Project

General

Profile

Actions

Bug #5760

open

statd could be too slow with notifications

Added by Marcel Telka over 7 years ago.

Status:
New
Priority:
Normal
Assignee:
-
Category:
nfs - NFS server and client
Start date:
2015-03-27
Due date:
% Done:

0%

Estimated time:
Difficulty:
Medium
Tags:
needs-triage
Gerrit CR:
External Bug:

Description

After the reboot (or after SM_SIMU_CRASH) statd tries to notify other statd:s on all monited machines that the status of the local machine changed. This notification might be very slow for various reasons (for example, the remote machine is down), so the whole notification process might last far more than the grace_period (by default 90 seconds) in lockd.

If this happens, the lockd might leave the grace period before all of the remote machines are notified, and before such machines had a chance to reclaim their locks.

No data to display

Actions

Also available in: Atom PDF