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 had this weird situation on my server where df reported 29G of disk space in use, but du only found 12G. Doing an lsof +L1 reported lots of deleted zap files kept open by my program (which uses bleve). Sure enough, when I restarted the service, the reports of df and du matched.
Is there any reason for those files being kept open, or is that a bug? Can I do something about it through configuration?
The text was updated successfully, but these errors were encountered:
@elektito This could likely do with the merger's policy not being strict enough for your use-case. As can be seen here, this is the default merger policy ..
You could try defining a more aggressive merger policy to achieve faster deletion of overwritten index content by passing scorchMergePlanOptions within kvConfig while creating/opening a bleve index. Here's an example ..
I had this weird situation on my server where
df
reported 29G of disk space in use, butdu
only found 12G. Doing anlsof +L1
reported lots of deleted zap files kept open by my program (which uses bleve). Sure enough, when I restarted the service, the reports ofdf
anddu
matched.Is there any reason for those files being kept open, or is that a bug? Can I do something about it through configuration?
The text was updated successfully, but these errors were encountered: