0

All my firmware is up to date on the server and the controllers.

Storage crashed over the weekend. I rebooted it and it detected that I put in two new disks last week (I did check that both disk completed the rebuilding process last week). After it booted into the OS I see that it gave me an information message. After 18 hours it is at 54% so it is looking healthy. But I need to replace 5 more disk in the msa. Should I wait for this message to finish before replacing more disks?

785 Background parity initialization is currently queued or in progress on Logical Drive 1 (15.0 TB, RAID 5). If background parity initialization is queued, it will start when I/O is performed on the drive. When background parity initialization completes, the performance of the logical drive will improve.

lbanz
  • 1,609
  • 5
  • 20
  • 30
  • 1
    `15.0 TB, RAID 5` what. RAID 5 gives you *one* disk of parity. If you lose another disk while the RAID is rebuilding, well, I hope you have good backups. Yes, wait, or even better, change the RAID level to 6 or 10. – tombull89 Oct 01 '12 at 09:29
  • Yes, this is a very old msa that was put in years ago. Got new replacement hardware but I need to keep it running for maybe a few months until we have the time. We will most probably have 30TB with raid 6. – lbanz Oct 01 '12 at 09:43
  • A single 30TB array? Or several RAID6 arrays that will give you 30TB combined? – pauska Oct 01 '12 at 10:20
  • Haven't decided yet At the moment we have 15TB raid5 and a 19TB raid6 which makeup the 30TB+ combined. But we know we will use microsoft DPM to backup the 30TB and 40TB is the max it can support. – lbanz Oct 01 '12 at 11:14
  • Which model MSA is this? Please include those details. – ewwhite Oct 01 '12 at 20:37

1 Answers1

2

Oh boy.

You should really really learn about the technology you're using BEFORE trying to do maintenance/recovery on it.

RAID5 consists of one parity partition alligned over all disks, which means that you can lose 1 disk and still have data available, while rebuilding a new disk (which should be a hot spare). Remove another disk while this is happening and your data is potentially lost.

Second of all, you have a 15TB RAID5 set, which I'm pretty sure is 1TB S-ATA disks? You're looking at a long, long, long period of time if you're going to replace 5 more disks, as the raid will need to rebuild completely every time you replace 1 disk (and remember - 1 disk at a time, wait for rebuild, then another disk and so on).

This is not the normal way of doing things. If the SAN is so old that you need to replace all disks then a new SAN should be purchased to migrate data to.

pauska
  • 19,620
  • 5
  • 57
  • 75
  • I replaced one disk at a time. I checked that it finish rebuilding before I put in a new disk. This server / msa was put in by a software vendor before my time and no one was really in charge of it until it kept falling over. They are all 1TB drives and we bought a new SAN. But don't have time to move them yet. As the user have a deadline to meet and they need to extract data off it so we will continue to put in new disks until we have the time to replace the entire storage. – lbanz Oct 01 '12 at 09:35