Netraid woes

[ray close]

Our netraid storage array was failing with intermittent drive failure, so a hp technician was called in to fix the problem. They replaced the backplane and controller card. This drive worked for a week and then 2 hard drives failed. I found that the technician had inserted 2 level 2 scssi and 2 level 3 scssi hard drives in the array (configured as a raid 5). We could not obtain level 2 scssi quickly so I replaced all drives with level 3 maxtors and re-installed everything. These drives have worked fine until I cycle power to the server, after which all drives indicate failure during bootup. I can revive it by selecting physical drive - makeonline, but I feel nervous about this scenario.

If I reset the server this problem does not occur, it only occurs on a power cycle. Any clues on how to fix this problem. Tech support have not been much help over here, they sent out the technician, cost us 3000 dollars and the problem (in a different form) still exists. I know that the array specifies level 2 scussi, and we could only obtain (on short notice) level 3 scussi. Is this the problem???? if so, will it cause more problems in its current mode of operation do you think (been running 1 month ok)??? Should we just scrap this array because of the history of problems we have had with it and buy something else????

Regards
Ray

[Martin Breidenbach]

If you paid for a repair and the problem isn't solved call HP again.

You don't say which server and netraid model you are using. Which models do you have ?

'level 2 scsi / level 3 scsi' - do you mean scsi-2/scsi-3 or ultra-2/ultra-3?

Anyway mixing drives should work. The drives are HP drives or 3rd party drives ?

You say the problem only occurs when the system is switched off and on. Do you get a message like 'firmware spinning drives' for some time ? If yes then one or some drives have a problem spinning up. In that case the drives should be replaced.

It may also help to update the firmware on all drives. HP support should have checked firmware versions - did they ?.

[Mark Young_2]

I agree with Martin. Your initial problem sounds a lot like a firmware and/or driver issue. I would be sure that the firmware on the NetRAID card and all of the HDD's is up to date. Also, you want to be sure you have the most current driver for the NetRAID card.

Have you been doing consistecny checks on the array? Your problem could also be a sign of a degrading array. If the array has been up and running for a long time without any consistency checks, your array may be on it's way to degredation. At this point I wouldn't advise doing one, as there is the possibility of causing more harm than good.

As for your current problem, again I'm with Marting. Are they HP specified drives for your server? IF not, non-HP drives do not play well in Netservers.

Mark

[ray close]

No they are not HP drives. They are 4 brand new maxtor atlas 10K ultra 3 scsi drives. When the initial 2 drives failed, none were in immediate stock and we needed replacements the next day, so these drives were purchased. The Tech said they would be okay, and we didnt have an option at the time anyway. After the repair we had 2 HP and 2 Maxtor drives in the array. The 2 HP drives failed, yet i could manually start them but knew the data would be inconsistent. The array operated for 1 day with severly degraded performance. Overnight I rebuilt replaced the faulty HP drives with the maxtors (and built a backup server just incase). Since then the array has functioned normally - untill I power cycle, after which all drives show failure, and they must be manually brought online in order to function. After being brought online the drives operate again without a problem. The spinup seems normal (about 2 - 4 seconds each) and the firmware versions were not checked. I will forward all model numbers to you tommorow once I am infront of the machine, but from memory the servers are HP LPR 550mhz, the external storage rack (12slots) has 2 bays of 6 slots allocated for each server and 4 slots each filled with 18gb hard drives.

I appreciate your help Martin and Mark
Regs
Ray

[Mark Young_2]

Hi Ray,

Well in my experience lots of people have had success in getting non-HP drives to work in their Netservers. Usually they try this in cases such as yours or when they are looking for a cheaper and/or larger solution.

The problem is that the Firmware on the non-HP drives and the HP drives will not work well together. I would suspect that is why the HP drives kept failing when you had the Maxtors in. HP does have at least one Maxtor manufactured drive that I know of, but HP writes and flashes their own firmware for these drives. In my experience, it is all about the Firmware.

On a Netserver, the most stable disk system is one in which all the componenets have HP firmware that is up to date along with the most current drivers. Throw regular consistency checks into the mix and you get pretty much as stable a system as possible. With non-HP drives it usually isnt a case of IF the drive is going to fail, it's a case of WHEN.

I'm not trying to give you a sermon, because as I said lots of people are successful with non-hp drives. You just have to be prepared for the inevitable failure.

Hope this helps. Sorry to lecture.

Mark

[Greg Carlson]

Ray,

Your solution is to put supported HP hdds in the server. SCSI 2 and 3 drives are compatable as long as they are HP drives. Netraid cards fw and drivers are written for HP hdd fw. Without the right hdd fw, you will be prone to communication errors.
Third party hardware in an OEM machine is a bad idea and ends up costing more in the long run. Third party works fine on a grey box with third party controllers.

I would suspect you can find P1166a hdds from http://www.pmount.com which are compatible 18gb hdds for most of the netservers from the LC2000 and LH3,4 until the end of the netserver line. P1169 if you are looking at common tray hdds.

Ciao,
Greg