OCR fails on batch ingest

[ajstanley]

I thought we had this one sorted, but if we are doing a batch import with images being tesseracted then the process manager attempts to run more than one instance, (I think it goes for five) which overwhelms the container, and instead of seeing a new OCR we get nothing but timeouts. A single process uses about 65% of the CPU cycles.

When I'm doing migrations I pull the existing OCR over instead of creating new, but this won't work if there is no existing OCR.

It continues to work well for one-by-each ingestion.

[noahwsmith]

Which container tag are you testing on, Alan?

[ajstanley]

alpha-15

…

On Thu, Apr 21, 2022 at 10:48 AM Noah W. Smith ***@***.***> wrote: Which container tag are you testing on, Alan? — Reply to this email directly, view it on GitHub <#251 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AADY2J2JEXOYYSZDFC5QZILVGFMA5ANCNFSM5T4DRDLA> . You are receiving this because you authored the thread.Message ID: ***@***.***>

-- Alan Stanley Community Outreach Coordinator Spinal Cord Injury, PEI ☎ *902-393-0017*

[noahwsmith]

And with a new clone of ISLE-DC, or no?

[ajstanley]

Fresh build as of a couple of days ago.

…

On Thu., Apr. 21, 2022, 12:56 p.m. Noah W. Smith, ***@***.***> wrote: And with a new clone of ISLE-DC, or no? — Reply to this email directly, view it on GitHub <#251 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AADY2J4MC4ROVBRSGFX5EVLVGF3C5ANCNFSM5T4DRDLA> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[ajstanley]

But let me run a few more tests before you do anything - I've run into some other problems on the server that may have obscured the real problem.