cname host is logged twice #53

New Issue

ungleich-gitea · 2021-11-20T11:25:14Z

ungleich-gitea commented

2021-11-20 11:25:14 +00:00

ander@limukas:~$ grep verbosity .cdist.cfg 
verbosity = INFO
ander@limukas:~$ cdist config cname.example.com
INFO: [272690]: cname.example.com: Starting configuration run
INFO: cname.example.com: Starting configuration run
INFO: [272690]: cname.example.com: Finished successful run in 23.65 seconds
INFO: cname.example.com: Finished successful run in 23.65 seconds
ander@limukas:~$ cdist config 127.0.0.1
INFO: [273185]: 127.0.0.1: Starting configuration run
INFO: [273185]: 127.0.0.1: Finished successful run in 23.87 seconds

only happens with addresses which are cnames. @poljakowski?

``` ander@limukas:~$ grep verbosity .cdist.cfg verbosity = INFO ander@limukas:~$ cdist config cname.example.com INFO: [272690]: cname.example.com: Starting configuration run INFO: cname.example.com: Starting configuration run INFO: [272690]: cname.example.com: Finished successful run in 23.65 seconds INFO: cname.example.com: Finished successful run in 23.65 seconds ander@limukas:~$ cdist config 127.0.0.1 INFO: [273185]: 127.0.0.1: Starting configuration run INFO: [273185]: 127.0.0.1: Finished successful run in 23.87 seconds ``` only happens with addresses which are cnames. @poljakowski?

poljakowski was assigned by ungleich-gitea

2021-11-20 11:25:14 +00:00

ungleich-gitea closed this issue

2021-11-20 11:25:14 +00:00

ungleich-gitea commented

2021-11-20 15:31:38 +00:00

mentioned in commit 06cc20aa28

mentioned in commit 06cc20aa28a3cdfc09dd5642fb66e1b24a8e8d2b

ungleich-gitea commented

2021-11-20 15:31:39 +00:00

closed via merge request !883

ungleich-gitea commented

2021-11-20 15:31:40 +00:00

good work! thanks!

ungleich-gitea commented

2021-11-20 15:31:42 +00:00

mentioned in merge request !883

ungleich-gitea commented

2021-11-20 15:31:43 +00:00

mentioned in commit f4e1bbc87e

mentioned in commit f4e1bbc87e9127ed4ac1789c700e1cbee02ff61c

ungleich-gitea commented

2021-11-20 15:31:44 +00:00

@ander Figured it out :) This has nothing to do with CNAME, but with cdist. as first part of hostname.

It's about logger hierarchy which is defined by dots in logger name.

https://docs.python.org/3/library/logging.html#module-logging

The name is potentially a period-separated hierarchical value, like foo.bar.baz (though it could also be just plain foo, for example). Loggers that are further down in the hierarchical list are children of loggers higher up in the list.

So we have cdist logger, created before target host processing.
Then we process host cdist.whatever. Logging creates logger hierarchy, which means cdist logger is parent of .whatever. When cdist.whatever logs something, this is also propagated to ancestors.
There is one ancestor defined, cdist logger, so it too writes log message.
That is why all are written twice.
For parallel part, cdist ancestor is defined without process ID in its format, so you have one line with process ID, and one line without it.

Since we don't use hierarchy with propagation in its full potential, I will try fixing this bug with setting propagate to False.

@ander Figured it out :) This has nothing to do with `CNAME`, but with `cdist.` as first part of hostname. It's about logger hierarchy which is defined by dots in logger name. https://docs.python.org/3/library/logging.html#module-logging > The name is potentially a period-separated hierarchical value, like foo.bar.baz (though it could also be just plain foo, for example). Loggers that are further down in the hierarchical list are children of loggers higher up in the list. So we have `cdist` logger, created before target host processing. Then we process host `cdist.whatever`. Logging creates logger hierarchy, which means `cdist` logger is parent of `.whatever`. When `cdist.whatever` logs something, this is also propagated to ancestors. There is one ancestor defined, `cdist` logger, so it too writes log message. That is why all are written twice. For parallel part, `cdist` ancestor is defined without process ID in its format, so you have one line with process ID, and one line without it. Since we don't use hierarchy with propagation in its full potential, I will try fixing this bug with setting `propagate` to `False`.

ungleich-gitea commented

2021-11-20 15:31:46 +00:00

cnameconfig.log

If you give me your id.pub, I can give you access for testing.

[cnameconfig.log](/uploads/c68460a24638ebdbb62ed66b74c8645d/cnameconfig.log) If you give me your `id.pub`, I can give you access for testing.

ungleich-gitea commented

2021-11-20 15:31:47 +00:00

@ander Do you have a trace log level output for cname case?