| *** TxGirlGeek has quit IRC | 01:36 | |
| *** TxGirlGeek has joined #starlingx | 02:45 | |
| *** TxGirlGeek has quit IRC | 03:14 | |
| *** TxGirlGeek has joined #starlingx | 03:38 | |
| *** TxGirlGeek has quit IRC | 05:20 | |
| *** billzvonar has joined #starlingx | 12:55 | |
| *** ijolliffe has joined #starlingx | 13:27 | |
| sgw | Morning all | 14:23 |
|---|---|---|
| bwensley | morning | 14:25 |
| billzvonar | or evening, depending where yer at | 14:25 |
| *** rcw has joined #starlingx | 15:02 | |
| *** TxGirlGeek has joined #starlingx | 16:01 | |
| *** openstackstatus has joined #starlingx | 16:06 | |
| *** ChanServ sets mode: +v openstackstatus | 16:06 | |
| *** billzvonar has quit IRC | 16:10 | |
| sgw | slittle1: so I know you ran out of time yesterday, but can you confirm you saw the download_mirror failure with bash/golang/python? | 16:20 |
| slittle1 | yes I did | 16:41 |
| slittle1 | trying several things to try an understand the trigger | 16:41 |
| slittle1 | various yum clean options, including 'clean all' fail to resolve the problem once it starts | 16:42 |
| *** vmrod25 has joined #starlingx | 16:45 | |
| slittle1 | turned on every debug and log option I can find... nothing useful | 16:46 |
| *** ijolliffe has quit IRC | 16:52 | |
| *** vmrod25 has left #starlingx | 16:52 | |
| sgw | let me know how I can help, as I know that I can test it ! | 16:53 |
| *** ijolliffe has joined #starlingx | 17:01 | |
| *** ijolliffe has quit IRC | 17:12 | |
| *** ijolliffe has joined #starlingx | 17:19 | |
| *** ijolliffe has quit IRC | 17:26 | |
| *** ijolliffe has joined #starlingx | 17:30 | |
| *** ijolliffe has quit IRC | 17:34 | |
| *** ijolliffe has joined #starlingx | 17:37 | |
| slittle1 | trying to compare './download_mirror.sh -c ./yum.conf.sample -n -g' vs './download_mirror.sh -c ./yum.conf.sample -g' in two different containers | 17:39 |
| slittle1 | do I get a different result? I think the '-n' version passed ... at least on the first attempt... but now I'm starting to doubt myself | 17:41 |
| sgw | I almost always run with -n | 17:46 |
| *** ijolliffe has quit IRC | 17:47 | |
| sgw | I am trying down a different issue in download_mirror where it just exit with $? = 1 after step #3 before #4 (where it remves i686 packages and count packages) | 17:48 |
| sgw | There is no exit there but it still does! | 17:48 |
| *** ijolliffe has joined #starlingx | 17:48 | |
| dpenney | That's because some of the scripts in stx-tools use "bash -e". So if any command in there has a non-zero rc, it exits the script | 17:53 |
| dpenney | as opposed to doing error checking with helpful log messages :) | 17:53 |
| dpenney | you could run with bash -x to see exactly where it's failing? | 17:54 |
| *** ijolliffe has quit IRC | 17:57 | |
| *** bwensley_ has joined #starlingx | 18:00 | |
| sgw | dpenney: that's the problem when I run with -x it succeeds! | 18:00 |
| slittle1 | sgw: bash -e | 18:01 |
| sgw | bash -e -x? | 18:01 |
| *** dpenney_ has joined #starlingx | 18:01 | |
| slittle1 | all you need to exit is a non-zero return code for any command and the script exits | 18:02 |
| *** dpenney_ has joined #starlingx | 18:03 | |
| slittle1 | sorry , don pointed it out before me | 18:03 |
| *** bwensley has quit IRC | 18:03 | |
| *** rcw has quit IRC | 18:04 | |
| slittle1 | ./download_mirror.sh fails with or without -n inside a container | 18:04 |
| *** rcw has joined #starlingx | 18:04 | |
| dpenney_ | Heisenbug... it only happens when you're not looking at it ;) | 18:05 |
| *** dpenney has quit IRC | 18:05 | |
| *** dpenney_ has quit IRC | 18:05 | |
| *** dpenney has joined #starlingx | 18:06 | |
| sgw | Ok, I tracked it down by running the command between step 3 and 4. the first find looking for Missing keys seems to exit 1 via the grep | 18:06 |
| sgw | find ./output -type f -name "*.rpm" | xargs rpm -K | grep -i "MISSING KEYS" > $LOGSDIR/rpm-gpg-key-missing.txt | 18:06 |
| *** ijolliffe has joined #starlingx | 18:06 | |
| sgw | Since the grep MISSING KEYS does not detect any missing keys grep exits status 1 | 18:08 |
| sgw | So any fully successful download will always exit before completion. | 18:08 |
| dpenney | I'm not a fan of "bash -e" in general... I always favour clean error checking/handling, because of things like this | 18:08 |
| sgw | I guess did Scott introduce the -e or was it was it always there? | 18:09 |
| dpenney | a number of the stx-tools scripts have had it since they were introduced | 18:10 |
| sgw | Strange we have not tripped over this before! | 18:10 |
| *** bwensley_ has quit IRC | 18:11 | |
| *** bwensley has joined #starlingx | 18:11 | |
| dpenney | I raised it early on when I first started trying these starlingx mirror tools, as I hit a few issues with some of the underlying scripts that would fail and cause a cascading silent failure, which would look like it passed from some parent script that didn't check return codes and didn't have the 'bash -e' | 18:12 |
| *** ijolliffe has quit IRC | 18:13 | |
| *** ijolliffe has joined #starlingx | 18:18 | |
| *** ijolliffe has quit IRC | 18:24 | |
| *** ijolliffe has joined #starlingx | 18:26 | |
| *** ijolliffe has quit IRC | 18:44 | |
| slittle1 | I agree we should get rid of 'bash -e' | 18:53 |
| slittle1 | I was hoping to do it as a stand alone update | 18:53 |
| slittle1 | I think I know part of the problem with compiler layer downloads | 18:54 |
| slittle1 | rpm is compiled both in std and installer | 18:54 |
| sgw | Finally had a chance to track down some issues I was seeing, I made some comments in tools review for you. | 18:56 |
| slittle1 | The std one is compiled packaged by the old rpm. The installer one is packaged by the new rpm. | 18:56 |
| slittle1 | great, thanks Saul | 18:56 |
| slittle1 | so I had two seemingly different versions of rpm going by the same name | 18:57 |
| sgw | slittle1: so there are 2 repos? | 18:57 |
| slittle1 | yes, one for each build-type | 18:58 |
| sgw | So actually bash, golang and python were packaged by the "system" rpm | 18:58 |
| slittle1 | yes | 18:58 |
| sgw | kind of the same problem that Don just saw with containerd? | 18:58 |
| slittle1 | Not what I want | 18:59 |
| slittle1 | I think I need to build 'installer' first, then 'std', and only publish 'std' | 19:00 |
| *** ijolliffe has joined #starlingx | 19:00 | |
| *** ijolliffe has quit IRC | 19:08 | |
| *** ijolliffe has joined #starlingx | 19:11 | |
| *** ijolliffe has quit IRC | 19:24 | |
| *** ijolliffe has joined #starlingx | 19:37 | |
| *** ijolliffe has quit IRC | 19:44 | |
| *** ijolliffe has joined #starlingx | 19:45 | |
| *** TxGirlGe_ has joined #starlingx | 19:59 | |
| *** TxGirlGeek has quit IRC | 20:01 | |
| *** TxGirlGe_ has quit IRC | 20:28 | |
| *** TxGirlGeek has joined #starlingx | 20:29 | |
| *** rcw has quit IRC | 20:37 | |
| *** rcw has joined #starlingx | 20:47 | |
| *** ijolliffe has quit IRC | 20:56 | |
| *** ijolliffe has joined #starlingx | 20:59 | |
| *** rcw has quit IRC | 20:59 | |
| *** rcw has joined #starlingx | 21:00 | |
| *** TxGirlGe_ has joined #starlingx | 21:14 | |
| *** TxGirlGeek has quit IRC | 21:15 | |
| *** ijolliffe has quit IRC | 21:15 | |
| *** ijolliffe has joined #starlingx | 21:36 | |
| *** ijolliffe has quit IRC | 21:57 | |
| *** rcw has quit IRC | 22:11 | |
| *** ericho has joined #starlingx | 23:04 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!