Misterious log entry 😄

[skeyby]

[error] 2020-06-25T07:52:37.644079Z [email protected] <0.3826.204> bd5c2120fc req_err(3717872069) unknown_error : normal
    [<<"chttpd:catch_error/3 L356">>,<<"chttpd:handle_req_after_auth/2 L322">>,<<"chttpd:process_request/1 L303">>,<<"chttpd:handle_request_int/1 L243">>,<<"mochiweb_http:headers/6 L128">>,<<"proc_lib:init_p_do_apply/3 L249">>]
[error] 2020-06-25T07:53:14.061720Z [email protected] <0.19345.203> 8a1fabea0b req_err(3717872069) unknown_error : normal
    [<<"chttpd:catch_error/3 L356">>,<<"chttpd:handle_req_after_auth/2 L322">>,<<"chttpd:process_request/1 L303">>,<<"chttpd:handle_request_int/1 L243">>,<<"mochiweb_http:headers/6 L128">>,<<"proc_lib:init_p_do_apply/3 L249">>]
[error] 2020-06-25T07:54:16.360572Z [email protected] <0.24324.200> bb2251ef97 req_err(3717872069) unknown_error : normal
    [<<"chttpd:catch_error/3 L356">>,<<"chttpd:handle_req_after_auth/2 L322">>,<<"chttpd:process_request/1 L303">>,<<"chttpd:handle_request_int/1 L243">>,<<"mochiweb_http:headers/6 L128">>,<<"proc_lib:init_p_do_apply/3 L249">>]
[error] 2020-06-25T07:54:16.360577Z [email protected] <0.11982.202> 6e161734ae req_err(3717872069) unknown_error : normal
    [<<"chttpd:catch_error/3 L356">>,<<"chttpd:handle_req_after_auth/2 L322">>,<<"chttpd:process_request/1 L303">>,<<"chttpd:handle_request_int/1 L243">>,<<"mochiweb_http:headers/6 L128">>,<<"proc_lib:init_p_do_apply/3 L249">>]

I have thousands of those in my couchdb2.log.

Is it something to worry about? "Unknown_error: normal".... mmmm

CouchDB 2.3.1 on FreeBSD 12.1

Thanks

[rjharmon]

This happens for me when a client process disconnects from an open _changes feed after 10s (or is it 10s of inactivity?). The "normal" is supposed to reassure you, but the ambiguity is almost as disturbing as the "unknown" part.

Maintainers, can this condition be detected more specifically so that the normal case can become a simple "client disconnect" message - or even lowered to [warn] or [info] severity?

[wohali]

We are tracking the actual bug report on this in #1600.

It does appear to just be a sporadic warning thrown when the client disconnects while CouchDB is waiting for a command, and can safely be ignored.

So far no one has reported this with CouchDB 3.x; try upgrading?

[skeyby]

@wohali Hi Joan, I'd like to, but Couch 3.x isn't available for FreeBSD yet.

I'm in contact with Dave Cottlehuber who's working on that (https://reviews.freebsd.org/D24615).