Document parsing jam

[GAowc-stack]

ragflow Local deployment The resolution of the uploaded file is stuck

[KevinHuSh]

The logs seems fine.
File parsing is one by one by default.

[troubleshootme]

how do we enable asynch parsing?

[ZhenhangTung]

how do we enable async parsing?

could you explain more?

[troubleshootme]

I believe I found the env vars pertaining to this: DOC_BULK_SIZE =50 EMBEDDING_BATCH_SIZE =25 Those are the settings I am currently using. The issue is that when using graph, it takes forever to parse and embed. I am using external providers such as OpenAI and JINA and the tokens they are receiving is very low. The vm I am running ragflow in at the moment is allocated 64 GB ram and 32 cores. Utilization is fairly low and I can't help but think things could move along much faster. I guess my question is, how do those 2 values above look and would I get any benefit increasing them any more? Are there any other config values that will allow ragflow to send much higher RPM to external embedding apis? Tony Bruno / Consultant ***@***.*** / 205.876.3715 Bruno Computer Support www.troubleshootme.com ( https://troubleshootme.com/ ) 205.876.3715 Sent via Superhuman ( ***@***.*** )

…

On Wed, Jun 18, 2025 at 1:33 AM, LeonTung < ***@***.*** > wrote: > > > how do we enable async parsing? > > could you explain more? — Reply to this email directly, view it on GitHub ( #5011 (reply in thread) ) , or unsubscribe ( https://github.com/notifications/unsubscribe-auth/AZAVWCPLKR4WJFBY32UI2ED3EEB3HAVCNFSM6AAAAABXHNRGU2VHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTGNJQGQ3DGMI ). You are receiving this because you commented. Message ID: <infiniflow/ragflow/repo-discussions/5011/comments/13504631 @ github. com>

[ZhenhangTung]

Pls check this PR to see if it can answer your question: #7845