Let me start by saying I have read the manual. I have a question that is not necessarily related to a failure of SME 8, I'm looking for best practices and I don't see anything I can understand that describes my issue.
I understand SME uses software RAID and I don't have a preference between a RAID controller and the software RAID in SME. I have a Dell PowerEdge R2970 with a PERC 5/i Integrated RAID controller. The server came that way and it is a hardware controller. What I decided to do was define each of my drives as individual virtual drives configured as RAID 0. This was my compromise because I can't pull the PERC card and plug into a motherboard sata connector and this configuration is more native to how the software RAID is working with SME - SME manages redundancy. I can't use smartctl because it's incompatible with the RAID controller which I don't like.
With my current config, I expected to have a RAID 5 install with 1 hot swap. This is what I get from mdstat.
# cat /proc/mdstat
Personalities : [raid6] [raid5] [raid4] [raid1]
md1 : active raid1 sda1[0] sdb1[1] sdc1[2] sdd1[3]
104320 blocks [4/4] [UUUU]
md2 : active raid5 sdd2[2] sdc2[3](S) sdb2[1] sda2[0]
975482368 blocks level 5, 256k chunk, algorithm 2 [3/3] [UUU]
unused devices: <none>
I can see from the above output I have 2 channels and it looks like I have RAID 5 + 1 just like the manual says I should. Unfortunately, I don't have a ton of experience with software RAID and would like someone's opinion. Does this configuration seem optimal for SME 8? I have read
http://wiki.contribs.org/Raidhttp://wiki.contribs.org/SME_Server:Documentation:Technical_Manual:Chapter5My next thought was I could just use the RAID controller to do all the work. It has the ability to monitor for failure and swap a drive as required. I will also regain the ability to check SMART data even though I have to down the server to do it. I am looking for fast, fast, fast.
What I don't want is a configuration that exposes me to risks I can't see right now. The reason I am moving my SME server to new hardware is because the old one has a punctured stripe. I was unable to properly diagnose the failure and ended up replacing all the drives just to have the same problem. To SME the array basically looks like a single drive and SME uses it like a single drive. I was wondering if it was possible that the puncture happened because I have it configured this way?
Does anyone have any experience with this type of configuration or any advise?