retry= is additive with the text-based mount interface

DISCUSSION Steinar H. Gunderson 3 participants 9 messages Apr 25, 2008 4 messages not in the archive

Subsystems (2)

NFS client NFS server

lore

Share & tools

lore

Steinar H. Gunderson

author

2008-04-25 08:05 UTC #0 lore raw

Hi,

(Via a user bug report)

It seems retry= is now additive with the text-based mount interface. In particular, "mount -o retry=0" still gives a two-minute timeout. Is this intentional?

/* Steinar */

Signature · 1 line

Homepage: http://www.sesse.net/

A message that is not in the archive (sent off-list or not yet received): 20080425080535.GA4999-6Z/[email protected]

Chuck Lever

2008-04-25 16:21 UTC #2 lore raw

Quoted 5 lines · Steinar H. Gunderson

(Via a user bug report)
It seems retry= is now additive with the text-based mount interface. In particular, "mount -o retry=0" still gives a two-minute timeout. Is this intentional?

No, it's broken. Can you test the attached patch?

--
Chuck Lever
chuck[dot]lever[at]oracle[dot]com

Steinar H. Gunderson

author

2008-04-25 16:35 UTC #3 lore raw

Quoted 4 lines · Chuck Lever

Quoted 3 lines
It seems retry= is now additive with the text-based mount interface. In particular, "mount -o retry=0" still gives a two-minute timeout. Is this intentional?
No, it's broken. Can you test the attached patch?

It seems it creates rather odd behavior:

fugl:~# mount -t nfs -v -o retry=0 10.0.0.11:/foo /mnt
mount.nfs: timeout set for Thu Jan  1 14:40:25 1970
mount.nfs: text-based options: 'retry=0,addr=10.0.0.11'

fugl:~# mount -t nfs -v -o retry=0 10.0.0.11:/foo /mnt
mount.nfs: timeout set for Sat Jan 10 17:07:36 1970
mount.nfs: text-based options: 'retry=0,addr=10.0.0.11'

/* Steinar */

Signature · 1 line

Homepage: http://www.sesse.net/

A message that is not in the archive (sent off-list or not yet received): 20080425163457.GA6952-6Z/[email protected]

Chuck Lever

2008-04-25 16:48 UTC #5 lore raw

Quoted 4 lines · Steinar H. Gunderson

Quoted 3 lines · Chuck Lever
Quoted 3 lines
It seems retry= is now additive with the text-based mount interface. In particular, "mount -o retry=0" still gives a two-minute timeout. Is this intentional?

Quoted 8 lines

Quoted 1 line
No, it's broken. Can you test the attached patch?
It seems it creates rather odd behavior:
fugl:~# mount -t nfs -v -o retry=0 10.0.0.11:/foo /mnt
mount.nfs: timeout set for Thu Jan  1 14:40:25 1970
mount.nfs: text-based options: 'retry=0,addr=10.0.0.11'

On my test system (Fedora 7) I get correct behavior with both /bin/mount and when using /sbin/mount.nfs directly.

--
Chuck Lever
chuck[dot]lever[at]oracle[dot]com

Steinar H. Gunderson

author

2008-04-25 16:59 UTC #6 lore raw

Quoted 5 lines · Chuck Lever

Quoted 3 lines
fugl:~# mount -t nfs -v -o retry=0 10.0.0.11:/foo /mnt
mount.nfs: timeout set for Thu Jan  1 14:40:25 1970
mount.nfs: text-based options: 'retry=0,addr=10.0.0.11'
On my test system (Fedora 7) I get correct behavior with both /bin/mount and when using /sbin/mount.nfs directly.

It seems the patch has an issue:

+ if (!parse_retry_option(&timeout, options, NFS_DEF_FG_TIMEOUT_MINUTES))

where parse_retry_option() _adds_ to timeout. But timeout is not initialized :-)

/* Steinar */

Signature · 1 line

Homepage: http://www.sesse.net/

A message that is not in the archive (sent off-list or not yet received): 20080425165921.GA7108-6Z/[email protected]

Trond Myklebust

2008-04-25 17:05 UTC #8 lore raw

Quoted 4 lines · Steinar H. Gunderson

Quoted 3 lines · Chuck Lever

Quoted 3 lines

fugl:~# mount -t nfs -v -o retry=0 10.0.0.11:/foo /mnt
mount.nfs: timeout set for Thu Jan  1 14:40:25 1970
mount.nfs: text-based options: 'retry=0,addr=10.0.0.11'

Quoted 8 lines

Quoted 2 lines
On my test system (Fedora 7) I get correct behavior with both /bin/mount and when using /sbin/mount.nfs directly.
It seems the patch has an issue:
+ if (!parse_retry_option(&timeout, options, NFS_DEF_FG_TIMEOUT_MINUTES))
where parse_retry_option() _adds_ to timeout. But timeout is not initialized :-)

Why would we ever want to do anything other than just set the timeout here, Chuck?

Trond

Chuck Lever

2008-04-25 17:19 UTC #9 lore raw

In reply to a message not in the archive

Quoted 7 lines · Steinar H. Gunderson

Quoted 5 lines · Chuck Lever
Quoted 3 lines
fugl:~# mount -t nfs -v -o retry=0 10.0.0.11:/foo /mnt
mount.nfs: timeout set for Thu Jan  1 14:40:25 1970
mount.nfs: text-based options: 'retry=0,addr=10.0.0.11'
On my test system (Fedora 7) I get correct behavior with both /bin/mount and when using /sbin/mount.nfs directly.

Quoted 6 lines

It seems the patch has an issue:
+ if (!parse_retry_option(&timeout, options,
NFS_DEF_FG_TIMEOUT_MINUTES))
where parse_retry_option() _adds_ to timeout. But timeout is not initialized :-)

Just change "*timeout +=" to "*timeout =" in parse_retry_option().

That was the bug I was trying to fix in the first place. Sigh.

--
Chuck Lever
chuck[dot]lever[at]oracle[dot]com

Steinar H. Gunderson

author

2008-04-25 17:38 UTC #10 lore raw

Quoted 3 lines · Chuck Lever

Just change "*timeout +=" to "*timeout =" in parse_retry_option().
That was the bug I was trying to fix in the first place. Sigh.

That fixes it -- thanks!

/* Steinar */

Signature · 1 line

Homepage: http://www.sesse.net/

A message that is not in the archive (sent off-list or not yet received): 20080425173833.GA7306-6Z/[email protected]

Chuck Lever

2008-04-25 20:38 UTC #12 lore raw

Quoted 6 lines · Steinar H. Gunderson

Quoted 3 lines · Chuck Lever
Just change "*timeout +=" to "*timeout =" in parse_retry_option().
That was the bug I was trying to fix in the first place. Sigh.
That fixes it -- thanks!

Thanks for testing. I'll post a corrected patch to Mr. Dickson in a bit.

--
Chuck Lever
chuck[dot]lever[at]oracle[dot]com