-
Notifications
You must be signed in to change notification settings - Fork 881
Documentation: description of known logging issues #2954
Conversation
@@ -129,6 +129,18 @@ $ journalctl -M rkt-bc3c1451-2e81-45c6-aeb0-807db44e31b4 -t redis | |||
|
|||
Additionaly, logs can be programmatically accessed via the [sd-journal API](https://www.freedesktop.org/software/systemd/man/sd-journal.html). | |||
|
|||
In some rare situations whena application inside the pod is writing to stdout/stderr instead of journal (i.e. nginx) there is no way to obtain logs. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
when an
29604dc
to
99b15a7
Compare
@@ -129,6 +129,27 @@ $ journalctl -M rkt-bc3c1451-2e81-45c6-aeb0-807db44e31b4 -t redis | |||
|
|||
Additionaly, logs can be programmatically accessed via the [sd-journal API](https://www.freedesktop.org/software/systemd/man/sd-journal.html). | |||
|
|||
Currently there are known 2 main issues with loggin in rkt: |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
s/loggin/logging/
@pskrzyns thanks, I had a pass on it. I think some portions should be reworded a bit, as the details are a bit tricky and should be written down correctly to avoid even more confusion. |
@lucab PTAL |
``` | ||
should be added to ```/etc/nginx/nginx.conf``` | ||
|
||
* If application inside the pod(i.e. newest etcd) writes directly to journald, it's resultig in its log entries being scoped by machind UUID. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Missing space before opening brace.
Also - "resulting".
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
s/machind/machine/
@@ -129,6 +129,28 @@ $ journalctl -M rkt-bc3c1451-2e81-45c6-aeb0-807db44e31b4 -t redis | |||
|
|||
Additionaly, logs can be programmatically accessed via the [sd-journal API](https://www.freedesktop.org/software/systemd/man/sd-journal.html). | |||
|
|||
Currently there are known 2 main issues with logging in rkt: |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
"two known main issues"
3ddd43b
to
3b3840e
Compare
``` | ||
should be added to ```/etc/nginx/nginx.conf``` | ||
|
||
* If application inside the pod (i.e. newest etcd) writes directly to journald, it's resulting in its log entries being scoped by machine UUID. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
If an application inside the pod writes directly to journald (e.g. some newer versions of etcd), the resulting log entries will not be written to stdout/stderr.
I don't really get the comment about the scoping.. doesn't that happen in every case? How is this any different to e.g. line 94 onwards above?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@jonboulle it isn't different, your re-wording is the correct one.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
"Some applications, like newer version of etcd, write directly to journald. Such log entries will not be written to stdout or stderr. These logs can be retrieved by passing the machine ID to journalctl:"
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
thx @joshix
@pskrzyns ping for a rework. |
@lucab so i should add etcd as example of application witch will cause problems in line 90 ? |
@pskrzyns I think that line 90 is ok, the journalctl part will work fine (it may just get some slightly different metadata). The issue to document here is that |
@pskrzyns ping, will you be able add the requested bits? thanks! |
3b3840e
to
f3e466a
Compare
Updated |
LGTM |
Documentation changes to cover some logging caveats.
Closes #2640
Closes #2417