[Linux-cluster] lock_dlm but local flocks = true?

[frank]

Hi Steve, thanks for your answer
but I have not put the "localflocks" mount parameter anywhere. Look at 
"gfs_tool df" output:

# gfs_tool df /mnt/gfs
/mnt/gfs:
   SB lock proto = "lock_dlm"
   SB lock table = "H-N:gfs01"
   SB ondisk format = 1309
   SB multihost format = 1401
   Block size = 4096
   Journals = 2
   Resource Groups = 200
   Mounted lock proto = "lock_dlm"
   Mounted lock table = "H-N:gfs01"
   Mounted host data = "jid=0:id=196610:first=1"
   Journal number = 0
   Lock module flags = 0
   Local flocks = TRUE
   Local caching = FALSE
   Oopses OK = FALSE

it says 'Mounted lock proto = "lock_dlm" ' because that is what I did. 
So why is it using "local flocks"?

Frank
> Date: Wed, 23 Dec 2009 13:15:46 +0000 From: Steven Whitehouse 
> <swhiteho at redhat.com> To: linux clustering <linux-cluster at redhat.com> 
> Subject: Re: [Linux-cluster] lock_dlm but local flocks = true? 
> Message-ID: <1261574146.2219.4.camel at localhost> Content-Type: 
> text/plain; charset="UTF-8" Hi, On Wed, 2009-12-23 at 13:24 +0100, 
> frank wrote:
>> >  Hi,
>> >  I'm running RHEL 5.4 with a GFS1 filesystem (over a SAN)
>> >  After checking the "ping_pong" test on the gfs filesystem I noticed
>> >  counters was too much high, like doing the test with a local filesystem.
>> >  But gfs filesystem is mounted with "lock_dlm" option. The only strange
>> >  thing I see is in "gfs_tool df" output because it shows "Local flocks =
>> >  TRUE" instead of "Local flocks = FALSE"
>> >  
>> >  How is this possible? Can I change this in some way?
>> >  Thanks for your help.
>> >  
>> >  
>>      
> You need to avoid specifying the localflocks mount parameter if you want
> the flocks to be true cluster locks,
>
> Steve.
>    


-- 
Aquest missatge ha estat analitzat per MailScanner
a la cerca de virus i d'altres continguts perillosos,
i es considera que està net.
For all your IT requirements visit: http://www.transtec.co.uk