| tomreyn | telescopeemoji3: i don't, and wont be able to test, but it'd be easier for others to respond if you told which image exactly (ideally by download location) you used and which how-to you followed and how you determined that this image "updates to rpi2". also note there is #ubuntu-arm | 00:06 |
|---|---|---|
| TJ- | currently building nginx (finally) for Eoan | 00:17 |
| teward | TJ-: that means you are about to fail randomly :P | 00:24 |
| TJ- | it's built, and installed, now I'm trying to restrict the container to 1 CPU although I'm not sure that'll help! | 00:30 |
| TJ- | argh, broken | 00:31 |
| sdeziel | TJ-: once it started fine, you can try "service nginx upgrade" and see if systemd lost track of the new main PID | 00:31 |
| sdeziel | oh shoot | 00:31 |
| telescopeemoji3 | https://www.ubuntu.com/download/iot/raspberry-pi-2-3 im using the 3 image | 00:38 |
| telescopeemoji3 | When it updates it changes the kernel to 2 | 00:38 |
| telescopeemoji3 | oh nm | 00:38 |
| telescopeemoji3 | i may be corrected | 00:39 |
| telescopeemoji3 | uname -a | 00:39 |
| telescopeemoji3 | Linux Raspberry 5.0.0-1006-raspi2 #6-Ubuntu SMP Thu Apr 11 18:04:26 UTC 2019 aarch64 aarch64 aarch64 GNU/Linux | 00:39 |
| TJ- | building again with added logging | 00:52 |
| TJ- | sarnold: nginx - do you know where the ngx_log_stderr() is supposed to go!? I'm not finding the output anywhere, even when starting nginx manually | 02:05 |
| sarnold | TJ-: I do not :( it might require setting a specific loglevel somewhere.. | 02:07 |
| TJ- | no, it seems somehow the debian build isn't including my code! | 02:10 |
| TJ- | I've added lots of ngx_log_stderr() with the string "child_pid" used, and "strings /usr/sbin/nginx" doesn't find it | 02:11 |
| sarnold | strings -a ? | 02:12 |
| TJ- | Aha! id' already done a fakeroot debian/rules build... so when I rebuilt the package it used the build products under ./debian/ ! Doing a 'clean' looks like it'll solve that | 02:14 |
| TJ- | so my adding log messages didn't get into the build products | 02:15 |
| TJ- | sarnold: we have data! https://paste.ubuntu.com/p/nksHVHcF2Q/ | 02:29 |
| sarnold | beautiful! | 02:31 |
| TJ- | sarnold: patch is: http://paste.ubuntu.com/p/sMT3Wnj6pH/ | 02:31 |
| sarnold | TJ-: nginx[10421]: | 02:31 |
| sarnold | TJ-: that never said 10422 :/ | 02:32 |
| TJ- | so it seems like ngx_create_pidfile() is actually writing the /wrong/ pid into the file | 02:32 |
| TJ- | so becuase of how I've coded it it looks like I've got to 'munge' the ngx_pid | 02:33 |
| TJ- | building it now | 02:42 |
| TJ- | I hope this is it; been hacking away for 24 hours now | 02:44 |
| TJ- | sarnold: woohoo! https://paste.ubuntu.com/p/zrBPvhq4SR/ | 03:03 |
| sarnold | nginx[27728]: nginx: This should be child process PID 27729, parent PID 27728 | 03:03 |
| sarnold | TJ-: something's still goofy :( | 03:03 |
| TJ- | nah, it's fine | 03:04 |
| TJ- | systemd uses the parent PID in its logging regardless of child processes | 03:04 |
| TJ- | the funniest is "nginx[27728]: nginx: Process PID 27729 exiting" because that is printed after ngx_pid was set to the child PID :) | 03:05 |
| TJ- | but whereas before this the package installation would hang during configure, it didn't hang this time | 03:05 |
| TJ- | I'll remove the log messages and attach the patch to the bug | 03:06 |
| sarnold | OH! so that part of the message isn't formatted by nginx? it's formatted by systemd? (and it gets it wrong, of course, becasuse why would it be doing the formatting.) | 03:06 |
| TJ- | well, I thought that was it, but just did a package upgrade and got this: "Jun 01 03:19:23 u1910 start-stop-daemon[18631]: /sbin/start-stop-daemon: warning: failed to kill -3: No such process" | 03:20 |
| sarnold | oh my | 03:21 |
| sarnold | why isn't there a process 3? I don't have one either :) | 03:21 |
| sarnold | if that had been -2 or -4 or so on you might have had a bad day :) | 03:22 |
| TJ- | it's -3, NGX_BUSY, the tell-tale | 03:26 |
| TJ- | that's because when ngx_inherited == true ngx_daemon() is NOT called, so child_pid doesn't get set | 03:27 |
| TJ- | I've just got to add another test on ngx_inherited | 03:27 |
| sarnold | the thing is, kill with a negative pid number STILL KILLS -- process groups | 03:28 |
| TJ- | I've fixed that now, doing a build test | 03:35 |
| TJ- | this only affected reloads where there was no fork() | 03:36 |
| TJ- | sarnold: sorted, works perfectly for starts/stops/restarts | 03:46 |
| TJ- | attached to bug report, going to bed now | 03:46 |
| sarnold | TJ-: beautiful! and .. the upgrade? | 03:46 |
| sarnold | aha | 03:46 |
| sarnold | TJ-: goodnight, good weekend :) | 03:46 |
| TJ- | yes, the upgrade too | 03:47 |
| sarnold | :D :D | 03:47 |
| FarhaadN | hi all | 07:32 |
| FarhaadN | i am using ubuntu server 12.04.5 LTS , and i know it's old, but i have problem on this | 07:33 |
| FarhaadN | df not working | 07:33 |
| lotuspsychje | we dont support eol versions FarhaadN | 07:33 |
| lotuspsychje | in this stage you can choose 12.04 ESM or !eolupgrade but reccomended would be clean install 16.04 or 18.04 | 07:34 |
| FarhaadN | can you help please from your experience? | 07:34 |
| FarhaadN | i cant upgrade server to newer version right now | 07:34 |
| lotuspsychje | FarhaadN: supporting an eol version is not how it works, as your system might suffer all kinds of security flaws | 07:35 |
| FarhaadN | i know that , server isolate from security attack, i am just check mount and storage | 07:47 |
| Greyztar | cant seem to install packages no more is there a problem with the mirrors? | 15:21 |
| Greyztar | nvm was my iptables rules o.0 | 15:31 |
| banditti | Hello all. I have a ubuntu server (18.04) with dual nic. (10.0.0.0 and 10.0.100.0). I am trying to get one only to connect to a qnap box via iscsi. when I use iscsiadm it logs in both IPs. Also, need to figure out how to auto boot the mapping. What am I missing? Thank you in advance | 19:01 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!