Sometimes the timeouts take as much as 60 seconds, and usually, after the timeline loads, the right column doesn't even appear. They are caused by very slow or unreachable external links.
Using setconfig.php, I can see there's a "http connect_timeout" (value=5) and a "http timeout" (value=60). I guess the second is the one that needs to be changed, because I can see several "Request timed out after 60 second(s)" in the logs.
I've tried to change the value to 10 seconds (with "php setconfig http timeout 10") but then I get in the logs "config option http/timeout is not an integer value: '10'". From there, only deleting the value ("php setconfig -d http timeout") restores to the default value and logs stop complaining.
Is there any other way to change this behaviour? Timeouts appear too frequently right now...
Sometimes the timeouts take as much as 60 seconds, and usually, after the timeline loads, the right column doesn't even appear. They are caused by very slow or unreachable external links.
Using setconfig.php, I can see there's a "http connect_timeout" (value=5) and a "http timeout" (value=60). I guess the second is the one that needs to be changed, because I can see several "Request timed out after 60 second(s)" in the logs.
I've tried to change the value to 10 seconds (with "php setconfig http timeout 10") but then I get in the logs "config option http/timeout is not an integer value: '10'". From there, only deleting the value ("php setconfig -d http timeout") restores to the default value and logs stop complaining.
Is there any other way to change this behaviour? Timeouts appear too frequently right now...
Since i wrote this bug report, i've managed to improve the situation by using DNS cache. Right now, this behaviour only happens from time to time, but it is there. It has just happened, leaving /public timeline without the right column, and only about 1/3 of the page loaded, after 60 seconds wait. This is what log says:
Since i wrote this bug report, i've managed to improve the situation by using DNS cache. Right now, this behaviour only happens from time to time, but it is there. It has just happened, leaving /public timeline without the right column, and only about 1/3 of the page loaded, after 60 seconds wait. This is what log says:
2021-07-29 15:27:04 LOG_ERR: [gnusocial.net:868.0c868370 GET /main/public] HTTPClient: HTTP HEAD https://proxy.vulpes.one/gemini/reisub.nsupdate.info/planetalibre/index.gemini - Request timed out after 60 second(s)
2021-07-29 15:33:47 LOG_ERR: [gnusocial.net:891.9c55e6a1 GET /main/public] HTTPClient: HTTP HEAD https://proxy.vulpes.one/gemini/reisub.nsupdate.info/planetalibre/index.gemini - Request timed out after 60 second(s)
XRevan86
Sometimes the timeouts take as much as 60 seconds, and usually, after the timeline loads, the right column doesn't even appear. They are caused by very slow or unreachable external links.
Using setconfig.php, I can see there's a "http connect_timeout" (value=5) and a "http timeout" (value=60). I guess the second is the one that needs to be changed, because I can see several "Request timed out after 60 second(s)" in the logs.
I've tried to change the value to 10 seconds (with "php setconfig http timeout 10") but then I get in the logs "config option http/timeout is not an integer value: '10'". From there, only deleting the value ("php setconfig -d http timeout") restores to the default value and logs stop complaining.
Is there any other way to change this behaviour? Timeouts appear too frequently right now...
Since i wrote this bug report, i've managed to improve the situation by using DNS cache. Right now, this behaviour only happens from time to time, but it is there. It has just happened, leaving /public timeline without the right column, and only about 1/3 of the page loaded, after 60 seconds wait. This is what log says:
2021-07-29 15:27:04 LOG_ERR: [gnusocial.net:868.0c868370 GET /main/public] HTTPClient: HTTP HEAD https://proxy.vulpes.one/gemini/reisub.nsupdate.info/planetalibre/index.gemini - Request timed out after 60 second(s)
2021-07-29 15:33:47 LOG_ERR: [gnusocial.net:891.9c55e6a1 GET /main/public] HTTPClient: HTTP HEAD https://proxy.vulpes.one/gemini/reisub.nsupdate.info/planetalibre/index.gemini - Request timed out after 60 second(s)