Sam, If all the clients connect to the same AFR on a server then it works fine. If there are different AFRs that are accessing same posix/storage volumes then there is a bug which is not handled (creating a file simultaneously on two clients might return success on both the clients or locking from two clients might return success on both the clients) this bug is being fixed after which you can have multiple AFR instances, but if this bug does not bother you then you can use multiple AFR instances. Krishna On Dec 11, 2007 8:35 AM, Sam Douglas <sam.douglas32@xxxxxxxxx> wrote: > [damn email interface... tab then space caused it to send early] > > volume brick1 > type storage/posix > ... > end-volume > > volume client1 > type protocol/client > ... > end-volume > > volume client2 > ... > end-volume > > volume afr > type cluster/afr > option replicate *:3 > subvolumes brick1 client1 client2 > end-volume > > volume server > # Export the bricks > end-volume > > -- > > Would there be any issues with multiple different AFR instances > working on the same storage volumes? > > -- Sam Douglas > > > > _______________________________________________ > Gluster-devel mailing list > Gluster-devel@xxxxxxxxxx > http://lists.nongnu.org/mailman/listinfo/gluster-devel >
