-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Hi all,
I''m going to start some tests with lustre. But before this i would
like to know if this scenario is acceptable, if some one can give a hand i would
appreciate.
I have 24 node in a cluster. 2 of them will connect to an storage array (fibre)
i also have another server that would access some of this data through one of
those 2 nodes. Lets call then NodeA and NodeB.
In order to have fail-over i''ve connected this 2 nodes in eth1 with a
crossover cable. The main idea is to have in NodeA a MDS server and in NodeB a
MDS fail-over. And have OSS in NodeA and NodeB maybe a active/active
configuration.
Does this scenario seems acceptable ?
I''m not sure on how i should divide the space for the OSS and MDS
servers ? should it be 75% OSS and 25% MDS ? or does the MDS server
doesn''t spend must disc space ?
Appreciate any help. Best regards
- --
Rui Ramos
=============================================Universidade do Porto - IRICUP
Pra?a Gomes Teixeira, 4099-002 Porto, Portugal
email: rramos[at]iric.up.pt
=============================================-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2.2 (GNU/Linux)
iQEVAwUBRRD+lr1uR0bdnTWSAQKp8Af/ebEmSYi4xUUtfnXX9jrqQe0bfYzlEC0M
927yo1m97ZgzvX5DAfkyOjmGAc2Mv5kcxsrKvG3f5lqwKUCTSedJHTXJapeMoAZE
PcZ523eyq6qSNBgKu/TnMRyPHzyYR0+0V74b9qwRZiVgjk2svnRpBvTsfPsPM50s
j/W7+sAMXyCG4FA1y3cDBqAary/+pRcKX+N8zWlaBpf+O4qc/Q4WRCNTZiK638oC
nFMFnbJoZMfhyMt2J0JA79V73o2KeR7LloMl+46dg3DWifQpesgZkahBEnydghjd
qGs9LJ249qiElYQ7JIeAbp6sC86ZvnS4J7Dy4ZOpomKCBP6UIWT3rw==kkV4
-----END PGP SIGNATURE-----