See http://web.archive.org/web/20200223214548/http://bugs.libre-riscv.org/query.cgi Produced by using https://web.archive.org/save to save http://bugs.libre-riscv.org/ and checking "save outlinks" (only appears when logged in to archive.org)
it's probably a good idea to check this soon before the logged errors on the server are deleted
(In reply to Jacob Lifshay from comment #1) > it's probably a good idea to check this soon before the logged errors on the > server are deleted there's only the access log. the error log is empty, and does *not* contain any errors for that date and time.
(In reply to Luke Kenneth Casson Leighton from comment #2) > (In reply to Jacob Lifshay from comment #1) > > it's probably a good idea to check this soon before the logged errors on the > > server are deleted > > there's only the access log. the error log is empty, and does *not* > contain any errors for that date and time. Seems odd, maybe we should try getting archive.org to access it again but log the whole connection this time, that way we could replay it and determine if it's our server that's causing the error
(In reply to Jacob Lifshay from comment #3) > Seems odd, maybe we should try getting archive.org to access it again but > log the whole connection this time, that way we could replay it and > determine if it's our server that's causing the error they claim a 50x error occurred however there is no such error in the logs.
(In reply to Luke Kenneth Casson Leighton from comment #4) > they claim a 50x error occurred however there is no such error in the logs. either our server is producing an error but not logging it (not impossible), or archive.org's internal stuff is failing. The idea was to try to reproduce the exact requests that archive.org sent to see which case occurred.
Now I'm getting 502 Bad Gateway from my local computer: curl --verbose --ipv6 --no-progress-meter http://libre-riscv.org/ &> curl.log produces: * Trying 2a00:1098:82:f::1:80... * Connected to libre-riscv.org (2a00:1098:82:f::1) port 80 (#0) > GET / HTTP/1.1 > Host: libre-riscv.org > User-Agent: curl/7.69.1 > Accept: */* > * Mark bundle as not supporting multiuse < HTTP/1.1 502 Bad Gateway < Server: nginx < Date: Thu, 02 Apr 2020 17:44:04 GMT < Content-Type: text/html < Content-Length: 166 < Connection: keep-alive < { [166 bytes data] * Connection #0 to host libre-riscv.org left intact <html> <head><title>502 Bad Gateway</title></head> <body bgcolor="white"> <center><h1>502 Bad Gateway</h1></center> <hr><center>nginx</center> </body> </html> Either we should fix the server's config or remove the IPv6 DNS entry, since it works through IPv4.
ngggh annoying. am going to assign this one to alain to investigate.
All sites need the lines: listen 80 ; listen [::]:80; I have done that with all of them.
(In reply to Alain D D Williams from comment #8) > All sites need the lines: > > listen 80 ; > listen [::]:80; fantastic. ah! i previously added ::443 and of course that wasn't it. thanks alain.
git.libre-soc.org does not have an IPv6 address. It should be 2a00:1098:82:f::1
(In reply to Alain D D Williams from comment #10) > git.libre-soc.org does not have an IPv6 address. > > It should be 2a00:1098:82:f::1 added to /etc/bind/libre-soc.db
Please add to /etc/bind/libre-soc.db an IPv6 address for lists.libre-soc.org
Please create DNS entries for inbox.libre-soc.org IPv4 & IPv6 Ta
(In reply to Alain D D Williams from comment #13) > Please create DNS entries for inbox.libre-soc.org > IPv4 & IPv6 > > Ta done - it's in /etc/bind/libre-soc.org.db
btw i'm just doing a snapshot of the offsite rsync backups that i do, before rsync'ing. 3.2 GB copied so far, another 3 GB to go. that way we have a comparison of before-and-after the libre-soc state, and can do "diffs" on the configs (etc.)
All done - except a quick tidy up - removing old config commented out. Done: * all libre-riscv sites redirect to a corresponding libre-soc site * all http:// sites redirect to a https:// site This might mean that web browsers will not auto fill passwords, since they are looking at different domains. Also: what to do with uelves.com ? What is this ?
(In reply to Alain D D Williams from comment #16) > All done - except a quick tidy up - removing old config commented out. > > Done: > > * all libre-riscv sites redirect to a corresponding libre-soc site > > * all http:// sites redirect to a https:// site > > This might mean that web browsers will not auto fill passwords, since they > are looking at different domains. yes. found that out already :) > Also: what to do with uelves.com ? What is this ? that's another domain i'm running, nothing to do with libre-soc.
the next one, before we get much further, is the mailing lists: https://bugs.libre-soc.org/show_bug.cgi?id=184 however at the same time, each list needs a public-inbox: https://bugs.libre-soc.org/show_bug.cgi?id=181 see /home/publicinbox and fortunately it's possible to have a per-list-filter despite all messages going into /home/publicinbox/Maildir to discuss at bug #184
(In reply to Alain D D Williams from comment #16) > All done - except a quick tidy up - removing old config commented out. > > Done: > > * all libre-riscv sites redirect to a corresponding libre-soc site > > * all http:// sites redirect to a https:// site Bugzilla needs to be updated to send out the correct url.
(In reply to Jacob Lifshay from comment #19) > Bugzilla needs to be updated to send out the correct url. well spotted. done (in bzla admin interface)
uelves.com is now done to the standard of the others. I have removed certbot - replaced with my scripts.
All now done