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
Hi.
I receive a lot of empty folders after removing metrics and it is bad for me, because I receive a lot of non exist fields in /find request. Now I fix it by find /mnt/xvdx/ -type d -empty, but maybe we should remove it after "delete" operation?
I can create PR if you agree with it.
The text was updated successfully, but these errors were encountered:
@jjneely, @azhiltsov there is two way how we can do that - with an additional flag to delete command, but I'm not sure because there is a lot of undefined behaviors:
Basically, to integrate the find command invocation into buckyd. I'd be up for that.
But yes, there are a bunch of corner cases and race conditions possible which is why I've been hesitant to add this. I've lost data because of deleting "empty" directories in my graphite cluster before.
The full incantation I suggest is
/usr/bin/find ${prefix}/storage/whisper -type d -empty -mtime +1 -delete
The -mtime setting there does a lot to sanitize possible race conditions, although its not race free.
Hi.
I receive a lot of empty folders after removing metrics and it is bad for me, because I receive a lot of non exist fields in /find request. Now I fix it by
find /mnt/xvdx/ -type d -empty
, but maybe we should remove it after "delete" operation?I can create PR if you agree with it.
The text was updated successfully, but these errors were encountered: