systemdlete | Why does runlevel tell me "N 2" when I think I am in multi-user (5?) level? | 02:05 |
---|---|---|
systemdlete | This is the situation immediately after boot. | 02:06 |
systemdlete | /var/run/runlevel does say 2 | 02:06 |
systemdlete | nvm. | 02:09 |
systemdlete | I keep forgetting this is rc, not sysvinit | 02:10 |
mason | systemdlete: Yeah, it's not quite as rigorously segmented as sysv. | 02:11 |
systemdlete | I'm having a long "hang" upon boot. I see cups and cups-browsed, then it hangs for a few minutes | 02:11 |
systemdlete | Finally, I see dovecot start. So I am guessing it is dovecot at fault here. | 02:12 |
sgage | my inittab shows default level is in fact 2 | 02:12 |
systemdlete | But starting dovecot from cmd line after login works instantly, so I am guessing it is dovecot | 02:12 |
systemdlete | sgage: ignore please | 02:12 |
systemdlete | it is correct; I was not | 02:13 |
sgage | AOK | 02:13 |
systemdlete | :) | 02:14 |
systemdlete | mason: Any idea about dovecot (or whatever is causing it to pause?) | 02:14 |
mason | systemdlete: Yeah, pop some echo lines into the scripts around there, and you can see exactly where you're hanging up. | 02:15 |
mason | start and end of each script, and you'll know right where you are | 02:15 |
systemdlete | For the moment, I have disabled dovecot, and I am rebooting to see if that's it | 02:15 |
systemdlete | Then I can debug dovecot or the like | 02:16 |
mason | systemdlete: Ah. Are you using ifupdown? If so, change allow-hotplug to auto. | 02:16 |
mason | systemdlete: I don't know why it defaults to allow-hotplug, but it's kind of bad. | 02:16 |
systemdlete | Oh... right. Maybe it is waiting on a wireless connection? | 02:17 |
mason | Or maybe it's something else, but that one bites me on just about every install. | 02:17 |
systemdlete | :) | 02:17 |
systemdlete | I mean, :( | 02:17 |
mason | Exactly. | 02:17 |
systemdlete | I confirm it hangs on dovecot, and probably for the reason you are saying. Now changing interfaces file... | 02:19 |
systemdlete | mason: My file is already set for 'auto' | 02:20 |
* systemdlete goes off to pollute dovecot startup with tons of traces... 8| | 02:21 | |
mason | Hm. Then yeah, instrument the start-up scripts and where you're hanging will pop right out. | 02:21 |
mason | Was it definitely dovecot? | 02:21 |
systemdlete | Just dovecot really | 02:21 |
mason | kk | 02:21 |
systemdlete | yep. | 02:21 |
systemdlete | rebooted with it disabled, and *snap!* it was back up instantly | 02:22 |
mason | so, if you've narrowed it down that far, you can replace dovecot with something that runs dovecot under strace, for example. | 02:22 |
systemdlete | dovecot is def the sticky wicket. | 02:22 |
mason | Like, move the actual binary and put in a shell script with the right name, where the shell script passed its arguments to an strace'd dovecot binary and saves the trace off to some pid-based name. | 02:23 |
systemdlete | set -xv first | 02:23 |
systemdlete | that will help zero in on which part of the script hangs, might not be dovecot itself. It does other setups | 02:24 |
systemdlete | It's hanging on doveconf, actually | 02:24 |
systemdlete | (that's why I wanted to do xv first) | 02:25 |
systemdlete | And then I can try it after login to see if I can replicate that behavior | 02:25 |
tuxd3v | hello guys, I need some help to configure my bluethoot headset | 02:39 |
tuxd3v | I installed blueman | 02:39 |
tuxd3v | I already paired the devices | 02:39 |
tuxd3v | for what I understand blueman needs dbus | 02:39 |
tuxd3v | I should have dbus running by now? | 02:40 |
tuxd3v | it fails to connect: Procoll not availlable??? | 02:41 |
tuxd3v | DBUS failed error | 02:42 |
tuxd3v | its what I get | 02:42 |
tuxd3v | blueman.bluez.errors.DBusFailedError:Protocol not availlable | 02:45 |
systemdlete | mason: Mystery solved (almost). strace of doveconf shows it hangs on call to getrandom(... in a VM! I recall there is a way to fix this, but I don't remember what | 03:25 |
rrq | haveged | 03:28 |
tuxd3v | its possible to use Alsa, in Mate Desktop instead of Pulse Audio? | 03:31 |
sgage | tuxd3v: that's what I'm doing right now. Got some nice music going on Pandora :-) | 03:55 |
sgage | You may have to install apulse to satisfy some programs, but no-pulseaudio works fine. | 03:56 |
systemdlete | rrq: I think that was it, but it looks like it has not started in time for dovecot. Dovecot is still doing its hangy thing on boot | 04:59 |
systemdlete | oh, wait | 05:00 |
systemdlete | looks like it wont start | 05:00 |
no_9 | I just installed ASCII, can't seem to install kaffeine or KDE... anyone know what's up with that? | 07:11 |
bpmedley | no_9 : Fresh install? | 07:12 |
no_9 | Yasss | 07:12 |
bpmedley | What error messages are you getting? | 07:13 |
no_9 | I can't install hex chat either, so I have to manually transfer | 07:14 |
no_9 | Package KDE standard is not available but is referred to by another package | 07:14 |
bpmedley | Any network issues? Can you ping a remote host like google.com? | 07:15 |
no_9 | I have no problem being online, I tried to get on here or via mibbit but apparently freenode doesn't like mibbit | 07:15 |
bpmedley | What's in /etc/apt/sources.list ? | 07:16 |
no_9 | I've already Run apt-get update, apt get upgrade, app gat get dist upgrade | 07:17 |
bpmedley | But, what's in the sources.list? | 07:17 |
no_9 | The sources list is what was on the web page for Devuan for ascii | 07:17 |
bpmedley | What's your apt-get install command? | 07:17 |
no_9 | Its amd64 btw | 07:18 |
no_9 | Apt-get install kde then i tried apt-get install kde* | 07:19 |
bpmedley | try "apt-get install task-kde-desktop" just guessing | 07:20 |
no_9 | Trying | 07:21 |
no_9 | Package task KDE desktop has no installation candidate | 07:22 |
no_9 | Likewise for KDE desktop | 07:22 |
no_9 | Wait that's not right | 07:22 |
no_9 | Package task KDE desktop is not available, but is referred to by another package | 07:23 |
no_9 | I get that same message for KDE standard | 07:23 |
no_9 | I really didn't want to use KDE, I just wanted to use some KDE apps, like kaffeine | 07:24 |
no_9 | Does devuan not have its own multimedia Source? Do you have to use a Debian release multimedia source? | 07:25 |
no_9 | Something is really not right, I just tried to install midnight Commander and I get package MC has no installation candidate | 07:38 |
ullet | are you encountering unsupported codecs no_9 | 08:17 |
no_9 | pk i got hexchat installed | 08:36 |
no_9 | ok too | 08:36 |
no_9 | i think i inaccurately copied the sources list 1st time around | 08:37 |
ullet | heh | 08:39 |
no_9 | nothing is so sure after 3-4 shiner bocks | 08:39 |
no_9 | what do you guys do for multimedia? | 08:44 |
no_9 | use debian jessie? | 08:45 |
ullet | devuan no_9 | 08:55 |
ullet | all my media plays fine for me | 08:55 |
no_9 | well whats your line in /apt/sources.list for multimedia? i couldnt find olne for devuan | 08:57 |
no_9 | one | 08:57 |
ullet | none | 08:57 |
ullet | just the three beowulf entries no_9 | 08:59 |
no_9 | link pls? | 09:00 |
no_9 | i thought beowulf was a cluster distro | 09:00 |
ullet | http://0x0.st/i2uP.list | 09:00 |
no_9 | if i mix that in w/ 64 bit ascii am i going to break everything to hell? | 09:01 |
ullet | well to upgrade from ascii to beowulf i switched the ascii lines to beowulf and apt updated / apt upgraded | 09:02 |
ullet | but there might be issues | 09:02 |
no_9 | WAIT a minute theres no multimedia spurces in that | 09:02 |
no_9 | sources | 09:02 |
ullet | yeah | 09:03 |
ullet | i see there's a forum post that mentions deb.devuan.org/merged/dists/ascii/contrib and deb.devuan.org/merged/dists/ascii/non-free | 09:03 |
no_9 | ive just used debian jessie multimedia | 09:03 |
ullet | i don't think devuan has a 'multimedia' repository | 09:04 |
no_9 | already got jack qsynth installed and rosegarden is installing... | 09:04 |
ullet | mhm | 09:05 |
ullet | so for nonfree stuff you might want smth like | 09:05 |
ullet | deb http://deb.devuan.org/merged ascii main contrib non-free | 09:05 |
ullet | deb http://deb.devuan.org/merged ascii-security main contrib non-free | 09:05 |
no_9 | i got ascii on a usb stick & it uses the swap partition on the hdd | 09:07 |
ullet | nice | 09:07 |
no_9 | i just wanted to see if it installed my faves then i was migratring it to the hard drive | 09:08 |
ullet | what's an example of a package that needs multimedia repo? | 09:08 |
no_9 | jack? | 09:08 |
ullet | in debian | 09:08 |
ullet | if i apt-cache search jack i see it | 09:09 |
no_9 | qhackctrl? | 09:09 |
ullet | you know apt-cache search? | 09:09 |
no_9 | bah | 09:09 |
no_9 | qjackctrl | 09:09 |
ullet | qjackctl also | 09:09 |
ullet | you know apt-cache search? | 09:09 |
ullet | use apt-cache search, no_9 | 09:09 |
no_9 | cant i feel in the dark til i find something? | 09:10 |
ullet | :) | 09:10 |
ullet | the assumption that you require a multimedia repo for jack or qjackctl is revealed to be incorrect by running apt-cache search for those packages | 09:10 |
no_9 | well i'll try that after this install finishes | 09:11 |
no_9 | beowulf is not for clustering? a distro by that name was for that... | 09:17 |
ullet | yeah that was a popular thing around 20 years ago | 09:24 |
ullet | beowulf is also a name for a devuan distro | 09:24 |
no_9 | well whats the most recent stable? | 09:25 |
no_9 | i thought it was ascii | 09:25 |
ullet | stable is ascii | 09:26 |
ullet | not sure why you'd want that | 09:26 |
no_9 | beowu;f is...testing? | 09:26 |
ullet | i thought it was equivalent to 'sid' from debian | 09:27 |
no_9 | oh shit--experimental | 09:27 |
no_9 | sid was rge kid who broke all his toys maaaaan | 09:28 |
no_9 | the kid** | 09:28 |
ullet | for general home-use i pick beowulf to not have ancient versions of software | 09:28 |
ullet | and since it's the environment i'm used-to, i used it for business too | 09:29 |
no_9 | stuff sometimes doesnt work for weeks on sid dood | 09:35 |
no_9 | i appreciate the efforts of those who use it | 09:36 |
ullet | k hm | 09:36 |
ullet | different people's usage patterns and preferences expose different problems | 09:37 |
no_9 | but you have to have something the equivalent of the old chevy when you got to go somewhere today | 09:37 |
no_9 | ie stable | 09:37 |
ullet | ok | 09:38 |
no_9 | always nice to know the beast will actually START | 09:38 |
_abc_ | I see ascii and so on has packages so old it lives up to the bad name debian has created for this "feature" of the distribution. Some things are updated (firefox libreoffice) others dangerously not so (irssi and so on). Is there a howto and a "devuan way", perhaps just a checklist, to maintain a package stream repo locally with updated versions of what one needs? | 10:30 |
Wonka | _abc_: you know about ascii-backports? | 10:32 |
_abc_ | I do but I think I have it enabled and still does not help? Let me check. | 10:32 |
Wonka | https://www.devuan.org/os/documentation/dev1fanboy/upgrade-to-ascii has the four useful lines | 10:33 |
_abc_ | Yes it is on and does not help | 10:34 |
_abc_ | ascii-backports is on | 10:34 |
Wonka | hm. cannot help further then; I'm running on unstable mostly... | 10:34 |
_abc_ | My sources.list looks pretty much like the one in your link, except I also had ascii-contrib on in the past, is commented now, and I also have the sources for all the other four enabled | 10:37 |
ejr | yesterday around this time i asked for advice in here about an fsck error message i get when i boot into devuan. it is related to an external encrypted device i had attached to my system some days ago. I was advised to run fsck on the decrypted device, which i did, but the error message persists (even thoug fsck said the device was clean) | 11:31 |
ejr | is there anything else i can do? | 11:31 |
fsmithred | ejr, maybe change the frequency of checks using tune2fs | 11:58 |
ullet | why did debian switch default user to 1001 instead of 1000 ? | 12:04 |
ullet | can i mount old drives mapping 1000 to 1001? | 12:04 |
fsmithred | where did you get 1001? or how? | 12:05 |
ullet | in crappy debian | 12:07 |
ullet | sorry it's offtopic | 12:08 |
fsmithred | not really | 12:08 |
ullet | but maybe others have the issue | 12:08 |
fsmithred | I just checked a buster to beowulf system, and user is 1000 | 12:08 |
fsmithred | who is 1000 on your system? | 12:08 |
ullet | during mount time, there is probably a way to map uid:gid | 12:08 |
fsmithred | or add your user to the group | 12:09 |
agris | !ops | 12:15 |
agris | Can a channel operator please privmsg me | 12:17 |
ullet | there is no 1000 on the system fsmithred | 12:19 |
agris | tobypus, Sigyn | 12:20 |
agris | MinceR, | 12:20 |
ullet | debian stretch 9.12 | 12:21 |
ullet | ullet:x:1001:1002:.... | 12:22 |
ullet | user2:x:1002:1003:..... | 12:22 |
openbsdtai123 | Is there a small too in devuan to create a bootable usbstick with EFI ? | 13:13 |
openbsdtai123 | (too=> tool) | 13:13 |
ullet | idk | 13:14 |
ErRandir | ullet: the 1st UID used can be adjusted in /etc/adduser.conf | 13:31 |
ullet | that's curious. because it shows FIRST_UID=1000 and GID as well | 13:33 |
ullet | maybe the vendor added a hidden user >_< | 13:33 |
ullet | thanks ErRandir | 13:33 |
ErRandir | or maybe they added a temp uid for the install and removed that later | 13:34 |
fsmithred | openbsdtai123, use dd or cat. The isos are already efi-bootable except for the minimal-live isos and the i386 desktop-live | 13:54 |
openbsdtai123 | thank you. I did cat of efi... I would like to mod it to boot slack and devuan using EFI (I know efi :( helas). I will try cat and then to add a partition. | 14:05 |
fsmithred | you want multi-boot with live systems, or installer isos? | 14:09 |
fsmithred | if just live systems, you can use refracta2usb (not in repo) | 14:09 |
openbsdtai123 | I saw that elilo could eventually do this... | 14:10 |
fsmithred | I've never used elilo | 14:10 |
fsmithred | for multi-boot usb, you're better off making a fat32 partition instead of imaging the isohybrid file. With the amd64 isohybrids, you can't make a second partition. | 14:28 |
ShorTie | justice ?? | 14:33 |
ullet | for great justice | 15:12 |
MinceR | there is no justice. there is just us. | 15:15 |
ham5urg | I wanted to install a devuan lxc container inside a devuan machine. But no template is available. Has anybody got it run? | 15:40 |
ullet | tell the lxc folks to make one? | 15:41 |
zatumil | maybe you can find a devuan dockerfile and use this in your lxc etc ... | 15:46 |
fsmithred | ham5urg, maybe this: https://dev1galaxy.org/viewtopic.php?id=617 | 15:49 |
fsmithred | oh, if you use that, don't use the mirror he used. Use deb.devuan.org instead. | 15:51 |
systemdlete | So it looks like beowulf has trouble launching doveconf during boot, but not after boot. I cannot get haveged to work, unless I run it from the command line. But it doesn't seem to matter since it doesn't run at boot anyway. | 21:46 |
systemdlete | A lot of people use dovecot, so it is helpful to attend to this. | 21:46 |
systemdlete | Again, putting an strace on doveconf during boot shows it hangs on getrandom() | 21:47 |
systemdlete | This is running as guest VM under vbox 6.1 | 21:48 |
fsmithred | looks like haveged runs here | 21:52 |
fsmithred | does it start after dovecot? | 21:53 |
systemdlete | That's my theory. I don't see where it is starting or failing to start in the boot. I may have missed it though. | 22:12 |
systemdlete | The log files show that audit is DENYing it to start | 22:12 |
systemdlete | beowulf-vm kernel: [ 949.375612] audit: type=1400 audit(1588475434.092:46): apparmor="DENIED" operation="mknod" profile="/usr/sbin/haveged" name="/run/haveged.pid" pid=4122 comm="haveged" requested_mask="c" denied_mask="c" fsuid=0 ouid=0 | 22:12 |
systemdlete | I did not explicitly install apparmor, though. | 22:13 |
systemdlete | I find it interesting that it works for YOU fsmithred. Do you have apparmor installed? | 22:13 |
fsmithred | I guess so - I see apparmor messages | 22:14 |
fsmithred | hang on.. (I don't have dovecot) | 22:14 |
systemdlete | ah. | 22:14 |
fsmithred | yeah, I do | 22:14 |
systemdlete | oh. | 22:14 |
fsmithred | have apparmor | 22:14 |
systemdlete | dovecot? | 22:14 |
fsmithred | I had to mess with haveged in my refracta live isos | 22:15 |
fsmithred | no dovecot | 22:15 |
fsmithred | is there an /etc/init.d/dovecot? | 22:15 |
systemdlete | can you install it (please) just to see if you get the hang also? (yes, there is an init for it) | 22:15 |
fsmithred | look at /etc/rc2.d/ | 22:15 |
fsmithred | which comes first, haveged or dovecot? | 22:16 |
fsmithred | S0... | 22:16 |
systemdlete | Both are S03 | 22:16 |
fsmithred | hm | 22:17 |
fsmithred | they're probably alphabetical when the numbers are the same | 22:17 |
systemdlete | But even if I twiddled with the ordering a bit (assign S02 to haveged, e.g) | 22:17 |
systemdlete | haveged does not start, period, when I run "service haveged start" from the command line AFTER boot | 22:17 |
systemdlete | I mean, happy to try it, if you think that gives us more info | 22:18 |
fsmithred | I think haveged runs and stops | 22:18 |
systemdlete | maybe safer to change dovecot to S04* | 22:18 |
fsmithred | I see a message at 29 seconds into boot | 22:18 |
systemdlete | but it does not launch a daemon | 22:18 |
fsmithred | but it's not running now | 22:18 |
systemdlete | yup | 22:18 |
systemdlete | run "service haveged status" | 22:19 |
systemdlete | does it say running? | 22:19 |
systemdlete | also, ps -ef |grep haveged | 22:19 |
fsmithred | not running, and not in ps output | 22:19 |
systemdlete | aha! | 22:19 |
fsmithred | lemme see what it does in live session | 22:20 |
systemdlete | So at least you and I are getting the same results for haveged, which is comforting | 22:20 |
systemdlete | (my install of beowulf is not too screwed up) | 22:20 |
systemdlete | fsmithred: There is a -F option to haveged that does work, at least at the command line, bare of the service scripts | 22:21 |
systemdlete | I can't seem to get that option to work with the service scripts | 22:22 |
systemdlete | script | 22:22 |
systemdlete | the problem is that the service scripts need the service (haveged) to fork itself off | 22:22 |
fsmithred | oh, haveged is running in live session | 22:22 |
fsmithred | because I started it - had to make a live-config script to run it so ssh host keys could be generated at boot | 22:23 |
systemdlete | So then the start-stop daemon would have to be set to do that instead. But it might work for me. | 22:23 |
systemdlete | Is that the problem -- missing keys? | 22:23 |
systemdlete | nvm | 22:23 |
systemdlete | I get it | 22:23 |
fsmithred | I have to generate new keys every time the live session boots... | 22:24 |
fsmithred | otherwise, everyone who uses that iso has the same host keys | 22:24 |
systemdlete | right, right -- I spoke too soon | 22:24 |
systemdlete | btw, is this in a VM or on HW | 22:24 |
systemdlete | fsmithred? VM or not? | 22:25 |
fsmithred | live session is in VM | 22:25 |
fsmithred | non-live is on hardware | 22:26 |
systemdlete | oh. | 22:26 |
fsmithred | that's this system I'm currently using | 22:26 |
systemdlete | ok | 22:26 |
systemdlete | so we might be comparing apples and oranges | 22:26 |
systemdlete | (I'm not sure how haveged works -- does it access randomizing hw if available? idk) | 22:27 |
fsmithred | we're comparing haveged in two different environments | 22:27 |
fsmithred | neither am I | 22:27 |
fsmithred | I just know it needs to run before those keys get made | 22:27 |
fsmithred | oh | 22:27 |
systemdlete | and before dovecot | 22:27 |
fsmithred | cat /etc/default/haveged | 22:27 |
systemdlete | right, did that already | 22:27 |
systemdlete | I have not modified it. Still says Daemon args are -w 1024 | 22:28 |
systemdlete | I tried adding that -F option but then the service doesn't background itself | 22:28 |
systemdlete | but I can get haveged to run directly from shell with -F, otherwise it exits and takes any children with it, apparently | 22:29 |
systemdlete | I think there is an option to start-stop that will do that, though | 22:29 |
fsmithred | maybe hack the do_start script to leave you some output in a file | 22:29 |
systemdlete | do_start? | 22:30 |
fsmithred | in /etc/init.d/haveged | 22:30 |
systemdlete | oh, you mean the function? | 22:30 |
systemdlete | I can do that. | 22:30 |
systemdlete | I was going to hack it a bit anyway, just to see if I could get the silly thing to work at all | 22:30 |
fsmithred | looks like it starts it twice - once with --test and once wit DAEMON_ARGS | 22:31 |
systemdlete | (I don't care -- I'm not really using beowulf VM for anything anyway) | 22:31 |
systemdlete | Yes. Not sure why, but it does do that. | 22:31 |
systemdlete | I think the first time is to get some info | 22:31 |
systemdlete | I've been toying with haveged for a day now | 22:32 |
systemdlete | not like it matters. I'm not going anywhere today... | 22:32 |
fsmithred | man start-stop-daemon explains what the return values mean | 22:33 |
systemdlete | That's rare in *nix | 22:33 |
systemdlete | (sarcasm only) | 22:33 |
fsmithred | lol, yeah | 22:33 |
systemdlete | the service script seems to return 0 always | 22:34 |
systemdlete | I've also tried running haveged exactly like the script, and it also returns 0 | 22:35 |
systemdlete | At least, I've never seen it return non-zero yet | 22:35 |
systemdlete | I put a set -xv to see how haveged and start-stop are being called | 22:35 |
systemdlete | So I could test it directly from shell | 22:35 |
systemdlete | And I also looked for I/O redirection, etc. (there was none, but I always check anyway) | 22:36 |
systemdlete | start-stop-daemon --start --pidfile /var/run/haveged.pid --exec /usr/sbin/haveged -- -w 1024 -v32 -F | 22:39 |
systemdlete | works | 22:39 |
systemdlete | but take the -F away and there is nothing! | 22:39 |
systemdlete | guess what fsmithred -- my theory works! | 22:39 |
fsmithred | ? | 22:40 |
systemdlete | If I run the start-stop daemon with -b (background the target) and pass -F to haveged, the service script finally works !!! | 22:40 |
systemdlete | It's stupid, clumsy, and ridiculous, but it works | 22:40 |
systemdlete | it's no way to run a country, but... | 22:41 |
fsmithred | so you can change the command in the init script | 22:42 |
fsmithred | or maybe in /etc/default | 22:42 |
fsmithred | that would be cleaner | 22:42 |
systemdlete | well, I can change the options to haveged, and I've done that yesterday when fooling with this | 22:43 |
fsmithred | yeah, you can't use the -b that way - wrong command | 22:43 |
systemdlete | but there is no similar option for the script itself | 22:43 |
systemdlete | this would be a... | 22:44 |
fsmithred | hack the init script and watch out for any upgrades to haveged | 22:44 |
systemdlete | ***hack***!!! | 22:44 |
systemdlete | yup | 22:44 |
systemdlete | sad | 22:44 |
systemdlete | hopefully upgrade will fix this to begin with | 22:44 |
systemdlete | maybe a new version of haveged that works more like one would expect | 22:44 |
fsmithred | check for bug reports in debian | 22:45 |
systemdlete | will do | 22:45 |
systemdlete | or is it the haveged project? | 22:45 |
fsmithred | start with debian - someone else may have reported it | 22:45 |
systemdlete | ok | 22:45 |
systemdlete | what is the meaning of the name? | 22:45 |
fsmithred | and if you file, they'll let you know if it needs to go upstream | 22:45 |
systemdlete | haveged | 22:45 |
systemdlete | ok | 22:45 |
fsmithred | havege is some algorithm, I think | 22:46 |
systemdlete | ah, ok | 22:46 |
systemdlete | HAVEGE (HArdware Volatile Entropy Gathering and Expansion) algorithum | 22:47 |
systemdlete | (love that "u" in the last word there) | 22:47 |
systemdlete | https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=941626 | 22:50 |
systemdlete | so it appears they know about this | 22:50 |
systemdlete | has something to do with switching to sysvinit from... uh... err... | 22:51 |
* systemdlete burps on himself | 22:51 | |
systemdlete | excuse me | 22:51 |
systemdlete | the trail of damage never ends | 22:52 |
systemdlete | for now, I will hack the script. I'm not going to bother with /etc/default at all. if the problem gets fixed, I'd rather make a clean break | 22:52 |
systemdlete | But... while I can get it to start now, it fails to quit when I run stop | 22:56 |
systemdlete | apparently, status_of_proc does not work for this case | 22:56 |
fsmithred | I run it without any options in the live-config script | 23:02 |
systemdlete | "it" -- haveged or start-stop | 23:02 |
fsmithred | haveged | 23:02 |
systemdlete | hmm | 23:03 |
fsmithred | you could just put a line in /etc/rc.local | 23:03 |
systemdlete | I probably don't need to worry about stopping it. And, if I need to, can always do killall or other manual op for now... | 23:03 |
systemdlete | when you say YOU run it etc, is this a custom ISO you are using? | 23:04 |
fsmithred | yeah | 23:04 |
systemdlete | well... | 23:04 |
systemdlete | I think my solution will suffice for the moment. | 23:04 |
fsmithred | yeah, I don't think you want it to stop | 23:05 |
systemdlete | although, I don't plan on using dovecot anyway, which is how this ball started rolling in the first place | 23:05 |
systemdlete | In my actual beowulf environment, on real hw, upgrading from ascii, someday, then yes. | 23:05 |
systemdlete | that is the only reason I am testing dovecot | 23:05 |
systemdlete | I suppose I might want to test dovecot a little just to see if THAT works... lol | 23:06 |
fsmithred | good idea | 23:07 |
systemdlete | I see I need to force haveged to run before dovecot. I should be able to rename it from S3haveged to S3_haveged | 23:09 |
systemdlete | crum | 23:10 |
systemdlete | nope | 23:10 |
systemdlete | dot, dash, equals might work | 23:10 |
systemdlete | oh wait | 23:11 |
systemdlete | underscore comes before lowercase, so that should work | 23:11 |
systemdlete | ok, man gotta plan | 23:11 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!