[Gluster-devel] Feature request(s)

[Gareth Bult]

>why do you have to restart glusterfsd? 

Well, of course I would rather not restart it .. however I change the client configs all the time at the moment, and the server configs a number of times per day .. (to test performance etc ..) 

I was also having to restart as a result of fuse/gluster crashes, but the combination of glfs7 and tla seems to have calmed this ... 

Gareth. 


----- Original Message ----- 
From: "Anand Avati" <avati at zresearch.com> 
To: "Gareth Bult" <gareth at encryptec.net> 
Cc: "gluster-devel" <gluster-devel at nongnu.org> 
Sent: Tuesday, January 1, 2008 7:41:16 PM (GMT) Europe/London 
Subject: Re: [Gluster-devel] Feature request(s) 






Couple of very small points that would be really useful; 

a. Could mount.glusterfs be made to take a list of target hosts rather than just one? 
(so a machine could successfully mount a glusterfs so long as "a" valid node was available, rather than a specific node?) 

I see this necessary too. noted it down. 



b. Being able to specify the size of an IO cache (and other parameters?) at mount time would be really useful for tuning .. 

you would want to wait for the meta xlator where such parameters can be re-configured at runtime with a /proc like interface. 



(at the moment it seems to require a glusterfsd restart which typically involves (for me) 10's of Gb's of file transfer) 

why do you have to restart glusterfsd? 

avati 

-- 
If I traveled to the end of the rainbow 
As Dame Fortune did intend, 
Murphy would be there to tell me 
The pot's at the other end.