Topic: server-manager and webmail not working after upgrade

[yelnic]

I recently upgraded a customer from 7 to 8 by doing a clean v8 install on a new machine and restoring the DAR2 backup from the v7 machine.  I had some issues with client PCs, so I left the v7 machine running until I could get back to it.  When I had time to complete the migration I did another restore of the DAR2 from the v7 machine and all seemed to go well.  Clients are happily using their email.

The problems start when I try to log into the server-manager as admin.  I get "invalid username or password" and I'm using the original combo that used to work.  I can ssh into the shell just fine and I reset the admin password using passwd to what it should be and still have no luck at all logging into the server-manager.

Also, when I try to access webmail I get a blank white screen.

From /var/log/httpd/error_log:

Code: [Select]

[Fri Sep 06 01:19:59 2013] [notice] SSL FIPS mode disabled
[Fri Sep 06 01:20:00 2013] [warn] RSA server certificate is a CA certificate (Ba
sicConstraints: CA == TRUE !?)
[Fri Sep 06 01:20:00 2013] [warn] RSA server certificate CommonName (CN) `mail.<domainname>.com' does NOT match server name!?
[Fri Sep 06 01:20:00 2013] [notice] Digest: generating secret for digest authent
ication ...
[Fri Sep 06 01:20:00 2013] [notice] Digest: done
[Fri Sep 06 01:20:03 2013] [notice] SSL FIPS mode disabled
[Fri Sep 06 01:20:03 2013] [warn] RSA server certificate is a CA certificate (Ba
sicConstraints: CA == TRUE !?)
[Fri Sep 06 01:20:03 2013] [warn] RSA server certificate CommonName (CN) `mail.<domainname>.com' does NOT match server name!?
[Fri Sep 06 01:20:03 2013] [notice] Apache configured -- resuming normal operati
ons
I can't find anywhere with a strange domain name.

I'm stumped.

[pmstewart]

yelnic

Have you tried logging in from mutiple machines or just one?? Using multiple browsers or just IE?? and if so what version of IE??

Have you actually tried logging in to the e-links browser from the console to see if that works?

[yelnic]

yelnic

Have you tried logging in from mutiple machines or just one?? Using multiple browsers or just IE?? and if so what version of IE??

Have you actually tried logging in to the e-links browser from the console to see if that works?

Thanks for the response.

I have tried multiple machines, Windows XP, Windows 7, Red Hat Linux 5, IE8, FF23, Chrome29.  Tried the e-links from the console too.  I have also scoured the forums and the documentation.

Any other ideas?

[janet]

yelnic

Did you search omn the error ?
"RSA server certificate CommonName (CN) `mail.<domainname>.com' does NOT match server name!?"

My (wild) guess is that the restore did not correctly or fully restore.
Check the db setting for CommonName

Maybe you should run the Configure this server from the console.

Did you restore to a clean installation of sme8 OS, the second time you did the restore ?

[yelnic]

yelnic

Did you search omn the error ?
"RSA server certificate CommonName (CN) `mail.<domainname>.com' does NOT match server name!?"

I searched on CommonName but didn't see anything that seemed to apply.

My (wild) guess is that the restore did not correctly or fully restore.
Check the db setting for CommonName

That's what I was thinking too.

Maybe you should run the Configure this server from the console.

Did you restore to a clean installation of sme8 OS, the second time you did the restore ?

No, I re-restored over the original restore and I have a fear that may be the cause of my problems.  Now I'm just looking for a solution.

I will attempt the console configuration.  If that doesn't work, do you have any other ideas?

Thanks for the reply.

[janet]

yelnic

I re-restored over the original restore...

Restore should ONLY be done to a clean install of sme 8 OS.
Ideally you should repeat the restore.
There may be a number of hidden issues depending what changes you made to the first install/restore.

You need to heed the error message(s) & fix the issue(s) identified. Take a look at the Certificates Concepts Howto re resetting the certificate.
As I already suggested check the db setting for CommonName.

[pmstewart]

yelnic,

not knowing your configuration or the full content you backed up and restored in the dar2 backup......it is difficult to know exactly why the second overwrite brought problems.

You installed clean sme8 and restored only part of the data a small amount at a time?? or all at once?? .....did you experience any problems the first go around??

I think Janet onto something as I've experienced the same cn problem before.....

[pmstewart]

I'm also experiencing problems logging into the console of at least one machine after updates even though I can ssh into the box and the web interface will let me in......bizzare to say the least......LOL

[yelnic]

yelnic,

not knowing your configuration or the full content you backed up and restored in the dar2 backup......it is difficult to know exactly why the second overwrite brought problems.

You installed clean sme8 and restored only part of the data a small amount at a time?? or all at once?? .....did you experience any problems the first go around??

I think Janet onto something as I've experienced the same cn problem before.....

I did a full restore again.  The first time it worked fine, but the clients weren't set up so I left the v7 server running instead of the v8 server.  When I was ready to change all twenty client machines I did another full backup on the v7 machine and a full restore (again) onto the v8 machine.  That's when the trouble started.

I can't find a command for the db settings for CommonName.  I've looked at the db command tutorial, but don't know how to do db for CN.

Thanks.

[janet]

yelnic

The 23rd forum search result for CommonName, and the answer is also in one of the Certificate howtos.
http://forums.contribs.org/index.php/topic,45241.msg219420.html#msg219420

I can't find a command for the db settings for CommonName.  I've looked at the db command tutorial, but don't know how to do db for CN.

[yelnic]

yelnic

The 23rd forum search result for CommonName, and the answer is also in one of the Certificate howtos.
http://forums.contribs.org/index.php/topic,45241.msg219420.html#msg219420

I did the db commands and still no luck.

In /var/log/admin_error_log I found the following information:

Code: [Select]

[Sun Sep 08 03:44:43 2013] [error] [client 127.0.0.1] Can't exec "/usr/lib64/htt
pd/modules/pwauth": Permission denied at /etc/e-smith/web/common/cgi-bin/login l
ine 57., referer: https://mail.<domain name>.com/server-common/cgi-bin/login
[Sun Sep 08 03:44:43 2013] [error] [client 127.0.0.1] Could not open pipe to pwa
uth: Permission denied at /etc/e-smith/web/common/cgi-bin/login line 59., refere
r: https://mail.<domain name>.com/server-common/cgi-bin/login
[Sun Sep 08 03:45:03 2013] [error] [client 127.0.0.1] Can't exec "/usr/lib64/htt
pd/modules/pwauth": Permission denied at /etc/e-smith/web/common/cgi-bin/login l
ine 57., referer: https://mail.<domain name>.com/server-common/cgi-bin/login?ba
ck=http%3a%2f%2fmail.priceattorneys.com%2fserver-manager%2f
[Sun Sep 08 03:45:03 2013] [error] [client 127.0.0.1] Could not open pipe to pwa
uth: Permission denied at /etc/e-smith/web/common/cgi-bin/login line 59., refere
r: https://mail.<domain name>.com/server-common/cgi-bin/login?back=http%3a%2f%2
fmail.priceattorneys.com%2fserver-manager%2f
[Sun Sep 08 04:05:46 2013] [error] [client 127.0.0.1] Can't exec "/usr/lib64/htt
pd/modules/pwauth": Permission denied at /etc/e-smith/web/common/cgi-bin/login l
ine 57., referer: https://mail.<domain name>.com/server-common/cgi-bin/login?ba
ck=http%3a%2f%2fmail.<domain name>.com%2fserver-manager%2f
I checked the permissions on /usr/lib64/httpd/modules/pwauth and they are the same (-r-sr-x---) as on another v8 install.

I looked at /etc/e-smith/web/common/cgi-bin/login (-rwxr-xr-x) and found this:

Code: [Select]

sub validate
{
  my ($username, $password) = @_;
  my $pwauth = -f "/usr/lib64/httpd/modules/pwauth" ? "/usr/lib64/httpd/modules/
pwauth" : "/usr/lib/httpd/modules/pwauth";
  unless (open(PWAUTH, "|$pwauth"))   <---<< Line 57
  {
    warn "Could not open pipe to pwauth: $!";
    return 0;
  }

Again, I checked it against another v8 install and it's the same.

Where can I go from here?

Thanks.

[janet]

yelnic

I advise you to lodge a bug report, the link is at the top of these forums.
You will have to be systematic with your answers & provide all details requested.
Many things have happened to your server & a more ordered diagnostic approach is needed to unravel your issue(s).

Please report the bug number back here.

[yelnic]

I advise you to lodge a bug report, the link is at the top of these forums.
You will have to be systematic with your answers & provide all details requested.
Many things have happened to your server & a more ordered diagnostic approach is needed to unravel your issue(s).

Please report the bug number back here.

I just remembered the one thing I have not had a chance to try is a configure through the console (I haven't been on site yet).  I will try that today and if I don't get good results I will lodge a bug report.

Thanks for all of your help.

[pmstewart]

For everyone's sake I hope you can log in at the console......I am going onsite to several of my locations tomorrow to see if console login works since the server I've upgraded and a test server here at my office don't allow me to log in at the console......

Good luck

**turns out there is only one server which has the console (tty1) login problem**

[janet]

yelnic

As you are going to lodge a bug report, then it is probably best you do not make any more configuration changes.
You erase the forensic trail by making more changes, so I suggest you report the issue first in bugzilla & only do what is requested of you by the troubleshooters & developers.

Do you have remote ssh access to this server ? Having that setup or allowed certainly saves you having to go on site for every little change or tweak or yum upgrade.

What was the sme 7.x version you upgraded from ?
It is recommended to upgrade to tge latest sme7.6 before doing a sme 8.0 upgrade.

Also what were the issues you had with your clients the first time you upgraded ? You said you needed to tweak the clients (I assume that means Window workstations), so instead you reverted to sme 7 ? Something was obviously wrong with your backup or restore or upgrade procedure right from the very start and this could be the underlying cause of your issues. This cannot be ignored when diagnosing what is going on here.

Also what contribs or other packages do you have installed ?
Have you checked that these are compatible with sme 8 ?

Also what custom changes eg custom templates & similar tweaks have you made to your sme 7 server, & have you reviewed and/or removed these in sme 8 ?

To pmstewart, you should lodge a seperate bug report & stop tacking on your problem to an existing forum thread. It may sound similar but the reasons could be different, so please lodge a bug report also giving as much detail as possible. Also provide the bug report number here.

[yelnic]

As you are going to lodge a bug report, then it is probably best you do not make any more configuration changes.
You erase the forensic trail by making more changes, so I suggest you report the issue first in bugzilla & only do what is requested of you by the troubleshooters & developers.

Done:  http://bugs.contribs.org/show_bug.cgi?id=7838

Do you have remote ssh access to this server ? Having that setup or allowed certainly saves you having to go on site for every little change or tweak or yum upgrade.

I do, but I can't find a terminal setting that displays the console correctly.  The console is completely jumbled, but the bash shell works fine.

What was the sme 7.x version you upgraded from ?
It is recommended to upgrade to tge latest sme7.6 before doing a sme 8.0 upgrade.

Yes, it was the very latest, I believe v7.6.

Also what were the issues you had with your clients the first time you upgraded ? You said you needed to tweak the clients (I assume that means Window workstations), so instead you reverted to sme 7 ? Something was obviously wrong with your backup or restore or upgrade procedure right from the very start and this could be the underlying cause of your issues. This cannot be ignored when diagnosing what is going on here.

I ran out of time to convert the Windows clients (20 of them) to the new settings for v8, FQDN, SSL, etc.

I really didn't "revert", I just left the old hardware running 7.6 and shut the new machine down for the day.

Also what contribs or other packages do you have installed ?
Have you checked that these are compatible with sme 8 ?

DAR2
User Vacations
AWStats
SARG Reports
User Panel Access
sme7admin

I believe everything was ok except for sme7admin when I checked them.  I removed sme7admin and went on.

Also what custom changes eg custom templates & similar tweaks have you made to your sme 7 server, & have you reviewed and/or removed these in sme 8 ?

I don't know enough about making custom templates or anything like that.  If it's not a contrib pkg, I don't do it.

I'd like to take a moment to thank you for all your time and input on this matter.

[janet]

yelnic

You need to tell the whole story in the bug report. Add a link to this forum thread & also add all the commands you issued when trying to fix the problem, you can see them in your bash history.

I can't find a terminal setting that displays the console correctly.  The console is completely jumbled, but the bash shell works fine.

There is a fix available for this, reported recently in the forums I think, caused by recent updates,  search on bash.

I ran out of time to convert the Windows clients (20 of them) to the new settings for v8, FQDN, SSL, etc.

This concerns me. Why was it necessary to change settings on workstations ? The backup & restore should have transferred all the original settings to the new server, that"s the whole point of doing it. Either something went very wrong or you are ignorant of the backup/restore concepts & requirements. Refer to tte Backup Concepts howto.
You say you used DAR2, do you mean the DAR2 contrib originally developed by Darrell May ?
Did you use/create a full system recovery type backup job ?

Did you install the DAR2 contrib on the new v8 server before you did the restore ?
Did you transfer the job/db details as per the DAR2 wiki article ?
Where was the backup, on a USB drive ?

All this info should go in the bug report so the experts are aware of the full history.

[yelnic]

You need to tell the whole story in the bug report. Add a link to this forum thread & also add all the commands you issued when trying to fix the problem, you can see them in your bash history.

I'm afraid they are probably all jumbled up as I have issued some commands repeatedly.  I will take a look.

This concerns me. Why was it necessary to change settings on workstations ? The backup & restore should have transferred all the original settings to the new server, that"s the whole point of doing it. Either something went very wrong or you are ignorant of the backup/restore concepts & requirements. Refer to tte Backup Concepts howto.
You say you used DAR2, do you mean the DAR2 contrib originally developed by Darrell May ?
Did you use/create a full system recovery type backup job ?

They were using the server IP address and ports 110/25.  It is my understanding that on v8 we need to be using the FQDN and ports 993/465, etc.

I ran the Darrell May contrib and followed the instructions to do a full system recovery, I believe.

Did you install the DAR2 contrib on the new v8 server before you did the restore ?
Did you transfer the job/db details as per the DAR2 wiki article ?
Where was the backup, on a USB drive ?

I followed the instructions I had that had me install DAR2 and copy the job/db files over for the restore.
I backed up to a directory on a RHEL5 server that we use for nightly DAR2 backups, but made a separate backup for this case.  I didn't have a USB drive handy.

All this info should go in the bug report so the experts are aware of the full history.

It's in there now as I just answered a bunch of questions in the bug report.

Thanks.

[janet]

yelnic

Assuming you have logged in as root do
cat .bash_history

post to the bug report

[janet]

yelnic

...I can't find a terminal setting that displays the console correctly.  The console is completely jumbled, but the bash shell works fine.

Please learn to search the forums & bugzilla for any issues before posting them here.
Many issues have been answered by the time you see the problem.

I think it could be answered by this recent post.
http://forums.contribs.org/index.php/topic,50098.msg251454.html#msg251454

[yelnic]

Please learn to search the forums & bugzilla for any issues before posting them here.
Many issues have been answered by the time you see the problem.

I think it could be answered by this recent post.
http://forums.contribs.org/index.php/topic,50098.msg251454.html#msg251454

I take some pride in using the forums to answer most of my questions without ever posting here.  I would not have found this post since I don't use midnight commander or putty, but I thank you for the information.  It worked to clean up my console display.

[janet]

yelnic

It worked to clean up my console display.

It seems that is only a temporary fix to undo some of the effects of a previous upgrade change.
There is a bigger issue going on in the bug reports & devinfo list.
See
http://forums.contribs.org/index.php/topic,50183.msg252037.html#msg252037

I see in the associated bug report for your original issue, that two main repos were not enabled.
Did you enable them & run the yum update.
What outcome ?

[yelnic]

It seems that is only a temporary fix to undo some of the effects of a previous upgrade change.
There is a bigger issue going on in the bug reports & devinfo list.
See
http://forums.contribs.org/index.php/topic,50183.msg252037.html#msg252037

I see in the associated bug report for your original issue, that two main repos were not enabled.
Did you enable them & run the yum update.
What outcome ?

I just used putty and did the UTF-8 change.  It works for the few times I may have to access the console remotely and I didn't want to change the default behavior of bash.

I have enabled the repositories, but have not done the yum update yet since I have not been able to make the default backup work.  I was using the (apparently old) DAR2 which worked fine to the RHEL5 machine, but the default DAR won't.  I am going to try to do a USB console backup on site tomorrow before doing the yum update just in case something goes wrong.

Thanks.

[janet]

yelnic

I wonder if this bug identifies your issues
http://bugs.contribs.org/show_bug.cgi?id=7806
It would be consistent I think with having wrong packages installed due to main base & updates repos not being enabled.

[yelnic]

I wonder if this bug identifies your issues
http://bugs.contribs.org/show_bug.cgi?id=7806
It would be consistent I think with having wrong packages installed due to main base & updates repos not being enabled.

Yes, it looks very similar.  My www group disappeared from /etc/groups which I assumed at the time caused the files in /home/httpd/html/horde/config/ to be assigned group apache.  After I manually added the www group back into /etc/groups and changed the group in /home/httpd/html/horde/config/ to www (based on comparing the group settings in another clean v8 install), webmail started working again.

Thanks.

[pmstewart]

Janet,

Was not trying to tack a new bug onto someone else's post - rather indicating this series of updates comes with multiple issues.

http://bugs.contribs.org/show_bug.cgi?id=7835