You have a raid 5 array and 2 discs drop out at once

looks like its gonna be a long sunday

oooh, burn.

I don't understand RAID, beyond 1, 0, stripped or mirrored, but two dropping out is that last thing you want.

Sorry your sunday is rapidly becoming your dads saggy :pants:

You have a raid 5 array and 2 discs drop out at once

looks like its gonna be a long sunday

Some years ago we had a server with mirrored discs. The primary failed at 6 a.m., so a replacement was ordered. The secondary failed at noon the same day. Full marks to IBM for consistency!

:confused:

:confused:

Raid 5 is designed to cope with 1 failed disc... so it will still run.

but loose 2 and loose the whole lot :thumbdwn:

Some years ago we had a server with mirrored discs. The primary failed at 6 a.m., so a replacement was ordered. The secondary failed at noon the same day. Full marks to IBM for consistency!

Thats the trouble , you are always advised to use identical disks but if they come from the same batch and have the same usage then the chances are they'll fail at the same time.

Not a HP NetRAID is it? I've had this happen several times on different NetServers using the NetRAID cards.

Sometimes when two HDs fail at once, one of them is only pretending to fail. If you can determine which one (or take a lucky guess), then you should be able to mark the disk back online in the RAID BIOS. You will sometimes be able to get your four disks back online and simply replace the genuine faulty one.

I would contact the manufacturer of the RAID array, and explain you think its very unlikely both disks have failed at once. They might have a secret trick for you to try.

Ross

Weeeelll

basically the two discs are part of a SAN array..... EMC if you want name n shame...

we have two san units one on main site, and a backup @ DR. we have been working today in the comms room, re-routing power so everything had to be powered down. and the discs showed up as failed when powered back up. and the hot-spare didnt have time to rebuild before the other one dropped.

we managed to recover the array, but its corrupted the data.. its also copied the corruption onto its mirror.

the problem we are faced with now, is what to do with the systems, we cant run the primary array until we find out why it went mad..

we have also just upgraded our CHES1000 link between sites to wavestream which as yet is still untested for large amounts of data.. management want to run our primary machine, on the backup array across this new circuit (after we restore data) we are a bit cautious as its still untested, and if the line goes wrong.. it will be business critical downtime..

we want to run the backup system on our DR site.. but i doubt we will get our way we wont know anything because its going to be another 2 hours before the wavestream is installed.

i need a drink :(

Unless you are very careful, I think that RAID5 can be worse than useless in that it lulls you into a false sense of security.

Many of the drive failures I see are secondary to another event, normally realated to power - PSU, UPS etc etc etc so two physical disks end up failing at the same time :(

raid 5 is great but always configure a hot spare that takes over from a failed drive ..... then when a second fails - which invariably happens within 24hrs.... you are safe.

raid 5 is great but always configure a hot spare that takes over from a failed drive ..... then when a second fails - which invariably happens within 24hrs.... you are safe.

I think thats what Col has got - but the hot spare didn't have enough time to rebuild before the second went :thumbdwn:

Chin up mate.

Raid 5 is designed to cope with 1 failed disc... so it will still run.

but loose 2 and loose the whole lot :thumbdwn:

:confused: :confused: :confused: :confused: :confused: Well that clears things up then.

I think thats what Col has got - but the hot spare didn't have enough time to rebuild before the second went :thumbdwn:

Chin up mate.

chin up? lol

its 20 past 8 and ive been here since 7AM!

have almost sorted the disc problem.. now faced with some gay Active directory issues where dns isnt looking upwards from the child domain to the parent.. this is another big issue to be sorted before i can go home

RAID5 here with a hot spare (although it's generally shut down) is a great way to do it.

Annoyingly we tend to not get enough budget for that at the office though - got it at home though

i finally gave up and got home at 10 last night.... and was in work for 8 next morning:thumbdwn:

I've been in a similar situation - one job I did lasted from 2pm Sunday tilll 10am Monday.

Hope you get it all sorted soon!