[01:35] <maozhou_> cjwatson： in the file of  "configs/development/launchpad-lazr.conf",  the path of  "error_dir“ is ”/var/tmp/lperr",but I can't find the folder "/var/tmp/lperr" ,why?
[01:37] <maozhou_> cjwatson：and I can‘t find any file like “OOPS....”
[04:21] <maozhou_> cjwatson： in the file of  "configs/development/launchpad-lazr.conf",  the path of  "error_dir“ is ”/var/tmp/lperr",but I can't find the oops file ,why?
[05:36] <wgrant> maozhou: If you're running a production instance with a development config, all hope is lost.
[06:34] <wgrant> wut
[06:34] <wgrant> pygit2 gives strs instead of bytes for a whole lot of things on Python 3.
[06:34] <wgrant> eg. ref names
[06:34] <wgrant> what does it do when there are invalid UTF-8 sequences, I wonder.
[06:35] <Spads> or new points in the astral plane 😨
[06:38] <maozhou> where is oops file, is there anyone knows?
[06:39] <wgrant> maozhou: We don't know how your installation is configured.
[06:39] <wgrant> But it sounds like you're using the development config on production, which is terrifying.
[06:39] <wgrant> Depending on how you've set it up, the OOPSes may have ended up in an ephemeral rabbitmq.
[06:42] <maozhou> cjwatson： in the file of  "configs/development/launchpad-lazr.conf",  the path of  "error_dir“ is ”/var/tmp/lperr", but I can't find opps file I need
[06:43] <maozhou> wgrant:in the file of  "configs/development/launchpad-lazr.conf",  the path of  "error_dir“ is ”/var/tmp/lperr", but I can't find opps file I need
[06:45] <wgrant> maozhou: The development configuration is not intended for a production environment.
[06:45] <wgrant> If you're using the default development config, OOPSes may end up in rabbitmq.
[06:45] <wgrant> If you are using rabbitmq for nothing else, consider stopping it. Then further OOPSes will fall back to the filesystem.
[06:47] <maozhou> wgrant: I am using the development config  for test.
[06:50] <maozhou> wgrant:  but ,how to stopping rabbitmq?
[06:51] <wgrant> maozhou: It depends on how you have set up the system. Examine your configuration file to identify where it expects to find rabbitmq, find that rabbitmq, and stop it.
[06:58] <maozhou>  wgrant:  If you're using the default development config, OOPSes may end up in rabbitmq ? where are the  rabbitmq? I just wan't to know the error information when I run "cronscripts/process-job-source.py IInitializeDistroSeriesJobSource'’
[07:00] <wgrant> maozhou: OOPSes will be sent to rabbitmq if it is configured and running. It can be challenging to get them out of rabbitmq, so it's easiest to find the rabbitmq that they're going to, stop that rabbitmq, then trigger the OOPS again. Then it will end up on disk.
[07:05] <maozhou> wgrant: I  understand,but I don't konw where is the configuration file to control rabbitmq.
[07:07] <wgrant> maozhou: The same configuration file that you mentioned earlier.
[07:11] <maozhou> wgrant: ok, I find it, let me try
[07:25] <maozhou> wgrant: I have stoped rabbitmq in file "configs/development/launchpad-lazr.conf",but when I trigger the OOPS again , I still can't find the OOPS file .
[07:28] <wgrant> maozhou: Have you restarted the application?
[07:32] <maozhou> wgrant: Yes I run "make run_all" again on my launchpad server
[07:51] <wgrant> maozhou: Have you confirmed that rabbitmq is not running?
[07:57] <maozhou> wgrant: how to confirmed that rabbitmq is not running? I have changed the value of  "launch:" to "False" which in the "configs/development/launchpad-lazr.conf"  of section "[rabbitmq]"
[07:58] <wgrant> maozhou: Are there any processes running with "beam.smp" in the commandline?
[07:58] <wgrant> If so, which user owns them?
[08:00] <maozhou> wgrant: what is "beam.smp" ?
[08:01] <wgrant> maozhou: beam.smp is the erlang process that hosts (among other things) rabbitmq.
[08:04] <maozhou> wgrant: is it the console which run "make run_all", if there are rabbitmq output on this console, it means the rabiitmq have not been stopped?
[08:06] <wgrant> maozhou: If make run_all starts rabbitmq, it would be reasonable to assume that rabbitmq has started.
[08:08] <maozhou> wgrant: so ,if the rabbitmq has been stopped ,the console of "make run_all" won't output messages?
[08:09] <wgrant> maozhou: run_all will always output some messages, but it shouldn't look like it's starting rabbitmq if it's not starting rabbitmq.
[08:18] <maozhou> wgrant: processes running with "beam.smp" in the commandline can judge the rabbitmq, but ,which commandline ?
[08:19] <wgrant> maozhou: In the output of ps
[08:54] <maozhou> wgrant: if there is processes running with "beam.smp" ， it‘s mean  the rabbitmq have not been stoopped？
[09:26] <wgrant> maozhou: Yes.
[09:42] <maozhou> wgrant: I reboot my launchpad server，before running “make run_all”, there is processes running with "beam.smp". need i run "invoke-rc.d rabbitmq-server stop" to stop  rabbitmq server?
[09:43] <wgrant> maozhou: That is worth a try.
[09:55] <maozhou> wgrant: I have ran "invoke-rc.d rabbitmq-server stop" to stop  rabbitmq server,and I have modified the value of  "launch:" to "False" which in the "configs/development/launchpad-lazr.conf"  of section "[rabbitmq]". But after run "make run_all", there is processes running with "beam.smp" . Nedd I kill it?
[10:02] <wgrant> maozhou: Who owns that process?
[10:03] <wgrant> It may also be the case that run_all disregards the "launch" option. I forget.
[10:03] <wgrant>     if len(requested_services) == 0:
[10:03] <wgrant>         return [svc for svc in SERVICES.values() if svc.should_launch]
[10:03] <cjwatson> run_all hardcodes a bunch of service names, I think.
[10:03] <wgrant> make run_all explicitly names all of the things it wants, so the launch option is not respected.
[10:03] <cjwatson> It should be pretty clear from the Makefile.
[10:04] <cjwatson> Yeah.
[10:08] <maozhou> wgrant: the user who run "make run_all" owns that process
[10:13] <maozhou> I killed the process running with "beam.smp" , then I  trigger the OOPS again.  I find the OPPS  file.
[10:13] <maozhou> thanks for your help :)
[20:56] <blr> morning
[23:38] <wgrant> Morning blr.
[23:39] <wgrant> blr: It turns out that turnip on Python 3 is more difficult than we anticipated, because pygit2 on Python 3 is quite brain-damaged.
[23:39] <wgrant> >>> r.listall_references()
[23:39] <wgrant> Traceback (most recent call last):
[23:39] <wgrant>   File "<stdin>", line 1, in <module>
[23:39] <wgrant> UnicodeDecodeError: 'utf-8' codec can't decode byte 0x80 in position 14: invalid start byte
[23:39] <blr> oh lovely >.<
[23:39] <wgrant> That's what happens if a repo has a ref that isn't in the filesystem's encoding.
[23:40] <wgrant> The low-level library tries to force Unicode onto something that isn't Unicode.
[23:50] <wgrant> https://github.com/libgit2/pygit2/issues/537, let's see what they say.