RAID Question
57 minutes ago, HawkJizbel said:@Oshino Shinobu @Quindor @Eniqmatic. Hey. Sorry guys. I didn't explain it well. So this is what is going on. We need a main file server with samba access for around 50 to 70 users. Raid 1 was just an example. I am not sure which raid to use for this situation. The storage will be around 6 to 8 TB. I want the primary server to have multiple redundancy (User data and data). I will also be setting up another local and a remote freenas server for backup.
What i also wanted to know is if raid is configured for redundancy, will raid take care of automatically redirecting the users from the failed drive to a working drive(While the system is running)? Or do we have to restart the system in order for the redirection of data to work?
Ah yes, that explains a lot more.
When using RAID, all the disks that are part of the RAID sort of become one. After that 1 (or more depending on type of RAID level) disks can fail and the volume as a whole (the RAID) will remain active like nothing happened. You replace the disk, it rebuilds (or resilvers in the case of ZFS) and you are back to your original redundancy! From a usage standpoint, you don't notice anything, at most you will take a temporary performance hit (With RAIDz1/z2, with mirrors only during resilvering).
When looking to build something for 50 to 70 users you need to take a lot of things into account. For instance, is 1Gbps enough? It will depend on what they will do with it. Word/Excel usage, sure, no problem. But if they are working with bigger image or video files sometimes, you need to start thinking about 10Gbps connections (for the server).
Then also the hardware becomes more important. For that amount of users you either need a bunch of 7200rpm disks and a decent cache or you need to think about 10K/15K disks (although I would just skip those now-a-days and directly go to SSD's). It all depends on budget. As a general rule of thumb, think about having 10% of your total volume size in SSD cache size when using cheaper 7200rpm disks.
Also, depending on chassis, it's better (when using RAID1/mirrors) to have 16x1TB (8TB effectively) then 4x4TB (8TB effectively). Although you would have the same amount of space, 16x1TB will give you much more I/O then 4x4TB would. Cache and such is great but it can never make up for actual disk I/O in the end.
What kind of hardware are you looking at and what is your budget? For this, something like a SuperMicro server would be perfectly suited. They also have the correct type of LSI HBA's, etc.
For ZFS CPU is somewhat important (Again depending on RAID level, compression (LZ4 highly recommended!, deduplication (NOT recommended), etc.) but memory is the most important (get lots of it!). A remote box to do replication is a great idea, that one can be a bit less in specifications.
Create an account or sign in to comment
You need to be a member in order to leave a comment
Create an account
Sign up for a new account in our community. It's easy!
Register a new accountSign in
Already have an account? Sign in here.
Sign In Now