<html><body><div style="font-family: times new roman, new york, times, serif; font-size: 12pt; color: #000000"><div><br></div><div>Hi,<br></div><div><br></div><div>There are some code changes in EC which is impacting response time of gluster v heal info<br></div><div>I have sent following patch to initiate the discussion on this and to understand why this code change was done.<br></div><div><a href="https://review.gluster.org/#/c/17606/1">https://review.gluster.org/#/c/17606/1</a> <br></div><div><br></div><div>============================<br></div><div> ec: Increase notification in all the cases<br> <br> Problem:<br> "gluster v heal <volname> info" is taking<br> long time to respond when a brick is down.<br> <br> RCA:<br> Heal info command does virtual mount.<br> EC wait for 10 seconds, before sending UP call to upper xlator,<br> to get notification (DOWN or UP) from all the bricks.<br> <br> Currently, we are increasing ec->xl_notify_count based on<br> the current status of the brick. So, if a DOWN event notification<br> has come and brick is already down, we are not increasing<br> ec->xl_notify_count in ec_handle_down.<br> <br> Solution:<br> Handle DOWN even as notification irrespective of what<br> is the current status of brick.<br> <br></div><div>============================<br></div><div>Code change was done by <a href="https://review.gluster.org/#/c/14763/">https://review.gluster.org/#/c/14763/</a> <br></div><div><br></div><div>Ashish<br></div></div></body></html>