-
Notifications
You must be signed in to change notification settings - Fork 195
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[FEA] Improved RMM internal logging control #564
Comments
Similar to |
The log is not specific to |
@jlowe can this be considered fixed now? |
While we can workaround the undesired |
Yeah planning to add config options in 0.17 |
This issue has been marked rotten due to no recent activity in the past 90d. Please close this issue if no further response or action is needed. Otherwise, please respond with a comment indicating any updates or changes to the original issue and/or confirm this issue still needs to be addressed. |
This issue has been marked stale due to no recent activity in the past 30d. Please close this issue if no further response or action is needed. Otherwise, please respond with a comment indicating any updates or changes to the original issue and/or confirm this issue still needs to be addressed. This issue will be marked rotten if there is no activity in the next 60d. |
This is still desired. It was originally targeted for 0.17, but I do not think it was implemented. |
This issue has been labeled |
Still desired. |
This issue has been labeled |
Still desired. |
@jlowe spdlog provides a way to add and remove sinks from already running loggers. We could provide an RMM specific way of specifying a new sink, but it would be less general than the API that spdlog already provides. Currently you can do this (not thread safe):
You can also add new sinks without deleting the existing sink(s). See gabime/spdlog#820 Is this sufficient, or do you think we need to provide RMM-specific wrappers for this functionality? |
Thanks, @harrism! I didn't realize we could replace a sink on an existing log. I think that should be sufficient to do what we need. |
I didn't either, but when I started to look at fixing this, I discovered it. |
Is your feature request related to a problem? Please describe.
Recently RMM has started emitting an
rmm_log.txt
file that logs internal RMM messages such as this (which occurred in an application that can recover from many RMM OOM situations and continue):There are cases where either the current working directory may not be writable or will not persist past the application lifetime, and thus the user would never see the logs. Therefore we need to be able to control where these logging messages are written, including the possibility of using existing file descriptors (e.g.; stdout or stderr).
Describe the solution you'd like
RMM could provide an API to set where internal log messages will be written. The initial implementation could just take a file descriptor to use. Being able to provide a spdlog sink or something similar would provide the most flexibility, as the application could implement some dynamic control behavior during logging if desired.
Describe alternatives you've considered
Compile-time flags would be less ideal since there is no flexibility. For example, if we need to compute the appropriate logging path at runtime, compile-time flags aren't going to work.
The text was updated successfully, but these errors were encountered: