[Krillin]

# We now use RAID.

Being an IT Professional over 10 years, I can honestly say RAID is NOT a good way to go for reliable backups. Why? If the 2nd drive is starting to fail, you NEVER know until the main one fails. I have seen clients loose to this scenario too many times. I do not care what kind of software you use to monitor. There isn't a good piece of software which monitors RAID. Your best bet for backup protection is BackupExec with a tape drive big enough to backup your data to take into consideration of future growth. There is more to this than meets the eye. It can be costly. But it is an investment that prevents this type of embarrassment.

I knew something was up when I went to go look at the status of CSSDM Crashing Insecure Server's kticket and the website was not found. I have not tried it lately but will try again and see if my bug report is missing.

But above is my two cents and my professional offerings to avoid this type of embarrassment in the near future and a heads up that RAID just doesn't work until it is too late.

[8088]

I don't think that anyone, including BAILOPAN, will consider RAID to be a backup measure. Also, any backup is worthless unless it's stored off-site. Regardless of the medium and used software.

[dataviruset]

I would prefer RAID to not have everything backed up off-site.

[devicenull]

Originally Posted by Krillin Being an IT Professional over 10 years, I can honestly say RAID is NOT a good way to go for reliable backups. Why? If the 2nd drive is starting to fail, you NEVER know until the main one fails. I have seen clients loose to this scenario too many times. I do not care what kind of software you use to monitor. There isn't a good piece of software which monitors RAID. Your best bet for backup protection is BackupExec with a tape drive big enough to backup your data to take into consideration of future growth. There is more to this than meets the eye. It can be costly. But it is an investment that prevents this type of embarrassment. I knew something was up when I went to go look at the status of CSSDM Crashing Insecure Server's kticket and the website was not found. I have not tried it lately but will try again and see if my bug report is missing. But above is my two cents and my professional offerings to avoid this type of embarrassment in the near future and a heads up that RAID just doesn't work until it is too late.

RAID in this case is not being used as the primary source of backups. It's being used so that one drive can fail, and it can be replaced without having to reconfigure everything.

So, lets say the server has two drives in RAID 1, and one fails.
Worst case: The other drive fails before the first can be replaced, and everything needs to be restored from backups.
Best case: The failed drive is replaced, and the system automatically rebuilds the array with no data corruption.

Let's say you have one drive, and it fails. Now you are faced with restoring everything from backups.

Given that the worst case with raid is no worse then the worst case without raid, I don't see any reason why it would be a bad thing.

RAID is not backups. It won't protect you from rm -rf / , but it can save you a ton of work.

[BAILOPAN]

I think everything that can be said about hardware or software here has been brought up in this thread, so let's let that discussion die. As much as I appreciate it, it's very noisy. Our problems have targeted questions and these for the most part have been answered privately or on IRC. We have all the data, and it's just a matter of time to make some decisions.

A quick status update is that the dead drive is heavily corrupted, so the recovery lab is taking a while. They think they'll have something this week. Based on those results I'll make another news post.

I'm also still working on the 500 errors.

[SeToY]

Do you know what you're going to pay for the recovery of the dead hdd?

[BAILOPAN]

Another update. They called to say "early next week", so looks like we might be another five days without bugzilla or so.

SeToY: It'll depend on how much they manage to recover.

[Spoony]

well, very intresting..

BAILOPAN, good luck with the recovering

[Dygear]

Wow, that really sucks man! Glad you got most of it back tho! This could of been huge if you did not have even the partial backups.

http://www.codinghorror.com/blog/200...eness-day.html
http://www.codinghorror.com/blog/200...-strategy.html
http://www.codinghorror.com/blog/200...be-afraid.html
http://www.codinghorror.com/blog/200...e-problem.html

[realchamp]

For 140$/month I can get a i7 hexcore with 3.33GHz and 24GB RAM and 100mbit/s with 5tb trafic/month..

http://www.hetzner.de/en/hosting/pro...ootserver/eq10

Just saying!