SCSI RAID and system keep crashing

[Walter Biscardi]

Which OS are you running?

Walter Biscardi, Jr.
Biscardi Creative Media
HD and SD Production for Broadcast and Independent Productions.

STOP STARING AND START GRADING WITH APPLE COLOR
The new Color Training DVD now available from the Creative Cow!

Read my Blog!

[Lisha Rigney]

OS 10.4.11

[Lisha Rigney]

…and the system crashed earlier today while opening FCP. I shut down the computer by holding down the power button and then turned off the RAID. Went to lunch, came back turned on the RAID, restarted the computer and the RAID will not show up on the desktop. I’ve restarted the system 4 times and the RAID will not show up on the desktop.

Peace,
Lisha

[Bob Zelin]

Lisha,
you will not like my smart ass answer.
Why would you continue to invest in a SCSI drive product in 2008, when inexpensive SATA and Firewire products can outperform a SCSI array ? On top of this, I never heard of the manufacturer of your SCSI array – it is not from a “video standard” company like Medea, HUGE Systems, Storcase, Rorke Data, Infortrend, etc. I am sorry that you REPLACED all 5 drives.

I have worked on SCSI arrays for many years, and have seen the best of them become worthless pieces of junk. Your MAC is fine, and the ATTO SCSI host card is fine. There are countless reasons why your drive can be failing. None of these comments below will help you, and the manufacturer of your chassis will deny anything I say anyway – you should have purchased a SATA or Firewire array from any of the manufacturers that you see advertising right here on Creative Cow.

SCSI arrays can be very fickle. I must assume that the SCSI ID’s of each drive have been set correctly at one point, because you said that it was working. I ASSUME that this is a JBOD, and not a box with parallel ATA drives with a SCSI interface. I ASSUME that you are using the Apple Disk Utility to stripe all 5 of these drives in a RAID 0 configuration – is this correct ?

No matter what you have been told, you either have a BAD SCSI DISK DRIVE, a BAD POWER SUPPLY IN THE CHASSIS, or a BAD MOTHERBOARD (or ribbon cable that interconnects the SCSI drives to the buss connector to the 68 pin interface on the back of your chassis). The only thing that you can do is to DELETE THE RAID GROUP, and test each drive out ONE AT A TIME.
When you delete the RAID, you will see 5 drives, which you can initialize (erase) one at a time. If you don’t see any drives, it’s your SCSI cable, or ribbon cable (or motherboard) in the back of the drives in the chassis. If you DO see 5 drives, erase each drive, and TEST EACH DRIVE one at a time. One will probably fail – this is your bad drive.
You have no ability to replace the ribbon cable (or motherboard), and you have no ability to test the power supply of your chassis – these are the responsibility of the manufacturer of your product. I am sure that they are anxious to blame EVERYTHING ELSE except their product. And remember, it is important to test the drives, because one bad drive can pull your whole system down.

SCSI is over, you should have purchase a modern product – one that is advertised right here on these user forums.

bob Zelin

[Bob Zelin]

Lisha –
I just went to the Caineng website. You do not have a JBOD – you have a SCSI chassis with a RAID 5 host controller. You will not be able to see the individual drives – the controller in the CAIN chassis controls the RAID 5 group, and the ATTO UL4D card sees the chassis as a single volume – the RAID 5 group is controlled by the hardware inside the chassis – the tests I just described will not work.

You have a problem with this chassis. I bet you $$$ that Cain does not even have a MAC with an ATTO card to test your system with. I dont’ know how you wound up with an outdated product like this, but I am sorry. These were the types of products that were made by HUGE and Medea at one time, and they are now obsolete. You are going to have to contact Cain Engineering, and complain to them, and tell them that your system is hanging up when trying to access their drive array.
They CANNOT do diagnostics unless they REINITIALIZE your drive array, in which case you will lose any media that you have on there right now.

for them to say “hey lady, we plug in our PC, and it tests out fine” means ABSOLUTELY NOTHING. They need to do comprehensive tests with this chassis – your dealer (was it MAC Hollywood) can’t do this for you.

bob Zelin

[Lisha Rigney]

Hi Bob,

Thank you so much for the detailed explanation!!!! It was a breath of fresh air to read!

In Oct of 2005 we purchased a G5 and RAID to accommodate a shoot using DVCPro50 and multi-cam/16-up. Moviola in Los Angeles provided the gear for the shoot, pushed the DVCPro50, AND suggested 2 outside techs to help set up the system. One of the techs convinced me that CAEN had the best product for the dollars. That’s how I ended up with the RAID… 🙂 I hate the thought of buying a new RAID after purchasing one 2 1/2 years ago. Or is that par for the course?

At one point I took my entire system to the CAEN office and set it up. After watching it crash while transferring files, rendering and during general editing they said it was a bad drive. As I mentioned in my first post, we purchased all new drives based on CAEN’s diagnosis. Since then, I’ve complained and they’ve told me they can’t do anything else/don’t know what else to do. When you say “comprehensive test with the chassis”. Can you give some examples?

I really appreciate your input!!!!!!!! Again, thank you so much!

Peace,
Lisha

[Bob Zelin]

Lisha,
you will not want my explanation. You response may be “it used to work 2 1/2 years ago, why can’t they fix it”. I used to use StorCase SCSI arrays – they were the best SCSI arrays on the market. Towards the “end of SCSI”, they came out with a product called the DS570 (another hi end SCSI company – Infortrend had the IFT-6310, another similar hi end product). Today is March 2008. All the “incredible, wonderful” StorCase DS570’s are dead, and StorCase is now out of business. Infortrend is still in business making wonderful Fibre Channel arrays, but all the IFT-6310’s are dead also. These were both indestructable RAID 5 SCSI arrays. (The Infortrend was also known as the AVID MediaRAID 8). All the AVID products from this line are also dead, with no support.

At some point, you have to say “that’s it” – there are WONDERFUL SATA arrays on the market today – from everyone from Dulce, Maxx Digital, Cal Digit, Sonnet, Ciprico, Rorke Data, and others – all that cost a fraction of what you paid for your SCSI array, and all that offer more storage, at a fraction of the price, with no headaches. The newest solutions from Maxx Digital, Dulce, Cal Digit and Sonnet are all RAID 5 as well, but with simplier SATA technology. Would this work better for you – YES. Should you have bought a new one of these, instead of spending money re-tooling your old array with NEW SCSI drives – OF COURSE.

As I said, you could STILL have a dead drive, but it could be the motherboard (or ribbon cables behind the SCSI drives). Most people would say “why bother, just get a newer, cheaper SATA array). But, it’s not my money – that’s up to you. Its easy for me to talk, when it’s not my money.

Please remember that the Apple XServe RAID is a much newer product than your old CAEN SCSI array, and the Apple XServe RAID is now obsolete, and is breaking down everywhere. But those that spent $14,000 for a new one last year don’t want to hear this.

I don’t know CAEN engineering, but the way to solve this issue may not be worth the labor and man hours from CAEN to do it – someone needs to KEEP THE CHASSIS there, and test it over and over (called a burn in) to see if it works, and keeps working. They will have disk drive exercise programs to do this – it can take a LONG TIME, and it may not be worth it to them. (or you).

Bob Zelin

[Lisha Rigney]

Okay… One more, or two more questions. 😉 I noticed that the stand-offs for the SCSI connection on the RAID are different sizes. Meaning one is slightly higher than the other. Would this make a difference? Overtime would gravity take over and the connection become insecure? And the second… I just started my computer and still no RAID on the desktop. The busy light on the RAID came on after about 10 minutes. Any thoughts on why the busy light would come on? It’s still not showing up on my desktop.

Thanks Bob and everyone!

[Lisha Rigney]

I turned off my computer. As is usually the case, when the RAID’s busy light comes on, the system will not shut down unless I push the button on the front of my G5. I disconnected THE NEW SCSI cable just in this morning from ATTO and reconnected it. This time I used the stand-off side that is higher as a guide. I screwed in the side that is higher first and screwed in the other side only to the level of the higher stand-off. I restarted the RAID and the computer and the RAID is showing up on the desktop. First things first, I’m backing up…

[Bob Zelin]

Lisha –
I am going to give you my professional advice. Please do not be offended by it (I offend so many people !).

You have been working on this problem, in desperation since February 28th. I don’t know how you can stay in business for being down for so long. I am sorry that you spent so much money for NEW SCSI drives, and you still have a “bum” system.

Lisha – this is my advice to you. GIVE UP. Go out, and buy a nice, inexpensive, hi performance SATA array from one of the many companies that are advertising on this user forum. Companies like Dulce Systems, Cal Digit, Maxx Digital, Sonnet Technologies, G-Tech, Pro Max, and others. These NEW products will outperform your SCSI array, without any aggrivation (you will have aggrivation with them in 3 years, then it’s time to change again).

As I stated before – I used to install LOTS of StorCase and Infortrend SCSI arrays – JUST LIKE YOU HAVE THERE – they were SCSI interface, with a RAID controller. They worked great for a few years, and then they FAILED. And no matter what when into them (Storcase had a big warantee, and kept changing out parts), the failure was EXACTLY what you are seeing now – the drives hang up, the busy light comes on, it takes forever to boot. You crash when the computer finally boots anyway.

You are saying right now – “I spent all this DAMN MONEY on this junk, and it used to work, and the idiots at the company just charged me all this money to replace all the DAMN DRIVES, and it STILL doesnt’ work – but it used to work – why can’t it be fixed”. Lisha – at a point, you have to give up. This is not a classis car (which has the same problems). My friends with old Porche’s have the SAME issues – it used to run, and they put all this money into it, and it STILL stalls, and has new problems all the time.

YES, certainly, the cables must be seated properly, and the 68 pin connectors must go ALL THE WAY IN, or it won’t work, and YES, you should check the 68 pin SCSI cables to make sure that none of the pins are bent, because this will show the same failure results. Your SCSI terminator may be defective, etc. etc. etc.

If you had purchased a Medea, a HUGE, a StorCase, a Rorke, you would experience the same problem – MOST PEOPLE DO (and some get lucky, and they are still working). Disney Public Relations had all their footage on StorCase DS570 RAID 5 protected SCSI arrays, and while they were in business, StorCase repaired the unit for free, and 3 months later they went out of business, and the DS570 died, and THEY LOST ALL THEIR FOOTAGE.

It’s March 10th – call your losses, buy a SATA array from one of the companies that you see advertise on these forums (except Lacie), and you will be ok, for very little money.

Bob Zelin