Concurrent PUT and new data acceptance in slicer #278

New issue

Open

opened 2025-12-28 18:07:39 +00:00 by sami · 1 comment

sami commented

2025-12-28 18:07:39 +00:00

Owner

Originally created by @roman-khimov on GitHub (Aug 11, 2025).

I'm always frustrated when I'm uploading big objects into NeoFS. One can clearly notice 64M boundaries where the process stops for a while and slicer PUTs its intermediate objects.

Describe the solution you'd like

Keep accepting data, PUT in the background (but only one part).

Describe alternatives you've considered

Keep things as is, this works.

Additional context

Kinda the opposite side of https://github.com/nspcc-dev/neofs-node/issues/2791.

This also requires more buffers which is a problem. But it can clearly help PUT performance.

Originally created by @roman-khimov on GitHub (Aug 11, 2025). ## Is your feature request related to a problem? Please describe. I'm always frustrated when I'm uploading big objects into NeoFS. One can clearly notice 64M boundaries where the process stops for a while and slicer PUTs its intermediate objects. ## Describe the solution you'd like Keep accepting data, PUT in the background (but only one part). ## Describe alternatives you've considered Keep things as is, this works. ## Additional context Kinda the opposite side of https://github.com/nspcc-dev/neofs-node/issues/2791. This also requires more buffers which is a problem. But it can clearly help PUT performance.

sami added the

enhancement

labels

2025-12-28 18:07:39 +00:00

sami commented

2025-12-28 18:07:39 +00:00

Author

Owner

@roman-khimov commented on GitHub (Oct 6, 2025):

Processing delays can trigger TCP slowdowns as well, so the performance penalty can be higher than expected.

@roman-khimov commented on GitHub (Oct 6, 2025): Processing delays can trigger TCP slowdowns as well, so the performance penalty can be higher than expected.