Tek-Tips is the largest IT community on the Internet today!
Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!
-
Congratulations Mike Lewis on being selected by the Tek-Tips community for having the most helpful posts in the forums last week. Way to Go!
RAID 5
- Thread starter DS4800SAN
- Start date
I think it depends on the controller card you are using check the documentation of your server/raid controller hw.
It's not really a good idea to have many drives in a RAID-5 array since it cannot survive more than one drive failure without losing the whole array
It's not really a good idea to have many drives in a RAID-5 array since it cannot survive more than one drive failure without losing the whole array
goombawaho
MIS
I wouldn't listen to the above advice. If you have a hot spare (one or more) in your array, RAID 5 can tolerate more than one drive failure without a problem.
And besides, anyone that doesn't have at least one drive ON HAND in case of a drive failure NOT counting any hot spares is treading on thin ice.
As soon as you get a failure and rebuild onto a hot spare, get that dead drive removed and replaced. This is your piece of mind. And then re-order that spare to put on the shelf.
If you live by these rules, you'll be pretty safe.
And besides, anyone that doesn't have at least one drive ON HAND in case of a drive failure NOT counting any hot spares is treading on thin ice.
As soon as you get a failure and rebuild onto a hot spare, get that dead drive removed and replaced. This is your piece of mind. And then re-order that spare to put on the shelf.
If you live by these rules, you'll be pretty safe.
goombawaho
MIS
That's not something I would worry about overly having managed a lot of servers in my day that were RAID 5. Never saw two drives go even remotely close to one another. Besides, with a hot spare on standby, you can lose two drives and not have a problem. But, choose your own worries is what I say.
Go ahead and suggest to him a different RAID level/approach and run it up the flagpole. He asked about RAID5 and I told him how to prepare for bad disks.
Go ahead and suggest to him a different RAID level/approach and run it up the flagpole. He asked about RAID5 and I told him how to prepare for bad disks.
Lets not forget some of these drives are getting a little old now, replaced a 6607 disk a while ago and whilst rebuilding another one went down. To make matters worse soon after that there was a power cut and had to shut system down, 2 more disks failed to spin up on ipl so thats a total of 4 bad disks in one day. got the in overtime tho.
goombawaho
MIS
Regarding having very old drives/servers in service. That is something that has to be addressed in a different way. It's not a technical problem. You have to get your management to agree to replace servers after a certain age due to the increased probability of failure and the decreased QUICK availability of parts/drives and support.
If you don't, now you're really talking about increasing the chance of server failure. To me, replacing a server every three to four years max is the smart thing to do to avoid these issues.
Management that says "run them until they die" is asking for just that - death. I know it's hard to get them to understand, but you have to liken it to old age in humans. Tell them a four year old server is about like a 60 - 70 year old man or a 10 year old dog. Not ready to die, but not a spring chicken either.
If you don't, now you're really talking about increasing the chance of server failure. To me, replacing a server every three to four years max is the smart thing to do to avoid these issues.
Management that says "run them until they die" is asking for just that - death. I know it's hard to get them to understand, but you have to liken it to old age in humans. Tell them a four year old server is about like a 60 - 70 year old man or a 10 year old dog. Not ready to die, but not a spring chicken either.
- Status