iliak
Contributor
- Joined
- Dec 18, 2018
- Messages
- 148
hi,
im new to the storage part of DevOps (a new responsibility at my work place)
we should receive a server that will be mostly used as read intensive jobs (approx 30-40 Gbs ) will be connected with 40gb network
we need to make 2 shared storage's :
1. 80% of the capacity to store the data files (the clients will read them)
2. 20% of the capacity as temporary scratch location( the output of the clients will be stored here for total time of one week)
what is the best configuration to do so
the reads are very have approx 30-40 Gbs (at peaks) up to 1k iops
the write are up to 1Gbs
i thought of making raid 5 (using all the 8 ssds), and make two volumes one for each.
this is not for production (that why 1 hdd fail should be enough)
what do you think?
im new to the storage part of DevOps (a new responsibility at my work place)
we should receive a server that will be mostly used as read intensive jobs (approx 30-40 Gbs ) will be connected with 40gb network
we need to make 2 shared storage's :
1. 80% of the capacity to store the data files (the clients will read them)
2. 20% of the capacity as temporary scratch location( the output of the clients will be stored here for total time of one week)
what is the best configuration to do so
the reads are very have approx 30-40 Gbs (at peaks) up to 1k iops
the write are up to 1Gbs
i thought of making raid 5 (using all the 8 ssds), and make two volumes one for each.
this is not for production (that why 1 hdd fail should be enough)
what do you think?