Pro Tools 10 - Disk access bug ?

[SteveH44]

Alex - We are seeing the exact same problem here. We have 17 workstations running Pro Tools v10 HD2 on OSX 10.6.8. Some workstation running qualified FW400 ext. drive, a couple running qualified FW800 drives. With 8 tracks in record for ensemble dialog records, machine starts to slow down when nearing 3-4 gigs of session data. Randomly we get the error -9073 (drives too slow), or beach balling, or just dropping out of record. Other times when going into record, PT stalls for 3 to 4 seconds before going into record. We saw the spikes you talk about when PBE is set to 512 or 1G. All of our machines log in to an OD server for user profiles, but this is not the issue, since it does the same thing locally. We've played with disk caching, but this does not seem to be the answer either, as some of our mixers have reported it not making a difference. All drives are local to the machine. We DO run SSD's as system drives only, and Avid is saying that they're not qualified, but I can not see why not, many users are using them trouble free. We do not record on them. So I'm thinking that there's a bug that Avid is not aware of yet. They state we should be disk caching, but caching is only required to accommodate various slower record drives, and a qualified local FW800 should not require it.

Steve
LA Studios, Hollywood, CA

[Shady Lane Studio]

Originally Posted by alex s

Hello,

Can someone please prove / disprove my findings by doing the following in PT10:

Record 24 tracks for 15 minutes.

Stop Recording.

Start Recording immediately.

Do you get a DAE error 9073 - DAE can't get audio from drives fast enough.

Thanks,

Al

I just upgraded from LE to PT10 and am having this same issue. I looked online and found your post. I have almost the exact same setup as you and can reproduce the problem every time; it's very frustrating. I record for a church service where I must stop and start recording frequently using up to 8 channels for the praise band. To me, this is a big problem that I need to find a work around for. At least I know I'm not going crazy! Thanks.

[Shady Lane Studio]

The "open Task Manager and select Pause all and Pause During Playback" trick seems to be working for the time being—until they acknowledge and fix the bug.

[Bill Denton]

If it's possible to do so, turn off disk caching while recording. You don't gain anything by using it.

I don't think what you're seeing is a bug, it's just a lack of understaning what's happening with your system.

Keep in mind that whatever is going into disk cache eventually has to be written to disk, and the perfect time to do it is when you've stopped recording. And since the disk and the cache are pretty busy during this write cycle it's going to take some time to "shift gears" and begin recording again.

Disk cache is primarily useful when you are editing, and need to quickly move around in the recording...

[Shady Lane Studio]

It's very possible I lack understanding in what's happening to my system, or Pro Tools for that matter. What would be helpful is a little guidance in light of how to change my disk caching, or whatever ideas you have to fix the problem.

[bradp]

hi all,
well, i must be lucky, the last live recording i did had one intermission, and using my fw400 glyph worked fine. i have another live concert this weekend and i did some tests this week.. YIKES.... after stopping a record (10min test, 170min test) and then after a minute, initiating record again, i got the disk error.(over and over after trying different settings)
i will try the task manager pause workaround tomorrow.
has anybody tried making a new playlist and recording from the start of session?
i would think that disk cache would not be a factor in this case as there's no audio to cache...
i would IMMENSELY appreciate any comments to improve things as i am now REALLY concerned if my weekend will be STRESSFUL!

*recording 2 stereo, 4 mono tracks at 88.2/24bit, interleaved on.
i found using 1 processor, 2048buff, no error protect, 60%cpu, dly comp off, made some difference. I do think the task manag. will help as well. I did initiate record quickly after stopping..
im using pt 10.0.0
if you need more specs, please ask

[mattpete]

This was fixed in 10.1.1 / 10.1.2.
The waveform rendering is completed after recording stops. This rendering requires disk access. In PT9 this was always "Paused During Playback" (or Record) no matter what the manual setting of "Pause During Playback" in the Task Manager window. But in PT10, this inadvertently got turned off. Turning on Pause During Playback manually would correct the situation. In PT 10.1.1, the "always pause" behavior for waveform rendering is re-enabled.

[bradp]

hi matt,
thanks for your quick reply. This puts some of my concern to rest. Ill be testing tomorrow! ;-)
you talk about 10.1.1 / 10.1.2.
Im only aware of the 10.0.1 update..
can you clarify?
also, could you tell me where on the avid site is the version information posted?
thx matt

[mattpete]

Brad, at the top level of the forum, there is a Technical Alerts category: http://duc.avid.com/showthread.php?p...94#post1920894

Always has link to the latest version available.

[bradp]

thanks matt,
here's some test info with:
pt 10.1.2, glyph portagig50 500gig fw800 drive(extended journaled). 44.1/24, 8tracks
<<mbp>>><HD>>><adapt fw800 to fw400><digi002r>
buff 1024, cpu 95%, no dly comp, streambuff 2:
after 12 min_ record stops, error -6086 (this is diff than i was getting with 10.0.0= -6093 and -9128)
then, change cpu to 512, all else same
have been in record for 56min and counting.
stopped at 60 min
initiate rec after existing clips
about 15 min later , -6093 error, falls out of record

Q: do i have the daisy chain order correct?
Q: does it seem logical that its working better at a lower cpu usage buff setting?

i cant seem to replicate the error, i will do more tests..