Bug 1469

Summary:	Mirrors website shows mirror to be unavailable while in fact it is reachable
Product:	Infrastructure	Reporter:	Remco Rijnders <remco>
Component:	Others	Assignee:	Olivier Thauvin <nanardon>
Status:	RESOLVED FIXED	QA Contact:
Severity:	normal
Priority:	Normal	CC:	marja11, misc, sysadmin-bugs
Version:	unspecified	Keywords:	PATCH
Target Milestone:	---
Hardware:	All
OS:	Linux
URL:	http://mirrors.mageia.org/
Whiteboard:
Source RPM:		CVE:
Status comment:
Attachments:	patch to replace // by / on uri check

Description Remco Rijnders 2011-05-29 05:20:47 CEST

On wednesday I registered http://mageia.webconquest.com/ as mirror. On http://mirrors.mageia.org/ it however shows in red, indicating the mirror is not available while it is.

The log on the server shows that a file with URL http://mageia.webconquest.com//mageia_timestamp is requested. Possibly the double // has something to do with this problem.

Comment 1 Michael Scherer 2011-05-29 10:25:48 CEST

Indeed, that's wrong as rfc 2616, 3.2.3 say that //foo should ( and could ) be different from /foo.

CC: (none) => misc

Comment 2 Michael Scherer 2011-05-29 10:32:17 CEST

Created attachment 495 [details]
patch to replace // by / on uri check

This should fix the issue, Nanar, can you check ?
( I am not sure if the substitution should be done without checking )

Comment 3 Nicolas Vigier 2011-09-26 22:51:37 CEST

any news on this ?

CC: (none) => boklm
Assignee: sysadmin-bugs => nanardon

Comment 4 Remco Rijnders 2011-09-27 07:29:12 CEST

misc came up with a fix for this (see his patch to this bugreport) but I don't think it ever went into production.

The mirror in question no longer runs thttpd (which I believe was the httpd resulting in the problem shown) so AFAIK this is currently unfixed but also not an issue for any of our current mirrors.

Keywords: (none) => PATCH

Comment 5 Marja Van Waes 2011-12-31 19:35:18 CET

(In reply to comment #4)
> misc came up with a fix for this (see his patch to this bugreport) but I don't
> think it ever went into production.
> 
> The mirror in question no longer runs thttpd (which I believe was the httpd
> resulting in the problem shown) so AFAIK this is currently unfixed but also not
> an issue for any of our current mirrors.

If it isn't an issue for any of the current mirrors, wouldn't it then be better to close as wontfix for now? It can then of course be reopened when the situation changes

CC: (none) => marja11

Comment 6 Marja Van Waes 2012-04-21 20:29:25 CEST

3 monthly ping

Comment 7 Remco Rijnders 2014-03-20 13:23:31 CET

Seems to have been fixed in the mean time. Thank you to the powers that be.

Status: NEW => RESOLVED
Resolution: (none) => FIXED

Nicolas Vigier 2014-05-08 18:05:09 CEST

CC: boklm => (none)