jerryshao commented on PR #13250:
URL: https://github.com/apache/gravitino/pull/13250#issuecomment-5749472172

   Hi @neoLsH, thanks so much for the thorough review — this is exactly the 
kind of feedback that's useful before merge. Here's where I landed on each 
point:
   
   **1. Byte-vs-line cap / O(S²) risk on a pathological single huge line** — 
you're right, this is a real gap and I'll fix it in this PR. Will bound the 
actual bytes read and returned (not just the line count), so a giant single 
line can't blow up the request thread or the response size.
   
   **2. Single atomic call + `truncated` flag vs. the current two separate 
`getJobStdout`/`getJobStderr` calls** — I don't think this rises to a real 
problem. Snapshotting stdout/stderr at slightly different moments for a 
still-running job doesn't meaningfully hurt debuggability in practice, and true 
atomicity isn't really achievable here anyway even with a single method — it 
would still mean two sequential file reads underneath. I'd rather keep the 
simpler two-method shape as-is.
   
   **3. Retention window mismatch (1h vs 7d)** — this is a limitation of the 
local job executor as it stands today, and I'm treating it as expected/known 
behavior rather than something to reconcile in this PR.
   
   **4. Multi-node ownership gap (non-owning node returns empty)** — also a 
local-job-executor-specific limitation. I'm planning to address this separately 
in a follow-up PR rather than folding it into this one.
   
   Thanks again for the careful read — will push the fix for #1 shortly.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to