quote: Scott Norwood

---

Note that mixing disk sizes and types will reduce performance. All disks in a RAID group should be the same model and have the same geometry for optimal performance.

---

There's quite some discussion going on about that. Mixing different drive models can, indeed cause some performance degradation, but in the case of a playback server, that's probably not your main concern.

And in practice: Even the big storage boys like NetApp, EMC, Dell EqualLogic, etc. are often mixing vendors and models on their storage solutions (mostly because of availability and price though).

Personally, I have found reliability usually to be the more important factor than performance
 

---

Posted by Adam Fraser (Member # 1074) on 12-30-2013, 01:36 PM:
 
A little off topic, but do RAID replacement drives require any work or formatting out of the box. Can I just pull one out of the package that I order from Newegg etc. to replace a bad drive and rebuild the array?

We have WD1003 drives and have been wanting to buy one as a spare for our GDC.
 

---

Posted by Marcel Birgelen (Member # 6801) on 12-30-2013, 01:46 PM:
 
RAID replacement drives should not to be formatted before being used.

Actually, I've encountered some problems in the past with drives that have been formatted before being used in a RAID, probably because the controller did recognize that there was already some data present and didn't want to overwrite that data.
 

---

Posted by Pete Naples (Member # 761) on 12-30-2013, 03:29 PM:
 
Steve, that'll be one of those 'refreshed' servers then! I wonder if it's got a matching set of eSATA cables internally?

At that age it must now be, it's time the drives were replaced as a set, you can up the capacity while your at it.
 

---

Posted by Bajsic Bojan (Member # 4928) on 12-30-2013, 03:54 PM:
 
while on topic, DCP2000 and 3x 3TB drives are still a no-no, right?

or did anyone succeed in doing that? Am asking as i will upgrade this year.
 

---

Posted by Marcel Birgelen (Member # 6801) on 12-30-2013, 04:17 PM:
 
I guess that depends on your mainboard. There should be two versions out there: One with an Intel Server board and the other one with the Supermicro X7SBE. That last one should actually be capable of supporting 3TB disks, as they're running them in AHCI mode with software RAID.
 

---

Posted by Steve Moore (Member # 4753) on 12-31-2013, 05:27 AM:
 
Yes Pete, there were "refreshed" or perhaps just dusted when purchased 2nd hand.
I see it has 14,000 hours on the drives according to the report (though we have been running the server since we purchased it for around 5 months)

I know I swapped out all my cables in the one at the cinema I work at on a daily basis, however I have not been inside the one causing problems (a 150 mile round trip) so before I set off to look at problems (I just had them email me a server detailed report, as they were having issues) I wanted to go armed with correct bits.

Cables were on my mind to replace whilst at it. I will see how I go. At this moment in time I would not have time to swap out all drives and re-ingest. Shame this didn't happen a week ago, could have done whilst it was shut for two days over Christmas!
 

---

Posted by Pete Naples (Member # 761) on 12-31-2013, 09:33 AM:
 
Dusted on the outside, if you're lucky.

I've had several with a mixture of orange and red internal SATA cables, a couple with not two hard drives the same and often as not they haven't had all the upgrades done. How something can be 'refreshed' when it's not even had the thermal upgrade fitted is beyond me.
 

---

Posted by Dave Macaulay (Member # 813) on 12-31-2013, 10:05 AM:
 
You should use an "enterprise" drive rather than a standard drive, these are not usually stocked at the local computer stores but easy to find online. Doremi will send you a current list of approved drive models on request, other drives will probably work but you may have issues. Some otherwise good non-spec drives will just not be seen by the system, I can't explain how that happens.
If the server is still under warranty (possible with 14000 hours on the drives if it's running 24/7) then get a free replacement drive from Doremi.
Replacement drives must be "clean" - either new and unused, or with the MBR wiped/zeroed. The server might accept some as-is used drives but I haven't had any such luck: I think you can reinitialize the RAID using them but "repair" without losing content is not possible. Wiping a drive is relatively simple, though.
 

---

Posted by Amanda Mundin (Member # 3386) on 12-31-2013, 10:12 AM:
 
We have had a recent problem with a DCP 2000 saying one of the hard drives was degraded, we thought it was the hard drive that had failed at first, the drive wasn't being recognized at all, so knowing that the internal SATA cables have had problems in the past, and having a spare old yellow one we swapped the cable, rebuilt the raid and all seems well at the moment, have been playing fine for the last few days and have ingested a couple of features with out problems.
Why do the internal SATA cables stop working?
Also are they anything special or could they be replaced with any SATA cable you can buy from Amazon or similar?
Are the Yellow cables any different to the Orange, to the Red?
 

---

Posted by Steve Moore (Member # 4753) on 12-31-2013, 03:02 PM:
 
It's beyond me too why sata cables would fail, especially when they are never unplugged.

David, these servers were bought 2nd hand so not sure if they will have warranty, they are the old black intel board ones. I have bought 3x WD 2TB brand new enterprise discs for my own, when i have a quieter week and little content in use, however this server is a fair way away and ideally i just want to swap the cables out and replace the degraded drive

Pete, what is or was the thermal upgrade?
 

---

Posted by Dave Macaulay (Member # 813) on 12-31-2013, 04:01 PM:
 
The thermal upgrade applies to the black front Intel SE7221 motherboard servers. It includes a Dolphin card cooling fan with plastic air guide, and a pair of smaller fans. You remove a couple of small fans as well. The idea is to improve CPU and Dolphin cooling and to reduce fan vibration at the RAID drives. All the old servers should have had it done quite some time ago. With this kit, the airflow is pretty much the same as on a new model server.
 

---

Posted by Carsten Kurz (Member # 5396) on 01-01-2014, 08:54 AM:
 
3TB drives are still not supported on Doremis, not internal/RAID, not external for ingest. Reason is that Doremis still run on a fairly old Linux kernel version. It's actually a driver/kernel, not a mainboard issue.

I guess though, that they are already working on support for larger drives.

- Carsten
 

---

Posted by Brad Miller (Member # 2) on 01-01-2014, 02:15 PM:
 
If you want larger drives, the Dolby DSS200 is the machine. It will take up to 4TB drives and be happy as can be. It's also a hardware raid, so you don't have to put tons of money into enterprise drives. I've ran torture tests on these machines with the most mixed-matched junk drives I could get my hands on and they just keep on trucking without even the slightest hiccup. Of every server and IMS type of solution out there, nothing else has come close to the quality and reliability of this model (including Dolby's other models).

Regarding Doremi, do NOT mix and match different types of hard drives. You will have problems. Even more important, don't use any drive not on their approved list. I've experimented with this and if you want reliability, their approved list is the ONLY way to get it.

The same thing goes for GDC as with Doremi, and Dolby's "cheaper" DSS220 model.
 

---

Posted by Steve Guttag (Member # 268) on 01-01-2014, 02:40 PM:
 
I have briefly mixed/matched drives in a GDC SX-2001 system (normal mediablock) without incident.

The key is to have identical read/write speeds. The drives have to function as a single unit.

Note, Dolby also has a list of "approved" drives.
 

---

Posted by Steve Moore (Member # 4753) on 01-01-2014, 03:41 PM:
 
Thank you everyone for your input. Extra capacity is not really an issue in any of our screen, it's reliable operation. Thank you Brad for your mention of your Doremi tests with "random" drives; I think that has put me off that idea. In the short term to get this server operational again, I will opt for locating a suitable 2nd hand drive and clear its MBR - and replace all SATA cables as a matter of course.

However - as the Doremi manuals and documents from their website do not make things 100% clear; if I was just to replace the leads and run a "Repair" on the raid would that leave content in place on the drives? This might be worth a try in the short term? It's just a time issue and I would rather not have to re-ingest if possible - two reasons I would have to go over on a night/early after they close to perform this and second deluxe in the UK claim back a lot of the drives within days of receiving and we may not have the content on site.

The joys of computers... you know I never had to replace a sata lead or a drive on any of my Cinemeccanica Victoria's EVER!
 

---

Posted by Brad Miller (Member # 2) on 01-01-2014, 04:18 PM:
 
At least you won't have to change the oil, belts or gears on your Doremi.
 

---

Posted by Dave Macaulay (Member # 813) on 01-01-2014, 10:35 PM:
 
Replacing the SATA cables doesn't change the RAID, you won't lose content. Do not power up the server with the drives removed or disconnected though.
If, after replacing the SATA cables, it still shows a drive as bad... you can try "repair". It may not let you do that though; if not, and you believe the bad drive is really OK, you can wipe its MBR and try a repair with it in place.
Only "initialize" will destroy the content.
Avoid "manage", that's for messing with stuff you shouldn't be messing with... if you want to add a 4th drive to the array (possible on the 4-bay servers) I think this is where you would do that.
 

---

Posted by Carsten Kurz (Member # 5396) on 01-02-2014, 07:10 AM:
 
Well, sure enough, with these very old drives, I would at least try to source 3 new/current consumer grade drives (not the very cheapest 'green' types, of course) in order to be prepared for a fatal drive loss. Put them on the shelf. In case you're running into trouble, it will be hard to source one or two old drives to be paired with the existing ones in short time.

You could prebuild a RAID from these three drives, mark their number/RAID sequence on them, and of something happens, your server will at least be up again for ingest within a few minutes.

Remember, these drives are old. Rebuilding a RAID creates a heavy load on all drives. It is not unlikely that rebuilding will cause another existing drive to fail.

- Carsten
 

---

Posted by Amanda Mundin (Member # 3386) on 01-02-2014, 11:26 AM:
 
Here is a current list of approved drives, recently sent from Doremi:

1TB
Hitachi 0A35772 HUA721010KLA330 HT1TB-SATA
Hitachi 0A38028 HDE721010SLA330 HT1TB-SATA
Hitachi 0A39289 HDU722010CLA330 HT1TB-SATA
Western Digital WD1002FBYS-02A6B0 WD1002FBYS
Western Digital WD1003FBYX-01Y7B0 WD1003FBYX

2TB
Western Digital WD2003FYYS-02W0B0 WD2003FYY
Western Digital WD2002FYPS-01U1B1 WD2002FYPS
 

---

Posted by Steve Moore (Member # 4753) on 01-02-2014, 11:51 AM:
 
thanks everyone for your input! I think best action is replace leads, see if it improves, try a repair if allowed. Failing that replace all for new at a week that no content is required and only one film needs to go on, though i realise fairly soon.

Brad, I was changing the oil in my Doremi... Just realised that they are USB ports on the front and not oil filling holes... no wonder the drives failed. Though in my defence, the manual doesn't mention "do not fill with oil"!

(of course I am joking!)
 

---

Posted by Marco Giustini (Member # 4544) on 01-03-2014, 05:13 PM:
 
Steve,

Apparently the yellow cables were poorly made and the contacts were eventually developing a tiny layer of oxide. Doremi are replacing them with better quality ones, but in theory you could just reseat them - to clean the contacts - and that should clear the problem.

Please also mind that CRC errors on the SMART table won't reset when the cable is replaced.
 

---

Posted by Steve Moore (Member # 4753) on 01-28-2014, 04:48 PM:
 
Well thanks for everyone's help so far, however the saga continues. I did the 200 miles trip today to our cinema armed with a 2nd Hand identical Hitachi drive, which I had tested and then wiped the MBR and also an brand new Western Digital Doremi approved 1TB drive. I also took new SATA leads.

I swapped out the leads, booted the server, no change.

I followed the Doremi "repair" process which takes drive "B" out of service, swapped the disk, let it mount that drive and it said "The disc replacement has been successful - The raid is rebuilding" and a green tick

However when you go back to the diagnostic tools, the Raid says it is degraded and no drive info comes up, nor does it say rebuilding raid and a % done
Then I tried removing the Hitachi and swapping for a Doremi approved Western Digital BRAND NEW sealed 1TB drive and did the same process and again it said successful however it did't actually rebuild the raid again.

I looked in the Bios and also did a detailed report which i uploaded to Doremi and received the report back. It did claim it could see either of the new discs with correct serial numbers. However in diagnostic tools a yellow dot and degraded Raid status.

After all that went into drive manage and it said the drive "B" operation not permitted - Device already disabled

I have a photo (not too clear of the screens if that helps



Sorry that was a long one. I'm a bit stuck now.

Just out of interest, could I power off the server and fit three brand new matching drives, reboot and try build a raid to see if it can do that. Then if that didn't work simply power down and put the three original drives back in and be back where I was; or would it need to re-register the old drive, so to speak, thus possibly rendering them useless.

Sorry that was a long Post, but any help would be appreciated.
 

---

Posted by Brad Miller (Member # 2) on 01-29-2014, 12:21 AM:
 
Steve please do not make a new thread, continue the discussion here. Unfortunately there are already posts on the new thread, so this one must be shut down.

Go here.
 

---