2019-05-15 01:05:19 UTC - Carlos Santana: We let Kubernetes handle the logs and forward them to a managed log service https://openwhisk-team.slack.com/archives/C3UDXSFA6/p1557882319038700?thread_ts=1557875726.037400&cid=C3UDXSFA6 ---- 2019-05-15 01:10:27 UTC - Rodric Rabbah: tada : Carlos Santana, Jin Choi, Dominic Kim, Rob Allen, James Thomas star : Carlos Santana, chetanm, Sugandha, Neeraj Mangal, James Thomas partyparrot : Dominic Kim, James Thomas whisking : Rob Allen, James Thomas, mmarth https://openwhisk-team.slack.com/archives/C3UDXSFA6/p1557882627038900 ---- 2019-05-15 01:39:01 UTC - Dominic Kim: Sure. ``` ${PATH}/wsklogs/*/*.log { su root root daily rotate 6 copytruncate compress delaycompress notifempty missingok minsize 30M maxsize 100M } ``` +1 : Jin Choi https://openwhisk-team.slack.com/archives/C3UDXSFA6/p1557884341039500?thread_ts=1557820193.034400&cid=C3UDXSFA6 ---- 2019-05-15 01:39:15 UTC - Dominic Kim: You can change the configuration as you want. https://openwhisk-team.slack.com/archives/C3UDXSFA6/p1557884355039700?thread_ts=1557820193.034400&cid=C3UDXSFA6 ---- 2019-05-15 02:05:36 UTC - Jin Choi: Thank you very much @Dominic Kim +1 : Dominic Kim https://openwhisk-team.slack.com/archives/C3UDXSFA6/p1557885936040100?thread_ts=1557820193.034400&cid=C3UDXSFA6 ---- 2019-05-15 02:12:31 UTC - Jin Choi: Thanks, I'll look further into your option once I deploy it on k8s. Unfortunately, for the time being, I am stuck with my Openwhisk deployed on bare metal servers.
I learned `copytruncate` option from logrotate tool and it seems charming in a sense, but I am concerned with periodical I/O peaks due to the heavy copying of the log files. You see, lots of open source solutions provide a way to reopen its log file descriptor so that we can simply move the file to another name after sending it a signal, say USR1. No doubt for better performance!. So the question is, does `controller` on Openwhisk provide such signal handling for such a purpose? If not, do you agree that the only option left for handling the ever-growing log files is `copytruncate` from logrotate? Or do you know a beeter option for me? https://openwhisk-team.slack.com/archives/C3UDXSFA6/p1557886351040500?thread_ts=1557875726.037400&cid=C3UDXSFA6 ---- 2019-05-15 04:26:49 UTC - chetanm: One possible option can be to tweak the logback config and let it write to a file under a mounted volume +1 : Jin Choi https://openwhisk-team.slack.com/archives/C3UDXSFA6/p1557894409041100?thread_ts=1557875726.037400&cid=C3UDXSFA6 ---- 2019-05-15 04:27:12 UTC - chetanm: Logback provide lots of support for log rotation with various policies etc thankyou : Jin Choi https://openwhisk-team.slack.com/archives/C3UDXSFA6/p1557894432041300?thread_ts=1557875726.037400&cid=C3UDXSFA6 ---- 2019-05-15 11:05:24 UTC - Carlos Santana: Before Kubernetes we used logstash, container output was redirected to a file on volume on the VM and then logstash would monitor the files process the new line and forward the logs (elasticsearch) thankyou : Jin Choi https://openwhisk-team.slack.com/archives/C3UDXSFA6/p1557918324044800?thread_ts=1557875726.037400&cid=C3UDXSFA6 ---- 2019-05-15 14:05:07 UTC - Matt Rutkowski: REMINDER: ~ 1 HOUR until the Apache OpenWhisk Tech. Interchange project meeting, * Day-Time: Every other Wednesday, 11AM EST (Eastern US), 5PM CEST (Central Europe), 3PM GMT, 11PM (Beijing) * Zoom: <https://zoom.us/my/asfopenwhisk> https://openwhisk-team.slack.com/archives/C3UDXSFA6/p1557929107045200 ----
