Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
Hi All,
First post, so please excuse me. I am running a system where we will
have a variable # of NFS/RAID servers (1-5), each running redhat 9 with
a 2.6.x kernel. I have drbd working on a preliminary set-up between just
2 of the boxes, which is fine for now. Previously in our system we did
not use drbd, and simply had tasks write to a single RAID, and
everything went fine, yet was not very redundant. Now I am running tests
that emulate our normal system which definitely loads the RAID as far as
writes and I/O is concerned. My tests were working fine the first time,
but since then I am getting kernel errors each time I try to re-run the
test. They appear to have something to do with tcp and my network
adapter (e1000's), but I have not captured those yet for posting.
Our system setup is to have a RH9 2.6.x box serve out RAID
partitions provided from a Chaparral box via Fibre, with another fibre
connection between the actual disk array and the Chaparral.
I know the setup for heartbeat and drbd preaches a serial cable for
heartbeat and a cross-over cable for synching, but that is not feasible
in my situation. Also, does anyone know how to handle the extra boxes
(beyond 2)? Perhaps I am too much of a drbd/HA-linux newbie, and
shouldn't be posting here, but all assistance is greatly appreciated. If
you need any more specific information I can probably provide that. I'm
using the latest drbd (I think 0.7.3).
Larry Anderson
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.linbit.com/pipermail/drbd-user/attachments/20040824/3ccfb9d2/attachment.htm>