Intermittent glitch with DSP100

[Peter Foyster]

I have had a couple of weird things happen with my Dolby DSP100 whilst setting up for a show tonight.

I went to ingest a trailer from a USB stick and it came up with an error and reset. It then ingested the trailer when I performed the operation again.

Later, when I was checking a feature and I set a new point in the movie to start from it came up with an error to say it had lost communication with the projector. After less than a minute it came good again without any input from me.

A patron said, a few weeks ago, that the picture went out for about a minute but then came back on. I could not explain what happened back then but am thinking that it might have been the same thing.

Any ideas as to what could cause this?

[Leo Enticknap]

To start with, we need to establish whether the intermittent fault is with the DSP or the DSS. I suspect the latter, because setting a new point in the movie, and ingestion, are operations that primarily involve the server, not the media block (player).

Possibilities that spring to mind include the RAID being almost (as in, more than around 80%) full, resulting in disruption to playback while it hunts around for space on the volume, or one of the drives going bad. Could you download a log from the projector, upload it to Dolby Log Analyzer, and post a link to the result here?

[Markus Lemm]

I had a similar issue with our DSP/DSS 100 a few years ago. The Lost Communication error came up a bunch of times. I just unplugged the Ethernet cable between the DSP and DSS and the cable that runs to the projector and re-seated them. Make sure you use good quality short Ethernet cables without kinks in them.

If that doesn't work then start looking at logs

[Leo Enticknap]

A little squirt of DeOxit on the RJ45 contacts can't hurt, either.

[Peter Foyster]

Thanks gents.
There was about 400Gb left out of a total 2Tb storage space. I still had a number of features that I was able to delete which brought the available space up to about 1.2Tb.

I then shut down the server/showplayer and, before rebooting, reseated all the Ethernet cables. I took the opportunity during the reboot to reset the BIOS clock which was about 3 minutes slow. I turned on the projector and reconnected the link to a so called NTP server that I had tried to set up on an old laptop with Leo's help last year. I'm pretty sure that the internal clock is not seeing the laptop but at least the laptop's time is very accurate!

When the system rebooted for the second time, the clock was about 5 seconds out and the internal clock about 10 seconds out. Considering that I have had the units for about 6 years now and don't know their history, 10 seconds is not bad, especially as I am not aware of any NTP source that they could have linked to. The BIOS clock only take a few weeks to lose time but a quick reboot and a reset is all it needs.

Perhaps more frequent rebooting might prevent some of these glitches from occurring. I tend to leave the entire system on (projector on standby) during the season and even for the 6 months that the cinema is closed during the winter months, turning the projector on for a few hours each week.

The reboot last night indicated that all 4 drives in the RAID were healthy with no reallocated sectors. The drives are barely 5 years old.

If I get a re-occurrence of the problem I shall download a log as suggested.

[Leo Enticknap]

You're doing exactly what I would with this equipment and usage situation: leaving it on during prolonged periods of hibernation, but rebooting immediately before the next period of active use.

Five years is approaching the point at which the reliability of spinning rust hard drives starts to decline significantly. Downloading a log package from your DSS100 and putting it through Dolby Log Analyzer every few weeks would, IMHO, be a good idea: that will flag up pretty clearly if any of those drives are starting to go bad.

Good to hear that a reboot seems to have cleared the glitchiness.

[Peter Foyster]

Any tips on the appropriate replacement 1 Tb drives to suit the DSS100? I guess I should have at least one, if not all four, in reserve. Steve G suggested the ones I have now but they are no longer available.

[Marco Giustini]

The qualified 1TB drives for the DSS100 are the following:

1TB Hitachi Ultrastar 1TB (Gemini - K) HUA721010KLA330
1TB Hitachi Jupiter-U 1TB HUA722010CLA330 (FW 39C)
1TB Seagate Constellation ES ST31000524NS (FW SN11)
1TB Seagate ES.2 1TB ST31000340NS (FW SN06)

That does not mean other models/brands won't work of course.

I did not know 1TB (and more!) have actually been qualified for the DSS100 model! Interesting.

[Steve Guttag]

There are some VERY old DSS100s, I've been told, that cannot take 1TB drives. I have never encountered them. I've had some bad experiences with Western Digital now, including the HA210 series (that which effectively replaced the HGST drives for 1TB and 2TB. I'm having pretty good luck, so far with Toshiba drives that Dolby has qualified for their current servers.

[Peter Foyster]

The drives were actually replaced in November of 2018, with Western Digital "Gold" WD1005FBYZ as recommended by Steve at the time, so they are only about 3 years old. I should probably buy at least one as a spare in case of an emergency but if Steve has now gone off Western Digital I guess I had better look at something else. I understand that the DSS100 can take up to a 2TB drive but I think I will stick to the 1TB. Its just a matter of finding the right one, that won't be obsolete a year from now. Then.. who am I to talk about obsolete, still using a DSS/DSP100!

[Steve Guttag]

Western Digital WD1005 has been folded into the HA210 series. Somewhere, I have a document from Western Digital indicating that one may have either part number (the current number is

HUS722T1TALA604). for what it's worth, I'm seeing GDC using HA210 series drives on their "Enterprise Storage" for the SR-1000. I'm having more issues with the 2TB drives than the 1TB. If you find any WD1005, they are bound to be old-stock. They'll probably be fine. I'm not having anywhere near the problems I had with Seagate in years past but I noticed an uptick in reliability problems on the WD drives as compared to the Hitachi/HGST ones.

[Peter Foyster]

Thanks Steve,
The drive is reasonably priced (A$145) so i might pick one up and save it as a spare. Chances are it won't get used for at least 2 years but it is better to be safe than sorry.

[Peter Foyster]

Well…it seems that I spoke too soon. Coming in for the final show of the weekend I noticed that the showplayer could not find the show store.
A reboot just resulted in two short beeps, followed by four short beeps, followed by two short beeps and a blank monitor.
I phoned a tech on the other side of Australia, who suggested a few things to try, to no avail.

I have a spare DSS100 and suggested to him that a swap might work. I changed over the hard drives, connected the replacement up and it booted fine.
I reloaded the KDM for the night’s feature and it has gone without incident.

Does anyone know what the sequence of beeps mentioned above represents and is it possible that the unit is salvageable?

I guess I will have to try and locate a couple of spare DSS100s to insure against a repeat performance.

[Steve Guttag]

Invariably, beeps mean that the memory has issues. Try reseating the sticks. Failing that, try them one at a time. Also, I've had low BIOS batteries do funky things on all sorts of systems. Someone at Dolby will know the beep code.

[Marcel Birgelen]

From the manual of the Supermicro board most likely involved:

Recoverable beep codes:
1 long and two short beeps - video configuration error
1 repetitive long beep - no memory detected
1 continuous beep w/Front Panel OH LED on - System Overheat

The manual doesn't list the non-recoverable beep codes, so the one you experienced seems to be a non-recoverable hardware error, unfortunately, without specification in the manual.
As Steve indicated, you should check the memory banks, maybe re-seat them. Otherwise, it looks like another low-level failure on the motherboard....