Doremi Memory Errors (recovered)

[Ryan Gallagher]

My season is over but on a couple occasions over the summer the DCP2K4 Doremi notifications interface reported a memory error and then quickly reported it had corrected or recovered.

is this something to address. Is it detecting ECC corrective events or something else?

It did not cause any show problems, but have time now to circle back.

If those refer to the ram, would it be worth booting into memtest for a extended test run?

[Leo Enticknap]

It certainly can't hurt. Pulling, contact cleaning, and then reseating the RAM boards is something else I'd try, too, plus uploading a "detailed report" log package from it to the Dolby Log Analyzer.

[Ryan Gallagher]

I have maintenance tomorrow too, will pull logs and upload to analyzer. It is also time to do annual cleaning/service, so good time to clean/reseat too.

[Ryan Gallagher]

Not seeing anything related to those errors in Analyzer.
https://loganalyzer.dolbycustomer.co...f95f141f6915d3

Only thing it surfaced I was unaware of was there being new firmware since our certificate extension installs.

And should I worry about ch 12+13 watermarking? We only use channels 9-16 when a festival is here and provides HI/VI systems.

But no hints on the memory errors. It has been rebooted since those occurred, so maybe the analyzer does not see them?

[Leo Enticknap]

I would have thought that any RAM errors should have been recorded in a log file that survives a reboot, but it would appear not. That log analysis basically gives your DCP2000 a clean bill of health, apart from asking for a Dolphin firmware update.

If the only thing you use 9-16 for is HI and VI on 15 and 16 (presumably because you want 7 and 8 for the 7.1 rear surrounds), you don't need to worry. In some setups, 13 is used for control signals for wobbly seats in the auditorium (DBox), and in Atmos-equipped houses, 14 is used to synchronize the regular AES and Atmos, so that the pix and sound won't drop out of sync if for any reason there has to be a fallback from IAB to regular AES3 audio. The digital watermarking on these channels could cause those functions to not work, but if you don't use either, this is not a problem in your theater.

[Ryan Gallagher]

I would have thought that any RAM errors should have been recorded in a log file that survives a reboot, but it would appear not. That log analysis basically gives your DCP2000 a clean bill of health, apart from asking for a Dolphin firmware update.

If the only thing you use 9-16 for is HI and VI on 15 and 16 (presumably because you want 7 and 8 for the 7.1 rear surrounds), you don't need to worry. In some setups, 13 is used for control signals for wobbly seats in the auditorium (DBox), and in Atmos-equipped houses, 14 is used to synchronize the regular AES and Atmos, so that the pix and sound won't drop out of sync if for any reason there has to be a fallback from IAB to regular AES3 audio. The digital watermarking on these channels could cause those functions to not work, but if you don't use either, this is not a problem in your theater.

Thanks Leo. I’ll still clean and reseat memory in the servicing, today was all about replacing the booth network switch, memtest if I get an opportunity and am bored, but mostly I’ll just watch for that error again and dump report again before a reboot.

Mostly correct, our 7.1 channels are remapped to free up 9-16 coming from the DMA8/CP650. And the AP20 analog inputs come from the CP200 iirc. Festivals just take the unused 9-16 from Doremi to get the HI/VI and CCAP working, leaving our AP20 weirdness alone.

[Elia Orselli]

Extending the analysis date period some error is visible. Maybe you saw this one?
immagine.png

[Ryan Gallagher]

Extending the analysis date period some error is visible. Maybe you saw this one?
immagine.png

Good catch, I didn't see the tools to look at a larger date period.

There is a good chance these memory errors did occur while we had a relatively full content drive in the thick of the summer series. Will be more mindful of pushing our storage limit... our DCP2K4 has NOT had the 4th drive added, something to consider also.

[Ryan Gallagher]

Looking at our whole summer series window, 6/1 to 8/31, other than predictable communication errors due to server only being on, CRU and USB drive issues at times, that storage quota one is really the only one that jumps out as possibly contributing to the memory notices that were popping up on screen. I should have made note of what date those occurred.

EDIT: If I look at just 9/4 I notice that our md0 storage was flagged as "active, recovering" with 66% complete. I doubt it was on long enough yesterday to finish... I should leave it on again soon overnight and see if that status changes.