git/list[1] front-page[2] threads[3] people[4] search[5] about
 

Re: [External Mail]Re: Partial-clone cause big performance impact on server

From
Derrick Stolee <derrickstolee@github.com>
Date
Aug 17, 2022, 13:41 UTC
Message-ID
<fc71e91b-7a66-5653-d723-e4df17bf2a9c@github.com>
In-Reply-To
<11d7fc5721c541d6bc44bf635517497b@xiaomi.com>
On 8/17/2022 6:22 AM, 程洋 wrote:
> But I still think the protocol still should tell the server which ref
> the blob is reachable.
> Because it would be really hard to implement any kind of ACL

I think this idea has merit on its face, but it wouldn't really solve the problem since the reachability query would still need to be done, just from a smaller set of references at first. If we were able to say "this blob can be found at path X at commit Y" then the server could do a commit-reachability query and a path traversal, which should be a lot faster.

However, it would be extremely difficult to plumb into the partial clone machinery. At the point where Git realizes it is missing a promisor object, that code is very generic and removed from any kind of walk from a reference. That is further complicated by the fact that the walk is probably from a local reference, which can be entirely different from the remote reference.

> But git is surely designed for open sources community. It makes senses
> this request will be rejected.

We try to keep all kinds of users in mind, so the fact that this applies to not-completely-open repositories is not a blocker.

One possible hurdle is the fact that this branch-level security is a feature of Gerrit, not a feature of Git itself. Optimizing Git to that special case that Git does not itself support is less valuable to the Git project itself.

My personal take is that the technical complexity required to make this faster paired with the limited scope means that this feature would have a difficult time getting accepted into the Git project. Perhaps a motivated contributor will find ways to overcome these obstacles and find other interesting applications that benefit a larger portion of Git users.

That's just my expectation. I'd be happy to read any patches that try to solve this problem.

Thanks, -Stolee

Previous: 程洋Next: Jeff King
Message 12 of 40 in “Partial-clone cause big performance impact on server”
  1. 程洋Aug 11, 2022
  2. Jonathan TanAug 11, 2022
  3. 回复: [External Mail]Re: Partial-clone cause big performance impact on server程洋, Aug 13, 2022
  4. 回复: [External Mail]Re: Partial-clone cause big performance impact on server程洋, Aug 13, 2022
  5. ZheNing HuAug 15, 2022
  6. 程洋Aug 15, 2022
  7. Derrick StoleeAug 12, 2022
  8. Jeff KingAug 14, 2022
  9. Derrick StoleeAug 15, 2022
  10. 程洋Aug 15, 2022
  11. 程洋Aug 17, 2022
  12. Derrick StoleeAug 17, 2022
  13. Jeff KingAug 18, 2022
  14. 程洋Sep 1, 2022
  15. Jeff KingSep 1, 2022
  16. 程洋Sep 5, 2022
  17. Jeff KingSep 6, 2022
  18. 0/3 speeding up on-demand fetch for blobs in partial cloneJeff King, Sep 6, 2022
  19. 1/3 parse_object(): allow skipping hash checkJeff King, Sep 6, 2022
  20. Derrick StoleeSep 7, 2022
  21. Jeff KingSep 7, 2022
  22. 2/3 upload-pack: skip parse-object re-hashing of "want" objectsJeff King, Sep 6, 2022
  23. Derrick StoleeSep 7, 2022
  24. Derrick StoleeSep 7, 2022
  25. Jeff KingSep 7, 2022
  26. Junio C HamanoSep 7, 2022
  27. Jeff KingSep 7, 2022
  28. [BUG] t1800: Fails for error text comparisonrsbecker@nexbridge.com, Sep 7, 2022
  29. Junio C HamanoSep 7, 2022
  30. rsbecker@nexbridge.comSep 7, 2022
  31. Jeff KingSep 7, 2022
  32. Junio C HamanoSep 7, 2022
  33. Jeff KingSep 8, 2022
  34. Junio C HamanoSep 8, 2022
  35. 3/3 parse_object(): check commit-graph when skip_hash setJeff King, Sep 6, 2022
  36. Derrick StoleeSep 7, 2022
  37. Junio C HamanoSep 7, 2022
  38. 程洋Sep 8, 2022
  39. Jeff KingSep 8, 2022
  40. Derrick StoleeSep 7, 2022

Read the whole thread, see it on lore, or plain text.

$ cat FOOTERMessages come from the public archive at lore.kernel.org/git, fetched every hour. The front page is chosen and written each morning by an AI editor and can be wrong; the threads themselves are the record. About and API. For agents: an MCP server at https://gitlist.dev/mcp, and any thread, story or person page as Markdown by adding .md to its URL (or sending Accept: text/markdown). Details in /llms.txt.