OK, fair enough. I think I will withdraw my request to 'fix' this. If
Solaris and other falvors of Unix (i.e. Aix) behave this way I think it
would not be good to change just Linux.
The other minor change I proposed earlier was that we allow unlock
operations during the grace period, and this will be useful in clustered
NAS heads.
What do you guys think about such a change?
The main goal here would be to prevent unlock operations from being
unnecessarily delayed when a cluster node is taking over another node and
is being set to the grace
period.
Juan
|---------+---------------------------->
| | Mike Kupfer |
| | <[email protected]|
| | g.sun.com> |
| | |
| | 11/20/02 10:23 AM|
| | |
|---------+---------------------------->
>-------------------------------------------------------------------------------------------------------------------------|
| |
| To: Juan Gomez/Almaden/IBM@IBMUS |
| cc: [email protected], [email protected], [email protected] |
| Subject: Re: [NFS] Re: Non-blocking lock requests during the grace period |
| |
| |
>-------------------------------------------------------------------------------------------------------------------------|
>>>>> "Juan" == Juan Gomez <[email protected]> writes:
Juan> However, I feel it is odd to block a client for about one
Juan> minutre when it issues "non-blocking" lock requests.
But if the server goes down, the call can end up blocking for
significantly longer than one minute anyway.
Juan> I have seen that Solaris code does so but still feels odd
Juan> and it may conflict with what most programmers expect
Perhaps, but there are other expectations to keep in mind. In
particular, when using NFS, the expectation (at least with hard
mounts) is that when the server goes down, the application will simply
wait until the server comes back. Your change would conflict with
that expectation.
Mike Kupfer [email protected]
Solaris File Sharing Speaking for myself,
not for Sun.
>>>>> " " == Juan Gomez <[email protected]> writes:
> OK, fair enough. I think I will withdraw my request to 'fix'
> this. If Solaris and other falvors of Unix (i.e. Aix) behave
> this way I think it would not be good to change just Linux.
> The other minor change I proposed earlier was that we allow
> unlock operations during the grace period, and this will be
> useful in clustered NAS heads. What do you guys think about
> such a change?
Until the server knows that the client has finished re-establishing
all state information, it should not accept any requests to modify
that state. Unfortunately, the only way it has of knowing that the
client is done is the grace period expiration...
Cheers,
Trond