SCSI Help

[thebeephaha]

SCSI Configuration Help

I got a Perc 3/SC which I think is a LSI Megaraid Express 500, I have two Seagate U160 18.2GB drives, a Fujitsu U160 18.2GB drive, and a 5 device cable with a terminator at the end.

So for the info: I've got the controller in a spare computer and I can see the card bring up its post screen but it gives me an error and when I press Ctrl+M it doesn't enter the config utility instead it just skips it and tries to boot the machine which fails as I don't have an array setup yet. (FYI I was going to try to make a RAID0 outta these)

I have the drives as such:

Controller:

Cable plug 1: Fujitsu with ID set as 1
Cable plug 2: empty
Cable plug 3: Seagate with ID set as 3
Cable plug 4: empty
Cable plug 5: Seagate with ID set as 5
Cable terminator

I also have the controller "J1 Termination Enable Jumper" set as "Permanently enable all onboard SCSI termination." as this is the default according to Dell.

Thanks.

 

[Deleted member 3]

I've had an issue like that before with an LSI card as well. Firmware update fixed it. If it doesn't you can always configure it from DOS/Windows. Downside of my controller was that the controller screams like a ***** when the array fails. Had to boot, flash etc while being deafened.

Anyway, the error simply means the attached drives and set configuration don't match, ie it probably has some array defined which isn't there.


The cable is a U320 cable btw, quite handy to have around.

 

[thebeephaha]

Wow, firmware update fixed the issue, I can get into the config utility just fine now!

So I tweaked around with the ID's some more to make the controller happy and detect all 3 drives but when I try to initialize an array it freaks out and the alarm you spoke of came on showing the two Seagates as "fail". Then I tried to clear data on the disks before making the array and, the two Seagate drives won't work. The controller says "error" for both of them. I can configure the Fujitsu and I can clear the data on it fine. But the Seagates, if I try to do anything with them the controller just says "error". I checked the drive info in the controller for them and it thinks they are fine, it is just if I try to change the data on them they don't work.

Any more ideas? I know the drives should be good as I got them from DaedalusHelios here and he said the Seagates were previously in a RAID0.

 

[DaedalusHelios]

I had all three of these in my old server at work and they ran fine 24/7. I was using them up until 4 days before I listed them here on techpowerup. They were all static bagged and well padded in transit.

I never used that controller with those drives before though. I had a bunch of regular SCSI hard drives hooked up to that controller in Raid within Windows 2000 Server though. So it all should be compatible.

I am glad we just made a step in the right direction.



There has to be something we have overlooked............

 

[thebeephaha]

Yea, they were shipped very well so I don't think that is the issue.

It has gotta be some sort of configuration thing, I just don't know what.

I did try just using the Seagates and not the Fujitsu thinking maybe they were clashing as in the Perc info it says the Seagates are SCSI 3 160 and the Fujitsu was SCSI 2 160 but it didn't make a difference.

This is mainly a learning experience so if I don't get it working I might just pass these along to someone else but if I do get them working I intend to use em.

Thanks for the help so far guys.

 

[DaedalusHelios]

Anyway, the error simply means the attached drives and set configuration don't match, ie it probably has some array defined which isn't there.

Maybe thats still the issue?



Otherwise, I would say it could be jumper settings????

PS, I have a second SCSI controller I could send you that might sort out the issue.

I wouldn't charge you for the card of course.

 

[Deleted member 3]

You have to remove all configuration data and create a new configuration (=array)


If required I can look into an LSI BIOS tonight and make some photos of the screens.

 

[thebeephaha]

I reset the controller and deleted all the previous array information. So I dunno, when it try to make an array it works, it is just at the initialization it freaks out or if I just wanna erase a drive. I think the controller is fine as the Fujitsu drive works, I just dunno what is the deal with the Seagates.

Let me take some photos of what I am doing on the controller.

 

[DaedalusHelios]

I reset the controller and deleted all the previous array information. So I dunno, when it try to make an array it works, it is just at the initialization it freaks out or if I just wanna erase a drive. I think the controller is fine as the Fujitsu drive works, I just dunno what is the deal with the Seagates.

Let me take some photos of what I am doing on the controller.

The last thing those drives had on them was Ubuntu Linux I think. If that helps.

 

[thebeephaha]

Here are the pics:

Drives are online, no configuration yet.

Fujitsu drive info:

Seagate drive info 1:

Seagate drive info 2:

Selecting drives for a new RAID0 array:

Array config:

Array created:

Initialization failed:

Controller shows the two Seagates "failed"

If I try to clear the data on the Seagates I get another "error"

 

[Deleted member 3]

Do those drives have any termination themselves? If so it has to be off. Only the end of the cable should be terminated.


On the bright sight, the screens seem identical to my own LSI controller

 

[thebeephaha]

Do those drives have any termination themselves? If so it has to be off. Only the end of the cable should be terminated.


On the bright sight, the screens seem identical to my own LSI controller

Does this tell you anything worthwhile?

J2 has motor start enable set, J5 has 2 on one drive and 4 set on the other, J6 is 1 for both.

 

[thebeephaha]

So I did some more goofing around and found if I put any one drive on the first connector of the cable closest to the controller, I can initialize or clear the data on that drive, even the Seagate.

Upon closer inspection it seems the connector on the second and third connectors are loose, like I can pull the rear tab off easily when disconnecting the drive. I don't know if that is enough to screw up data transfers to the other drives but this might be the issue, just a bad cable. As I have tried just about every other configuration that you guys and some people over at Hard Forum suggested as well.

If I try putting any drive past the first connector by itself it fails too, so I think this enough evidence to point at a bad cable.

Thoughts?

 

[DaedalusHelios]

So I did some more goofing around and found if I put any one drive on the first connector of the cable closest to the controller, I can initialize or clear the data on that drive, even the Seagate.

Upon closer inspection it seems the connector on the second and third connectors are loose, like I can pull the rear tab off easily when disconnecting the drive. I don't know if that is enough to screw up data transfers to the other drives but this might be the issue, just a bad cable. As I have tried just about every other configuration that you guys and some people over at Hard Forum suggested as well.

If I try putting any drive past the first connector by itself it fails too, so I think this enough evidence to point at a bad cable.

Thoughts?

I totally agree and there must be a way to get that cable back to normal.....

 

[thebeephaha]

http://cgi.ebay.com/New-SCSI-Ultra-...ryZ31494QQssPageNameZWDVWQQrdZ1QQcmdZViewItem

Would something like this work as a replacement? Cheapest one I could find.

 

[DaedalusHelios]

http://cgi.ebay.com/New-SCSI-Ultra-...ryZ31494QQssPageNameZWDVWQQrdZ1QQcmdZViewItem

Would something like this work as a replacement? Cheapest one I could find.

I am 95% sure it is, but Dan would know 100%.

 

[Deleted member 3]

Make it 100%
basically any 68pin cable can run at U160, the additional twists are for U320 speeds. Obviously running at lower speeds never is an issue.

 

[thebeephaha]

Update guys, I got a new cable and the drives work and I put them in a RAID0.

Things seem to be working okay but if I run a disk benchmark I get a read error and when I check the controller log one of the Seagate drives has some media errors. I ran a scan for bad sectors but it didn't show any. Ideas?

Things seem to be working okay, got Windows on it and such so I'm guessing it is just a glitch.

Edit:

I can run a quick benchmark with HDTach but the long bench gives the read error and so does HDTune.

The quick bench shows 60MB burst, 60MB average read with a access time of 8.5ms. (write back cache on, adaptive read ahead, cached I/O)

Seems a bit slow but these drives are fairly old I guess.

 

[DaedalusHelios]

I am glad it was just the cable.

 

[thebeephaha]

^ U got a heatware, I wanna give u good feedback for the drives u traded me and the help you provided.

 

[DaedalusHelios]

Sure, my heatware is Echelon V

If you would like another SCSI controller to see if the speed might be reduced because of the card. I have another SCSI card I could send you free of charge.

I wouldn't be able to ship it out til Monday though, as I am on vacation right now.

 

[thebeephaha]

That's quite a generous offer. I would at least like to pay for shipping though. PM me with a paypal or something and we can work it out.

 

[DaedalusHelios]

That's quite a generous offer. I would at least like to pay for shipping though. PM me with a paypal or something and we can work it out.

OK, I will send you the actual cost of shipping, after I ship it, if you would like.

PM me your address.