You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I started using this a while ago at work, and wondered if you would be interested in any of this commit. I haven't done a PR, because I made other local changes at the time and there is a caveat about logging.
I added a MONGODB_QUIET option, on seeing how much log output the progress reporting generates when running under GKE, but then didn't actually use it.
In fact even the version of the mongo tools affects it- eg dumps taken using the mongo 4.2 tools inside the 0.2.0 docker image, in my testing were not understood by a 4.0 mongorestore even though the backup was taken from a 4.0 server.
Keep the log output as it goes (using tee), and in the event of an error, emit the last line of that on Slack. I had noticed that the errors outputted to Slack otherwise amounted to "something went wrong", but gave me little idea what went wrong.
The caveat of what I did here, is that the 2>&1 where I pipe the output to tee, also means the logs aren't differentiated between stdout/stderr when output to the console, and therefore aren't differentiated between info/error in my GKE logs.
I needed to put quotes on something in the helm yaml to get it to deploy.
The text was updated successfully, but these errors were encountered:
I started using this a while ago at work, and wondered if you would be interested in any of this commit. I haven't done a PR, because I made other local changes at the time and there is a caveat about logging.
MONGODB_QUIET
option, on seeing how much log output the progress reporting generates when running under GKE, but then didn't actually use it..tar.gz
extension of the output archives, as it's not a tarball, it's actually a binary format specific to mongodump.tee
), and in the event of an error, emit the last line of that on Slack. I had noticed that the errors outputted to Slack otherwise amounted to "something went wrong", but gave me little idea what went wrong.2>&1
where I pipe the output totee
, also means the logs aren't differentiated between stdout/stderr when output to the console, and therefore aren't differentiated between info/error in my GKE logs.The text was updated successfully, but these errors were encountered: