically confirmed my guesses. Thanks. | 11:11 |
| cruncher | hi all, im a long time debian (sid) user, but im really getting ***** by systemd. My question is, how far behind are the pacakages compared to debian (sid)? | 13:14 |
| onefang | Our package mirrors redirect most packages to Debian package mirrors, and only supply the packages we change. So how far behind mostly depends on how long it takes our systems to deal with package updates. Blink and you'll miss it. | 13:16 |
| cruncher | onefang, and what about those you change? i dont suppose there is a list of the changed packages? | 13:22 |
| onefang | I think there is, someone else might know where we keep it. I should be sleeping soonish. | 13:25 |
| CueXXIII | cruncher: grep-available -sPackage -FVersion devuan | sort -u => https://www.oetec.com/pastebin/plain/b4XfoSS4 | 14:21 |
| Xenguy | CueXXIII, Unfortunately he's gone | 14:22 |
| Xenguy | Missed my answer by minutes, but all too common these days | 14:22 |
| CueXXIII | ah… he can read the backlog | 14:22 |
| Xenguy | True, the channel is logged | 14:23 |
| Xenguy | It would be interesting to know how common that actually is : -) | 14:24 |
| CueXXIII | but i switched to devuan shortly before the time_t transition and did not have any noticeable lags in the forked packages | 14:26 |
| CueXXIII | the time_t transition was a bit finicky, but i guess so was it in plain debian, too | 14:26 |
| CueXXIII | oh, and i missed devuan native packages like eudev - use: grep-available -sPackage -FOrigin Devuan | sort -u | 14:30 |
| Xenguy | https://pkgmaster.devuan.org/bannedpackages.txt | 14:48 |
| freem | is there is any practical difference between eudev and systemd-udev? | 18:23 |
| freem | this one was the most annoying daemon to implement decently under runit supervision... but that is unrelated to systemd, I guess (https://p.mort.coffee/UUH): basically there's a need to signal it that it needs to do it's damn work | 18:25 |
| freem | (I didn't invented this, I picked it on some blog post somewhere on the interweb) | 18:26 |
| freem | I have regularly been thinking about trying to find an alternative to (systemd-|e)udev, but I'm not knowledgeable enough about initramfs and /dev population | 18:28 |
| gnarface | freem: they're not supposed to be that different, but sometimes they'll have different bugs | 18:30 |
| freem | notably for a server, especially virtual ones, automatic discovery of devices when one is plugged in seems... a bit too much :) | 18:30 |
| gnarface | well, you can always just use mknod to populate /dev by hand | 18:31 |
| freem | yes, but I never went as far enough as to find doc on how to actually do this | 18:31 |
| freem | tbh it was mostly a concern to me when I was a bit more in minimalist systems... I still prefer those, but I don't really try too hard to get them | 18:32 |
| gnarface | it's not complex, but i don't know myself where you get the major/minor numbers from in a generalized sense | 18:33 |
| gnarface | i've only had to use it occasionally on a one-off basis | 18:33 |
| freem | yes, that's the thing | 18:33 |
| freem | using a simple command is not complex, but knowing what to feed it can be | 18:34 |
| gnarface | i needed to switch some systems from udev to eudev after daedalus for some reason, can't remember specifically what, but it might be worth a try if you're having some issue with udev | 18:34 |
| gnarface | maybe you'll get lucky and it'll "just work" | 18:35 |
| freem | I have no issue with it AFAIK, except for the weird workaround in the script | 18:35 |
| gnarface | i don't use runit either, but some people around here do. stick around | 18:36 |
| freem | this hack to make udev commands run after the daemon actually started have this bad taste of silly design to me :) | 18:36 |
| freem | but it was like this before systemd anyway | 18:36 |
| freem | just curious, what are you using as init/supervision (if you use a supervision tool, ofc)? | 18:37 |
| gnarface | just the sysvinit defaults | 18:37 |
| freem | I thought devuan switch to openrc by default? | 18:38 |
| gnarface | not doing any weird daemon supervision stuff, i supervise them myself | 18:38 |
| gnarface | no, openrc has never been the default | 18:38 |
| freem | I was a dev not so long ago, so too lazy to do it myself :p | 18:38 |
| gnarface | openrc has been added as an option to the installer though | 18:38 |
| freem | I like the "write a 1-liner script and forget about it" thingy | 18:38 |
| gnarface | so you can install with it instead of having to switch after the fact | 18:39 |
| freem | I see. Well, I just use deboostrap when I need to install anyway. | 18:39 |
| freem | my script for install requires less (almost none) user intervention in the middle of the install process, which is something that annoys me with debian-installer | 18:40 |
| freem | perhaps it's none, actually, I don't remember, long time since I used it | 18:41 |
| gnarface | the installer supports preseeding, but if you have a script already there's probably little point in exploring that | 18:41 |
| freem | I already tried preseed... it's quite complicated for not much imo | 18:41 |
| freem | I tried it when I tried to automate installs at work... and it's how I started to write my own script instead | 18:43 |
| freem | but I guess YMMV | 18:45 |
| freem | for example a beginner probably can't setup a diskless system and write such a script that easily | 18:46 |
| gnarface | yea, debconf states don't seem super well organized, it doesn't work great for everything, i was just mentioning it for the record | 18:47 |
| mazes_83 | hi | 23:26 |
| mazes_83 | is there a specific channel for raspberry pi ? I noticed they use eudev, this gentoo project is not more maintained, do they plan something ? | 23:27 |
| mazes_83 | I didn't even took time to watch at my PC's with devuan to look if it was same case | 23:28 |
| mazes_83 | :) | 23:28 |
| fsmithred | mazes_83, try #devuan-arm | 23:34 |
| mazes_83 | ok, ty, I'll still watch if it's same case on 'classic pc' | 23:35 |
| mazes_83 | same | 23:36 |
| clemens3 | i think eudev got some updates again lately | 23:38 |
| mazes_83 | ouch, I believe it began as a gentoo project, as gentoo dropped it recently, I thought it was dead. My bad | 23:42 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!