summaryrefslogtreecommitdiffhomepage
path: root/test/test_ruby_application.py
diff options
context:
space:
mode:
authorMax Romanov <max.romanov@nginx.com>2020-12-29 19:01:24 +0300
committerMax Romanov <max.romanov@nginx.com>2020-12-29 19:01:24 +0300
commitd65a66f9d813294917822554311281c5e1a7126b (patch)
treefe375efda23817f43a1ec7a30ff2250b58477e89 /test/test_ruby_application.py
parentd3d6864bdc64f34924e686ff65da704b29aaaa93 (diff)
downloadunit-d65a66f9d813294917822554311281c5e1a7126b.tar.gz
unit-d65a66f9d813294917822554311281c5e1a7126b.tar.bz2
Libunit: processing single port message.
This partially reverts the optimisation introduced in 1d84b9e4b459 to avoid an unpredictable block in nxt_unit_process_port_msg(). Under high load, this function may never return control to its caller, and the external event loop (in Node.js and Python asyncio) won't be able to process other scheduled events. To reproduce the issue, two request processing types are needed: 'fast' and 'furious'. The 'fast' one simply returns a small response, while the 'furious' schedules asynchronous calls to external resources. Thus, if Unit is subjected to a large amount of 'fast' requests, the 'furious' request processing freezes until the high load ends. The issue was found by Wu Jian Ping (@wujjpp) during Node.js stream implementation discussion and relates to PR #502 on GitHub.
Diffstat (limited to '')
0 files changed, 0 insertions, 0 deletions