Koozali.org: home of the SME Server
Obsolete Releases => SME 8.x Contribs => Topic started by: larieu on October 14, 2012, 09:27:08 PM
-
SME server 8 (from beta stage - but actualised to SME8 latest)
I had phpmyadmin from smecontribs 7
but I didn't use it
now I tried to update it to the latest one
yum update --enablerepo=smecontribs
(smecontribs point to sme8contribs; sme7contribs point to sme7 contribs)
after this and classical singnal-event....
I was not able to access any https webpage (https/mydomain/webmail, https://mydomain/wiki ...)
but
I found in logs errors that are related to php-pear-xxx
and I have tried
yum reinstall php-pear-Auth-SASL php-pear-Cache php-pear-Date php-pear-DB php-pear-File php-pear-HTTP php-pear-HTTP-Request php-pear-Log php-pear-Mail php-pear-Mail-Mime php-pear-MDB2 php-pear-Net-DIME php-pear-Net-FTP php-pear-Net-SMTP php-pear-Net-Socket php-pear-Net-URL php-pear-Services-Weather php-pear-SOAP php-pear-XML-Parser php-pear-XML-Serializer php-pear-XML-Util
php-pecl-Fileinfo
and then signal-event ....
now I am not able to login to server/manager
actually I receive on attached display this error (not in ssh - only on display)
httpd: bad group name www
on boot the only error is
Loading egw_domain.sql into mysql Failed
apparently the email, ftp, ldap works ok
I am able to access ssh but no webpages
in /var/log/httpd/admin_error_log
I have only (apparently apache is running)
[Sun Oct 14 22:07:07 2012] [warn] Init: Session Cache is not configured [hint: SSLSessionCache]
[Sun Oct 14 22:07:07 2012] [notice] Digest: generating secret for digest authentication ...
[Sun Oct 14 22:07:07 2012] [notice] Digest: done
[Sun Oct 14 22:07:08 2012] [notice] Apache configured -- resuming normal operations
I have a full backup (backup to workstation) but I am not able to access it through server-manager
and also an affa backup to another station
that means I can recover some files if I know which ones
How can I find more errors to be able to find the root of the problem?
-
more on this
I have mailarchiva installed and that seems to work ok
but no egroupware, no awstats,
from ssh if I try telnet on port 80
Trying 127.0.0.1...
telnet: connect to address 127.0.0.1: Connection refused
mailarchiva works on port 8090 and the rest on prot 80 / 443 don't
apparently this ports are blocked
how to enable access to this ports from console
-
I have found that also roamng profiles don't work (and samba is also down)
and also the ppp connection (this seems to timeout)
if I scan the server from LAn I found services only on ports
21,22,25,53,139,143,389,465,515,636,993
no 80 443
-
now I tried to update it to the latest one
yum update --enablerepo=smecontribs
(smecontribs point to sme8contribs; sme7contribs point to sme7 contribs)
after this and classical singnal-event....
I was not able to access any https webpage (https/mydomain/webmail, https://mydomain/wiki ...)
but
I found in logs errors that are related to php-pear-xxx
[/code]
you should file a bug, no contrib/update should do this
anyway, please report the errors and the list of updated rpms
and I have tried
yum reinstall php-pear-Auth-SASL php-pear-Cache php-pear-Date php-pear-DB php-pear-File php-pear-HTTP php-pear-HTTP-Request php-pear-Log php-pear-Mail php-pear-Mail-Mime php-pear-MDB2 php-pear-Net-DIME php-pear-Net-FTP php-pear-Net-SMTP php-pear-Net-Socket php-pear-Net-URL php-pear-Services-Weather php-pear-SOAP php-pear-XML-Parser php-pear-XML-Serializer php-pear-XML-Util
php-pecl-Fileinfo
why?
now I am not able to login to server/manager
actually I receive on attached display this error (not in ssh - only on display)
httpd: bad group name www
did you search the forums and bugzilla? I remember something about that..
do you have any other repo enabled other than centos/SME ones?
-
bug - I'll report
for moment I want to find some way to have apache on port 80 and roaming profile running
no other repo enabled (only standard ones)
httpd: bad group name www
Yes from there I had the idea to reinstall php-pear-xxxx
apparently apache start but some group perms ion / config files with changed permission don't let me to access it on standard ports
how to drop/restart firewall for test this?
-
bug - I'll report
asap, please..
for moment I want to find some way to have apache on port 80 and roaming profile running
if you report a bug, you'll also find the solution
no other repo enabled (only standard ones)
another good reason to open a bug ASAP
httpd: bad group name www
Yes from there I had the idea to reinstall php-pear-xxxx
you did not report the error message.. you made a decision, we don't know why, you broke your server and we are not able to help you if you don't help us
apparently apache start but some group perms ion / config files with changed permission don't let me to access it on standard ports
did I say you to open a bug NOW?
-
http://bugs.contribs.org/show_bug.cgi?id=7124
done
can at least someone tell mow to safely disable firewall (and the put it back) ?
to test if the problem is related to this
-
if you update only php-pear related rpms, firewall and its rules are untouched, so you can safely ignore it
-
unfortunately I didn't observed immediately that the servr-manager is broken
and I hat tried other things, too
but now I am not able to remember everything
-
history
will help you
-
that the problem
I don't understand why I have some missing history
The history show me only something like up to 2 days ago
and after the problem
and I am sure that I have done something else there
for example yum update --enablerepo=smecotnribs
I don't find it
or the long yum reinstall php-pear.... one
I have done something in that period which broken the server-manager
what I remember is that after signal-event
I was forced to go to IPKVM to issue one "cancel" after something like half an hour of not booting
there I have noticed the
httpd: bad group name www
otherwise I don't see it in any place
-
larieu
You might be better off rebuilding the server with a fresh install of sme8 and restoring from backup & reinstalling contribs, compared to the time & effort (& technical unsureness on your part) that will be involved to troubleshoot this & fix your problems.
Either way there will be considerable effort required.
If you do decide to do a restore, then do a test restore first (on different hard disks) to prove your backup is good, BEFORE wiping the current system.
-
my problem is that the users had tons of mails from the moment I have the last trusted backup (now are almost 4 days)
most probably I'll need to proceed like last time when I had a corrupted hdd
I'll prepare a new temporary one
over week-end I'll stop everything
I'll copy all relevant data from this server to the temp one
(users, mails, databases... = /home , /etc/e-smith , /etc/users.+ group + shadow... like affa does)
then when all is ok on the temporary one I'll use affa to put back on the real server
it exist any other verified method to move only users(+ userdata) + databases?
no any contrib/ or settings
on actual server I can do an USB backup
but I don't know if from backup I can recover only specific data
now just for testing
how safely drop firewall and the how to put ti back - to test this also
-
larieu
show output of
service httpd-e-smith status
service httpd-admin status
too see what iptables rules are configured see ouput of
iptables -L
-
# service httpd-e-smith status
down: /service/httpd-e-smith: 0s, want up
# service httpd-admin status
run: /service/httpd-admin: (pid 4184) 71551s, normally down; run: log: (pid 2114) 71568s
httpd-e-smith it is down
but
#service httpd-e-smith start
Starting httpd-e-smith: [ OK ]
# service httpd-e-smith status
down: /service/httpd-e-smith: 1s, want up
tail -f /var/log/httpd/admin_error_log
[Mon Oct 15 07:56:49 2012] [notice] Digest: done
[Mon Oct 15 07:56:49 2012] [notice] Apache configured -- resuming normal operations
[Mon Oct 15 08:09:09 2012] [notice] Graceful restart requested, doing restart
[Mon Oct 15 08:09:09 2012] [notice] Digest: generating secret for digest authentication ...
[Mon Oct 15 08:09:09 2012] [notice] Digest: done
even if I have issued restart
no relevant data in error log messages or in access log
the intricate thing is that my mailarchiva works (default contrib on port 8090)
that mean the base apache is up (httpd-admin)
I have used
iptables --list
( is equivalent - I think ) - everything seems ok there
-
Moving to contribs section of the Forums.
-
larieu
service httpd-e-smith status
down: /service/httpd-e-smith: 0s, want up
Now that web services has been started did you run a port scan again ?
AFAIU without the appropriate service running the ports will not be open (& accessible).
that mean the base apache is up (httpd-admin)
httpd-admin is your server manager
httpd-e-smith is your basic web server
Do you have another instance running ?
Show output of
service httpd status
When you say you did the classic "signal-event", that is rather unclear, as there are many possible signal-event's you could have done.
Do you mean
signal-event post-upgrade
signal-event reboot
???
-
A bug report has been opened in Bugzilla, it is being attended by a Developer, please follow there. Duplication of information is counter-productive.
Thanks.
-
Do you have another instance running ?
Show output of
service httpd status
Mary, if you want to help troubleshoot, please do so in the bug tracker. Not here. You are not helping by undermining the bug tracker.
"0s, want up" almost always indicates a service which is crashing immediately on startup.
"httpd: bad group name www" - this is very likely the core problem. Why does this system not have a group 'www'?
All followup to the bug tracker please.
Note: thread locked by moderator as requested.