I used the ideas from this post to work on the images but we dont have any task scheduler for that workflow. The jobs almost always say "failed" but task completion is 610/610 for example. There are no errors in DCC event logs. How do we prevent OCR from getting stuck without manually ending these processes?
https://answers.laserfiche.com/questions/75843/LfDcc-stuck-at-25-cpu-basically-one-core
Question
Question
DCC: ocr processes keep cpu at 100% long after job is done
Answer
Hi Ben,
Opening up a support case is the best course of action in this situation, as tech support can help narrow down the cause in your particular system.
Replies
A little more info. There is one worker machine which never has any ocr processes running in the morning.
It is only the scheduler/server that hits 100% cpu and it has 8 LFOmniOCR.exe processes that need to be manually terminated in the morning.
SO, my theory is that OCR gets stuck on documents that it cant OCR such as maps and blank pages. One OCR process takes up one core so processesing is reduced each time DCC hits a doc that cant be OCRed until all 8 cores are full and then it fails. We have never seen it finish all 4000 docs sent to it.
Yes just waiting for customer to renew LSAP now. Thanks
what was the resolution to this issue?
The machine in question did not met the minimum specifications for use with DCC, and thus troubleshooting did not continue:
CPU: 2.93 GHz or faster multi-core processor
Memory: 1 GB per processor core (2 GB per processor core recommended)
Furthermore, this issue was occurring in the beta version of the product, rather than the full version now available.
If you are encountering the same issue, and your machine meets these requirements, we recommend opening a support case through your reseller to troubleshoot the issue.