Film-Tech Forum ARCHIVE: DOLBY 100 HARD DRIVE

This is topic DOLBY 100 HARD DRIVE - ERRORS in forum Digital Cinema Forum at Film-Tech Forum ARCHIVE.

To visit this topic, use this URL:
https://ft-forum.com/ft/cgi-bin/ubb/ultimatebb.cgi?ubb=get_topic;f=16;t=003666

Posted by Allan Barnes (Member # 5178) on 04-29-2019, 09:10 PM:

I have two older hard drives (500gb / 1tb) on different Dolby 100 servers.... I am seeing "high" error rates. Is this a sign of a dying drive? And therefore should be replaced. There is NO problems on screen.

Posted by Peter Foyster (Member # 9599) on 04-29-2019, 11:02 PM:

No doubt you will get a highly technical explanation, however I thought I should mention my own experience.

I had a very similar problem late last year (Dolby DSP100 Show Player stops suddenly whilst playing trailer 01-19-2018).

The system showed a “small” error rate which I just ignored at the time and despite my attempts to blame content being responsible for the Show Player stalling it came down to bad sectors on one or more hard drives on the DSS100.

By the time I ended up doing something about it I had lost one drive and another was showing signs of failing during a screening. I was only just able to get through the night.

I purchased four new hard drives (Western Digital WD1005FBYZ) and replaced all four drives in the raid, one at a time, over a couple of days.

I upgraded to the latest available software and have just completed the summer outdoor cinema season without a problem.

The response and assistance I got from the guys on Film-Tech was nothing short of amazing.

Posted by Allan Barnes (Member # 5178) on 04-30-2019, 01:07 AM:

Thanks Peter for the advice.

Posted by Marcel Birgelen (Member # 6801) on 04-30-2019, 04:41 AM:

The sign of a dying drive is not so much a "high" error rate, but an increasing error rate.

Magnetic hard drives are designed to cope perfectly with a limited amount of errors, so a certain error rate is to be expected. Without that ability, not a single hard drive would last.

There are limitations though. While drives are able to remap data from a damaged part of the hard drive surface to another, reserved part, the reserved parts for reallocation will eventually run out. When that happens, your drive essentially cannot cope with any more hard errors.

In any case, you should look at hard drives as consumables though. Although they can last quite a while, chances of them failing increase significantly over their age. Opinions about when it is time to change drives prematurely differ, but I consider a drive up for replacement when it hits the 3 year threshold. Most drives will last longer than 3 years and it also depends on the amount of usage that's on those drives, but you'll see an increased failure rate, so I consider anything after 3 years as borrowed time.

The good thing is that hard drives are one of the cheaper components to replace in your average DCI system.

Posted by Jack Ondracek (Member # 1466) on 04-30-2019, 09:27 AM:

I had that problem on one of my Doremi servers. "Beauty and the Beast" froze twice, like someone had hit the pause button. Our service people said the same thing about increasing error rates.

We found the drives were available locally, so it was easy to get ahold of them. We replaced the drives (3 each), taking the opportunity to note how long the old ones lasted, and doubling the size of our arrays at the same time. It cost about $300 (US) to do each of our servers.

Hard drives are now on our maintenance / replacement list, along with lamps and UPS batteries.

Posted by Jim Cassedy (Member # 4115) on 04-30-2019, 09:52 AM:

I sent you a PM with a copy of Dolby's 'official' list of
approved replacement hard drives for your serverthingy.

Posted by Marco Giustini (Member # 4544) on 04-30-2019, 11:10 AM:

What exactly is this ‘error rate’ that you see? And where do you see it?

Posted by Jim Cassedy (Member # 4115) on 05-01-2019, 12:06 AM:

quote: Marco Giustini
What exactly is this ‘error rate’ that you see? And where do you see it?

I'm not sure about the DSS-100, but on the DSS-200, under "THEATRE
DEVICES" in the "SYSTEM" tab, you can get a 'bad' (reallocated) sector
count, which gives some idea of the state of the drives in the RAID.

Here's a screenshot I took on a system I was troublshooting last year

Posted by Marco Giustini (Member # 4544) on 05-01-2019, 12:44 PM:

That-s what I think but the SMART table also shows an "error rate" which could be meaningless on some models of HDD.

Posted by Karl Belter (Member # 10813) on 05-02-2019, 09:10 AM:

This is actually how I got involved originally with our theater. You want to deal with this soon. If you don't, you run the risk I ran into: a punctured array. RAID 5 stores the data + check value spread across all the drives. Lose 1 and your OK. A punctured array is where you have a bad spot on 1 drive, and near the same time, another drive totally fails. Now for that one bad spot your missing 2 parts, so you cannot ever get it back.

As others have said - swap out drives 1 at a time and rebuild to get all clean drives.

If that don't work - it is re-install time.

Posted by Brian DeCiancio (Member # 1422) on 05-02-2019, 11:08 PM:

I've replaced drives in plenty of DSS100 (I have 9 in use) and always go for increased size and a fresh install of the OS when drives fail.

Besides the obvious benefit in reliable/new drives, boot up time is lessened and inconsistent errors are eliminated by the clean install. Just be sure to have all content handy or on another network connected server.

What is the drop-dead date on these servers? The certificates will expire ?sometime in 2011?, correct?