Bug 5107 - old local drives mount causes delay for new mounts
Summary: old local drives mount causes delay for new mounts
Status: NEW
Alias: None
Product: ThinLinc
Classification: Unclassified
Component: Local drive redirection (show other bugs)
Version: trunk
Hardware: PC Unknown
: P2 Normal
Target Milestone: MediumPrio
Assignee: Peter Åstrand
URL:
Keywords:
Depends on: 4900 1271
Blocks:
  Show dependency treegraph
 
Reported: 2014-04-22 17:35 CEST by Pierre Ossman
Modified: 2020-04-07 11:21 CEST (History)
1 user (show)

See Also:
Acceptance Criteria:


Attachments

Description Pierre Ossman cendio 2014-04-22 17:35:59 CEST
If you umount a local drive and there is still something accessing it, then the kernel will not fully let go of it. It gets removed from /proc/mounts, but not /proc/fs/nfsfs/volumes. You can also see "server not responding" in dmesg.

The problems start when you connect again and get the same port number. The server believes it's the same NFS server (same address and port), so it will try to reconnect things for a while. This causes a very long and annoying delay before giving up and moving on with the session start up.

Not sure what to do about it though. Random ports? Fiddle some more with the timeout delays? Drop -l from the umount and have just -f? Or simply move to FUSE?

Note You need to log in before you can comment on or make changes to this bug.