Hard Drive Issue

[Scott Greczkowski]

It appears we are having a hard drive issue with one of the hard drives on our database server.

While the database is a RAID 10 setup it appears one drive is going bad which is causing the system to lockup. The data is fine the issue seems to be with the raid controller card trying to rescan the failing drive.

I am ordering a new drive to be sent to our ISP and will have it installed ASAP.

Sorry for the downtime tonight which lasted several minutes.

 

[lakebum431]

I was wondering what was going on. Glad you figured out the problem Scott.

 

[sundude90]

Raid Servers shouldn't do that when one of the Hard Drive is going bad. Even if one is going bad, it still should function. I think your Raid server as some kind of configuration problem for it to do that. Please keep up posted, Scott.

Josh

 

[yourbeliefs]

Raid Servers shouldn't do that when one of the Hard Drive is going bad. Even if one is going bad, it still should function. I think your Raid server as some kind of configuration problem for it to do that. Please keep up posted, Scott.

Josh

Yeah... isn't the whole purpose of a RAID is for seamless backup/switchover capabilities?

I'd look into getting a new controller card as well, if you said that was the thing getting stuck on the bad data. I'd figure the card would try a few times, realize the drive wasn't working, and move on to the next one.

 

[sundude90]

Yeah... isn't the whole purpose of a RAID is for seamless backup/switchover capabilities?

I'd look into getting a new controller card as well, if you said that was the thing getting stuck on the bad data. I'd figure the card would try a few times, realize the drive wasn't working, and move on to the next one.

That's exactly what I thought. We have several of them at school. We have several Apple's and a few Dell's. We have never had a problem like what your having Scott at school.

I would look into this more and see if there is a card issue as well.

Josh

 

[Smith P.]

It's pretty simple task - look into System log.

 

[Scott Greczkowski]

The log says... (these are the last two)

20090112234834 - Controller 0
WARNING - Sector repair completed: port=3, LBA=0x1A8339BD

20090113073242 - Controller 0
ERROR - Drive timeout detected: port=3

New drive has been ordered and will be installed tomorrow.

I haven't figured when I want them to install it though. I want to be able to remove the board from public view while they do it so LER's email box dont fill up with a million database errors while the database server is down.

 

[sundude90]

i still think that there is a Controller Error there. But I don't know about those Error Codes, so I can't really help you, Scott.

I would think that there is a way to bring the Board down and replace the Hard Drive without it spitting out a million database errors.

What kind or Raid Drive is it? (Model & Make?)

 

[Smith P.]

"20090113073242 - Controller 0
ERROR - Drive timeout detected: port=3"
doesn't sounds good. Should be more info there - at least HDD ID; then if it that one what went south, then you know it, if not - expect more problems soon.

 

[Hall]

i still think that there is a Controller Error there. But I don't know about those Error Codes...

A quick Google search suggests it is a hard drive issue. This log entry is from the RAID controller telling you that it detected an error and re-mapped/moved the data to a good location.

 

[sundude90]

A quick Google search suggests it is a hard drive issue. This log entry is from the RAID controller telling you that it detected an error and re-mapped/moved the data to a good location.

We know that the code means that it is a Hard Drive issue, but what we are saying that it should not freeze up or bring down the server if there is a hard drive issue with one of the hard drive. RAID servers should not freeze or crash if there is a hard drive issue.

 

[Scott Greczkowski]

You are incorrect, you would however be correct if we were in a hot swappable configuration which we are not.

The drive should be at the ISP today, I am waiting to hear from them to schedule a time to replace it tonight.

I am also thinking at the same time (before they replace the drive) to backup the databases and upgrade our software to vBulletin 3.8 Official Release as we are on a beta version. We can kill 2 birds with one stone.

I will post more when I know more.

 

[Smith P.]

Yeah, go ahead - time to get stable version of vB !