> ons 2012-08-29 klockan 12:42 -0600 skrev Alex Rousskov: >> Hello, >> >> I saw bogus ERR_ZERO_SIZE_OBJECT responses while testing Squid v3.1, >> but the same problem ought to be present in v3.2 as well. >> >> A compliant proxy may retry PUTs, but Squid lacks the [rather >> complicated] code required to protect the PUT request body from being >> nibbled during the first try, when pconn races are possible. > > +1 on the patch from me.
Committed to trunk as r12319. >> Today, we have a choice: >> >> A. Fail some PUTs. Close fewer pconns (current code). >> B. Handle all PUTs. Open more connections (patched code). >> >> If this patch is accepted, performance bug 3398 will resurface. Henrik, >> do you think committing the patch is the right decision here even though >> it will reopen bug 3398? > > Yes, but with a plan to fix it correctly. ... >> The bug title is wrong. There is a long discussion in the bug report >> about what the bug is really about. I think a better bug title would be: >> "persistent server connection closed before PUT". > > yes. > > Fix the bug title, reopen it with comment above and restore the check. Done. Quality patches preserving PUT bodies (via 100-continue and/or buffering) to optimize Squid pconn handling are welcomed. Thank you, Alex.