glinc
join:2009-04-07
New York, NY | reply to milnoc
Re: Fail exactly, they should have all hdd's set up with raid-1 |
|
| said by glinc:exactly, they should have all hdd's set up with raid-1 I've seen a RAID 1 controller puke and take both drives out. Or even worse, corruption on 1 drive immediately gets replicated to the other.
And, yes, I've seen even RAID 5 arrays go tits-up.
Of course people need to realize that even with tape backups, a full system restore of lots and lots of TB pretty much never happens. Where I work we have many, many (90 maybe) TB of data. It has been calculated that a full database restore from tape would take almost a month. I forget if that takes into account the validation work or not. Needless to say we have a redundant system somewhere else.
However, I wonder where it was said it was a SAN - the articles I've saw said "server"....
Working in IT for as long as I have, I've seen it all. Most likely what happened this time was there was:
1) No $$ for that spare SAN.
2) The risk was supposed to be so low as to not cause an issue.
My bet is on #2 - I've seen way too much of those types of things cause issues. Including work on 1/2 of a fully redundant SAN take the whole thing down where the vendor is still scratching their heads years later. |
|
 FBGuyyippee ki yayPremium
join:2005-03-19 | reply to glinc
even better would be to have several different sites all doing the same thing. and then have backups of those. |
|
bjbrock
join:2002-10-28
Mcalester, OK | reply to itguy05
Properly configured RAID1 uses 2 controllers. |
|
|
|
 tmh
@verizon.net | reply to itguy05
said by itguy05:And, yes, I've seen even RAID 5 arrays go tits-up. We had a 16 bay Promise VTrak die after a power surge (yes, it was on a UPS). All but one drive went south, so even RAID 5 was useless.
We did have backups fortunately. |
|
