Closed hmrodrigues closed 9 years ago
You should hard reset your browser's cache and restart Webmin.
@qooob already did it. It says to many redirects
What about putting cursor to URL line in Firefox and hitting enter? That shouldn't be the theme's case really..
@qooob since was on a virtual machine, I recover from a snapshot without webmin and after the update of the theme the problem happened again
Could you please try Chrome browser? Did you also try rebooting your machine?
In case it happens in Chrome, please make a screenshot from Chrome with DevTools opened.. I really have no idea why this could be happening.
Worth checking:
http://www.howtogeek.com/forum/topic/firefox-error-the-page-isnt-redirecting-properly
I can confirm the same issue Running CentOS7 Tested using Firefox and Safari on Mac (cleaned cache and cookies)
There is also Webmin cache. It could get stuck. Clear server side cache as well.
https://support.mozilla.org/en-US/questions/938531
What is your Webmin version, guys?
I cleaned everything that could be cleaned, same behaviour. Webmin 1.760. Below is a snapshot of the webmin debug log (all logging enable), and the log goes on with same entries until firefox/safari stops trying to load the page.
4719 [19/Sep/2015 13:38:29.004928] - 10.10.0.58 - START "script=index.cgi" 4719 [19/Sep/2015 13:38:29.005235] - 10.10.0.58 - READ "/etc/webmin/miniserv.conf" 4719 [19/Sep/2015 13:38:29.005738] root 10.10.0.58 - READ "/usr/libexec/webmin/authentic-theme/config" 4719 [19/Sep/2015 13:38:29.007453] root 10.10.0.58 - READ "/usr/libexec/webmin/lang/en.UTF-8" 4719 [19/Sep/2015 13:38:29.008967] root 10.10.0.58 - READ "/etc/webmin/custom-lang" 4719 [19/Sep/2015 13:38:29.009169] root 10.10.0.58 - READ "/etc/webmin/custom-lang.en.UTF-8" 4719 [19/Sep/2015 13:38:29.018060] root 10.10.0.58 - READ "/usr/libexec/webmin/authentic-theme/lang/en.UTF-8" 4719 [19/Sep/2015 13:38:29.019280] root 10.10.0.58 - READ "/etc/webmin/authentic-theme/custom-lang" 4719 [19/Sep/2015 13:38:29.019509] root 10.10.0.58 - READ "/etc/webmin/authentic-theme/custom-lang.en.UTF-8" 4719 [19/Sep/2015 13:38:29.020702] root 10.10.0.58 - READ "/usr/libexec/webmin//defaultacl" 4719 [19/Sep/2015 13:38:29.020952] root 10.10.0.58 - READ "/etc/webmin/root.acl" 4719 [19/Sep/2015 13:38:29.021193] root 10.10.0.58 - READ "/etc/webmin//root.acl" 4719 [19/Sep/2015 13:38:29.021422] root 10.10.0.58 - READ "/etc/sysconfig/network" 4719 [19/Sep/2015 13:38:29.021788] root 10.10.0.58 - CMD "cmd=hostname" 4719 [19/Sep/2015 13:38:29.025883] root 10.10.0.58 - READ "/usr/libexec/webmin/virtual-server/module.info" 4719 [19/Sep/2015 13:38:29.073748] root 10.10.0.58 - READ "/usr/libexec/webmin/server-manager/module.info" 4719 [19/Sep/2015 13:38:29.085647] root 10.10.0.58 - READ "/etc/webmin/module.infos.cache" 4719 [19/Sep/2015 13:38:29.108561] root 10.10.0.58 - READ "/etc/webmin/installed.cache" 4719 [19/Sep/2015 13:38:29.119455] root 10.10.0.58 - READ "/usr/libexec/webmin/acl/module.info" 4719 [19/Sep/2015 13:38:29.119916] root 10.10.0.58 - READ "/etc/webmin/webmin.cats" 4719 [19/Sep/2015 13:38:29.120124] root 10.10.0.58 - READ "/etc/webmin/webmin.descs" 4719 [19/Sep/2015 13:38:29.120768] root 10.10.0.58 - READ "/usr/libexec/webmin/change-user/module.info" 4719 [19/Sep/2015 13:38:29.121586] root 10.10.0.58 - READ "/etc/webmin/authentic-theme/favorites.json" 4719 [19/Sep/2015 13:38:29.132471] root 10.10.0.58 - READ "/usr/libexec/webmin//module.info" 4719 [19/Sep/2015 13:38:29.133591] root 10.10.0.58 - STOP "runtime=0"
Can you manually downgrade to let's say version 14 to see what will happen then? https://github.com/qooob/authentic-theme/raw/master/.build/authentic-theme-14.00.wbt.gz
I guess you could manually remove Webmin session by deleting /var/webmin/sessiondb.pag
I don't see how this issue is related to the theme.
I downgraded to version 14. Works without any problems as before.
Can you also clear the cookies? If not just run it in incognito mode, click Ctrl+Shift+N
(for Chrome), does it work then?
Tried the incognito mode v14: works v16: same behaviour, redirect loop
Reply from Firefox Forums:
To solve this problem there is no need to delete all of your cookies, simply delete the ones associated with your google and youtube accounts. Here's how:
- Click Tools/Options/Privacy
- Click "Remove Individual Cookies"
- In the Cookies Window type "accounts" (no quotes) into the Search box
- Select all the cookie names starting with: accounts.google.com and accounts.youtube.com
- Click Remove Cookie
You should now be able to reach your Gmail or Youtube login page as usual. This is how I was able to solve my problem. Deleting other google.com and youtube.com cookies did not help, only after I deleted the account cookies I was able to get back to the login page.
I tried different versions of Firefox in Linux, OS X, Windows - I can not reproduce the bug.
Please be kind making video screencast for the whole process of signin in, while keeping DevTools opened. (you can send it to me privately in case you are worried about URLs or make it in full-screen mode).
I have no clue why this is happening and I have very strong premonition that it's not theme related!
I hope your premonition is right :) Anyhow, I did a screen cast so you can see what I see, below is the link https://youtu.be/c7mjJFtxY9g
I've been doing some testing on various browsers, and so far I haven't been able to reproduce this issue.
I'll continue reviewing the issue, though I may ask Jamie to see if he has any input.
Ok, I think I know where the problem lies! You shouldn't use that module to change the theme.
Please goto Webmin->Webmin Configuration->Themes
. After you set Authentic Theme globally and see nice login (Authentic Theme login screen) - you will be fine!
Try it and let me know. Make sure that in Webmin->Change Language and Theme
you have default theme set first. Then do procedure above!
Helped?
Well... Tried the above, exactly as described by you. Now I get the nice looking login, but after the login goes again in the loop.
That should not be happening! :|
I can not reproduce it on and browser, in any OS.
Well I'm pretty sure it's not the browser... It's something happening within webmin+authentic theme. I will try running some stack traces on webmin se wtf is doing
Do you have any dependencies on any perl modules or anything within webmin ? What can differ between our environments is that mine is a fresh one, with webmin 1.760 installed. I've tried removing the theme completely also clean all traces of it from the fs/cache and installed directly the v16... still no go. cleaned again... installed v15.51... that one works fine.
That is so wired. Reboot the system!! I think personally, it's stuckness of Webmin cache somewhere..
Did you try Chrome?
Tried Firefox, Safari, Chrome ( Mac OS X ) v15.51 works fine v16.00 redirect loop I suspect some dependency to something which was left around in webmin before 1.760, because I installed directly the latest version... otherwise, don't know how big are the changes between 15.51 and 16. Forgot to mention, I also rebooted the webmin machine...
Changes are huge! Installed directly? Using Webmin->Webmin Configuration->Themes
and then install theme?
Yes
I don't know why it's happening. Will see what others will say on testing! Please wait..
I don't imagine you have a machine experiencing this issue, that we could remotely access and log in to? A test system, perhaps?
@qooob You are not alone there :) You can try to reproduce my environment: Clean CentOS7 install (up to date), then install the webmin-1.760 rpm... That should come with v14 of your theme... Just upgrade to the latest one... I think you will get the same results like me. I can also provide the list of rpm's installed.
@andreychek I can bring one machine up for testing and grant access to that one.. (if I will reproduce the issue)
@zamolxe what is the output of ls -lsaZ /usr/libexec/webmin/authentic-theme
and ls -lsaZ /etc/webmin/authentic-theme
, post it please!?
You have it below... selinux is disable on my end. ls -lsaZ /usr/libexec/webmin/authentic-theme /etc/webmin/authentic-theme/ /etc/webmin/authentic-theme/: total 12 drwxr-xr-x root root ? . drwxr-xr-x root root ? .. -rw-r--r-- root root ? config
/usr/libexec/webmin/authentic-theme: total 300 drwxr-xr-x root root ? . drwxr-xr-x root root ? .. -rwxr-xr-x root root ? authentic-lib.cgi -rwxr-xr-x root root ? authentic.pl -rwxr-xr-x root root ? buttons.cgi -rw-r--r-- root root ? CHANGELOG.md -rwxr-xr-x root root ? changelog.pl -rw-r--r-- root root ? config drwxr-xr-x root root ? images -rwxr-xr-x root root ? index.cgi drwxr-xr-x root root ? lang drwxrwxr-x root root ? lib -rw-r--r-- root root ? LICENSE -rwxr-xr-x root root ? navigation.cgi -rw-r--r-- root root ? README.md -rwxr-xr-x root root ? session_login.cgi -rwxr-xr-x root root ? settings.cgi -rwxr-xr-x root root ? settings-editor_read.cgi -rwxr-xr-x root root ? settings-editor_write.cgi -rwxr-xr-x root root ? settings-favorites_save.cgi -rwxr-xr-x root root ? settings-upload.cgi -rwxr-xr-x root root ? settings-upload_save.cgi -rwxr-xr-x root root ? sysinfo.cgi -rw-r--r-- root root ? theme.info drwxr-xr-x root root ? unauthenticated -rw-r--r-- root root ? VERSION.txt
Everything looks alright!
I've got some updates: Trying to query the root (/) ends up with the loop, but if I try to access only a module, in my case the dhcpd configuration (/dhcpd/) that works just fine. I suspect something in authentic-lib.cgi in "sub init {"...
What do you mean by Trying to query the root
? How do you query it?
ahh.. looking..
Ok, can you please find the line 2008 in authentic-lib.cgi
and add after &header($title)
the following:
prt($ENV{'REQUEST_URI'})
Now, please go to sub init
and completely comment out/remove this:
# Redirect user away, in case requested mode can not be satisfied
if ( $ENV{'REQUEST_URI'} ne '/'
&& $ENV{'REQUEST_URI'} ne '/?virtualmin'
&& $ENV{'REQUEST_URI'} ne '/?cloudmin'
&& $ENV{'REQUEST_URI'} ne '/?mail'
&& $ENV{'REQUEST_URI'} ne '/?dashboard'
&& index( $ENV{'REQUEST_URI'}, 'xhr' ) lt 0 )
{
my $webmin
= ( $ENV{'HTTPS'} ? 'https://' : 'http://' )
. $ENV{'HTTP_HOST'} . '/';
print "Location: $webmin\n\n";
}
elsif (
( $t_uri_virtualmin != -1 && !&foreign_available("virtual-server") )
|| ( $t_uri_cloudmin != -1 && !&foreign_available("server-manager") )
|| ($t_uri_webmail != -1
&& (&get_product_name() ne 'usermin'
|| ( &get_product_name() eq 'usermin'
&& !&foreign_available("mailbox") )
)
)
)
{
print "Set-Cookie: redirect=0; path=/\r\n";
my $webmin
= ( $ENV{'HTTPS'} ? 'https://' : 'http://' )
. $ENV{'HTTP_HOST'} . '/';
print "Location: $webmin\n\n";
}
Afterwards, reload the page. What is the output of the printed $VAR1
that we added to the line 2009
?
Are you sure is prt($ENV{'REQUEST_URI'}) and not print ?
Yes, I got that manual function there..
It's better to find printed stuff..
HTTP/1.0 500 Perl execution failed Server: MiniServ/1.760 Date: Sun, 20 Sep 2015 11:53:28 GMT Content-type: text/html; Charset=iso-8859-1 Connection: close Error - Perl execution failed
Undefined subroutine &main::prt called at /usr/libexec/webmin/authentic-theme/authentic-lib.cgi line 2009.
Like I said :) It's cache issue! In the end of the file there should be
sub prt {
my ($____v) = @_;
use Data::Dumper;
print '<div style="color: red">';
print Dumper $____v;
print '</div>';
}
Do you have this in your file? Did you really restart Webmin? Did you restart machines? (please say as it is, I need it for the future reference)
You could get error about not having use Data::Dumper
.
Take a look at this line: https://github.com/qooob/authentic-theme/blob/master/authentic-lib.cgi#L2358
OK instead of prt($ENV{'REQUEST_URI'})
do this:
print '++++' . $ENV{'REQUEST_URI'} . '++++';
What is the output?
In case you have /
from $ENV{'REQUEST_URI'}
, just manually update your .lib file and restart Webmin.
If I query https://1.2.3.4:10000 the output is / and the loop is gone.
But what do you query on regular basis?
This is what you should do?
Usually I go directly yo https://1.2.3.4:10000/ and sometimes directly to https://1.2.3.4:10000/dhcpd There is nothing fancy about the queries
Second will never work!!
What do you get on the print if you query the first example: https://1.2.3.4:10000/
?
Actually, if I query https://1.2.3.4:10000/fsdump/ it works without left menu. I wouldn't recommend doing this. Why not just query https://1.2.3.4:10000
... However https://1.2.3.4:10000/
works also just fine in my case.
Ok.. I found out the root cause... After installing perl-Data-Dumper to get your function working... Everything worked without any issues. Just to confirm, I removed perl-Data-Dumper... got back the loop... Can someone else confirm this ?
After updating to version 16, after the login gives the error The page isn't redirecting properly. Both Firefox and Midori give the error.
I'm using CentOS 7