[Gluster-devel] HA Roadmap Question...

Victor San Pedro vsanpedro at bioalma.com
Tue Jul 8 08:51:37 UTC 2008


I would like to ask one question about your road map. To be more precise
about *1.5* *HA* "*brick hot-add/remove/swap* - /live storage hardware
maintenance/"/ /item.

As I asked you some time ago, our aim is to manage big size files that
would be open once and remain in this state the rest of time.
When we turned off one of our gluster file servers (A Server), the other
included on the AFR (B Server) was able of continuing running the file.
But, on following... If (A Server) would be turned on again and (B
Server) would be turned off instead, the system crashed because the file
lost the open file descriptor.

"*brick hot-add/remove/swap* - /live storage hardware maintenance/". Has
this item meant that the lost of the open file descriptor would be
resolved in version 1.5 , or  the meaning would be that other storage
server or volume would be added without restart  the client glusterfs


P.D: At the end I managed to have gluster works with STRs over an AFR.
Anyway, I think that Unify over AFR configuration is the one that better
fits our needs. Nevertheless, I will make some other further tests in
order to make a final decision. Thanks again for your help and your

More information about the Gluster-devel mailing list