[Cluster-devel] [PATCH 2/4 Revised] NLM - set per fsid grace period

[Cluster-devel] [PATCH 2/4 Revised] NLM - set per fsid grace period

[Wendy Cheng]

This change enables per NFS-export entry lockd grace period. The 
implementation is based on a double linked list fo_fsid_list that 
contains entries of fsid info. It is expected this would not be a 
frequent event. The fo_fsid_list is short and the entries expire within 
a maximum of 50 seconds.  The grace period setting follows the existing 
NLM grace period handling logic and is triggered via echoing the NFS 
export filesystem id into nfsd procfs entry as:

shell> echo 1234 > nlm_set_grace_for_fsid

-- Wendy
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 002_nlm_grace_p.patch
Type: text/x-patch
Size: 15230 bytes
Desc: not available
URL: <http://listman.redhat.com/archives/cluster-devel/attachments/20070405/253eb376/attachment.bin>

[Cluster-devel] Re: [NFS] [PATCH 2/4 Revised] NLM - set per fsid grace period

[Olaf Kirch]

On Thursday 05 April 2007 23:52, Wendy Cheng wrote:
> This change enables per NFS-export entry lockd grace period. The 
> implementation is based on a double linked list fo_fsid_list that 
> contains entries of fsid info. It is expected this would not be a 
> frequent event. The fo_fsid_list is short and the entries expire within 
> a maximum of 50 seconds.  The grace period setting follows the existing 
> NLM grace period handling logic and is triggered via echoing the NFS 
> export filesystem id into nfsd procfs entry as:

Your patch creates a new per-export structure that is private to lockd.
It may be easier to put the grace period value into the export entry,
and check for that in nlm_fopen. It feels more natural than adding
a whole new shadow export table just for the benefit of lockd.
(Especially when you consider future extensions like being able
to search the list by dev_t or uuid or whatever)

Olaf
-- 
Olaf Kirch  |  --- o --- Nous sommes du soleil we love when we play
okir at lst.de |    / | \   sol.dhoop.naytheet.ah kin.ir.samse.qurax

[Cluster-devel] Re: [NFS] [PATCH 2/4 Revised] NLM - set per fsid grace period

[Wendy Cheng]

Olaf Kirch wrote:
> On Thursday 05 April 2007 23:52, Wendy Cheng wrote:
>   
>> This change enables per NFS-export entry lockd grace period. The 
>> implementation is based on a double linked list fo_fsid_list that 
>> contains entries of fsid info. It is expected this would not be a 
>> frequent event. The fo_fsid_list is short and the entries expire within 
>> a maximum of 50 seconds.  The grace period setting follows the existing 
>> NLM grace period handling logic and is triggered via echoing the NFS 
>> export filesystem id into nfsd procfs entry as:
>>     
>
> Your patch creates a new per-export structure that is private to lockd.
> It may be easier to put the grace period value into the export entry,
> and check for that in nlm_fopen. It feels more natural than adding
> a whole new shadow export table just for the benefit of lockd.
> (Especially when you consider future extensions like being able
> to search the list by dev_t or uuid or whatever)
>   
I'm traveling so may not be able to do intelligent responses until next 
Monday.

However, I should have made it clear that these patches are minimum 
changes to keep NFS V3 failover going due to large amount of existing 
installations and users. The real efforts and future enhancements should 
be placed on NFS V4. That's the reason most of the code in this patch 
set are stand-alone addition and I have been trying not to mingle the 
changes with main-line nfsd-lockd code.

Ideally we would like to avoid this grace period eventually if we can 
(by transferring the locks from one server to another server without NFS 
clients' involvement). 

-- Wendy