Topic: mitel-networks-server login? not MSE 7.3???

[rm]

I am now looking at a "mitel-networks-server" login prompt and I am lost.

Running SME 7.3, yesterday, in an attempt to free some disk space, I moved several projects from the server to DVD, and looked at installed modules. In a opportunistic mode I uninstalled (individual packages) several ispell languages... That didn't work out too well.

After a reboot the server is no longer accessible as it was. So I put in the CD and ran the installer. However, it was the 7.2 installer. After that I ran the 7.3 installer (fresh download).

Now I'm looking at a "mitel-networks-server" login prompt. Login as 'admin' is denied (text flashes for a microsecond before returning to login prompt), login as 'root' does work. Yum doesn't work "network is unreachable".

Can anyone help me along getting the system up and running properly again?

[rm]

To specify, the machine is only accessible via keyboard/monitor. The machine cannot be reached remote. So I cannot get into my samba/e-smith/shares which is really the problem I'm now trying to solve.

[cactus]

To specify, the machine is only accessible via keyboard/monitor. The machine cannot be reached remote. So I cannot get into my samba/e-smith/shares which is really the problem I'm now trying to solve.

I suggest you launch a bug in the bugtracker.

[Stefano]

Hi

I suspect that the OP has done something like

Code: [Select]

yum remove rpmname

and because of dependencies, many of e-smith-* and smeserver-* rpms are gone..

it's not the first time I see this situation.

my 2c
ciao
Stefano

[cactus]

Hi

I suspect that the OP has done something like

Code: [Select]

yum remove rpmname

and because of dependencies, many of e-smith-* and smeserver-* rpms are gone..

it's not the first time I see this situation.

my 2c
ciao
Stefano

Such a failure should be fixed by trying to reinstall from CD like OP did.

[Stefano]

Such a failure should be fixed by trying to reinstall from CD like OP did.

of course

[rm]

Thanks for looking into this. The machine seems to boot normal, however, it hasn't got any network. 'ifconfig eth0' replies 'Device not found'. (Onboard Intel in a Dell Poweredge sc440).

I have already booted from 7.3 CD several times, no failures, always the choice 'install new' or 'upgrade existing MSE 7.3', all to no avail.

[cactus]

Thanks for looking into this. The machine seems to boot normal, however, it hasn't got any network. 'ifconfig eth0' replies 'Device not found'. (Onboard Intel in a Dell Poweredge sc440).

I have already booted from 7.3 CD several times, no failures, always the choice 'install new' or 'upgrade existing MSE 7.3', all to no avail.

And what did you choose?

[rm]

Each time the choice was to upgrade existing install, since I would guess that a new install would ruin the data that is still on the hard disk, like a few gig of project files.

So, how do I return to it's original configuration, or at least get the network card to communicate..?

[David Harper]

If all you need to rescue is files, then why not try this:

1. Boot to single user mode as described here

2. Attach a USB hard drive and mount it

3. Back up your files to the HDD

4. Power down the server

5. Reinstall

6. Copy the files back on once the reinstall is complete

[pfloor]

Each time the choice was to upgrade existing install, since I would guess that a new install would ruin the data that is still on the hard disk, like a few gig of project files.

So, how do I return to it's original configuration, or at least get the network card to communicate..?

Have you logged in as admin and chosen "Configure This Server" and made sure all the settings are correct?

[FreakWent]

Have you logged in as admin and chosen "Configure This Server" and made sure all the settings are correct?

No he hasn't:
"Login as 'admin' is denied (text flashes for a microsecond before returning to login prompt)"

[rm]

If all you need to rescue is files, then why not try this...

Thanks for the suggestion, and yes, I did grab the major projects via an USB-stick, they are all still in /home/e-smith/files/ibays/
I don't have an USB hard disk at hand, but this would probably also work by means of an extra sata HDD..?

However, the machine did function as PDC for a Windows domain, implying that several people have roaming profiles with e-mail and application settings. My thought is that reinstalling does imply that their profiles get lost. Would that be a correct assumption from my end?

The thing is, loggin into the server, it is now named 'mitel-networks-server' where before this it was named differently. Hence my post title. I'm not a Linux expert (duh, I would not be in this mess to start with), but the situation seems to suggest that the initial configuration (i.e. server name, possibly even the users) are overwritten or lost...

[David Harper]

The roaming profiles are stored in /home/e-smith somewhere. If you dig around you will find them -- or I can find them for you if you wait until tomorrow when I get back to my test server

Anyway, when you have recreated the user accounts on the new server, you can copy their roaming profiles back into the correct location.

Sure, this would probably work with an extra HDD. Just so long as you don't extend the RAID array to it.

[dgs]

Thanks for looking into this. The machine seems to boot normal, however, it hasn't got any network. 'ifconfig eth0' replies 'Device not found'. (Onboard Intel in a Dell Poweredge sc440).

I'm curious here, I've never been able to get correct operation from the onboard NIC on an SC440 with 7.3, the setup identifies the card but no data is ever transferred. The Dell SC430 and T105 onboard NICs work perfectly but I've never had success with SC440.

Have you ever had activity across this NIC?

[David Harper]

Samba profiles are stored in /home/e-smith/files/samba/profiles/

If you restored them, you would need to change their user & group ownerships from root:root to user:user for each profile, to allow the users to access and save their profiles.

Code: [Select]

chown -R exampleuser exampleprofile
chgrp -R exampleuser exampleprofile


[rm]

Have you ever had activity across this NIC?

Yes I have. As it was shipped by dell, I just pushed the CD in and installed. At the time it was 7.1 or 7.2, and now ... hold on. Could it be that my install was still at 7.2, since I did burn a 7.3 CD after my action that initiated this post?

Where/How can I determine when 7.3 was installed on the server? Would the logs show that? And where do I need to search?

[David Harper]

/root/install.log should contain your installation log. The file date will tell you when you installed SME Server.

Code: [Select]

[root@xerxes ~]# ls -l /root/install.log
-rw-r--r--  1 root root 24187 Oct  8 07:45 /root/install.log
[root@xerxes ~]#


[rm]

I have been looking at these files, the /root/install.log only logs the most recent actions, and seems to be overwritten by each install. No news there.

However, from upgrade.log.20081106084133
Upgrading 23 packages, amongst which the un-installed aspells an locale files (de, es, fr, it, sv).
warning: /home/e-smith/db/accounts created as /home/e-smith/db/accounts.rpmnew
warning: /home/e-smith/db/configuration created as /home/e-smith/db/configuration.rpmnew

That's interesting... in /home/e-smith/db/ amongst others I find(have to type them over):
accounts - Nov 6 - 1237 byte
accounts.rpmnew - Jul 4 - 0 byte
accounts.rpmsave - Nov 1 - 5923 byte
configuration - Nov 6 - 6708 byte
configuration.rpmnew - Jul 4 - 0 byte
configuration.rpmsave - Nov 6 - 8319 byte
networks - Dec 24 2007 - 0 byte
networks.rpmsave - Oct 31 - 298  byte

Each file states "DO NOT MODIFY THIS FILE"... yet the .rpmsvae files seem to hold (parts of ?) the initial configuration, with network settings that are now not functional. Can I (Since I do not have access to the admin menu) fiddle around here? Rename the .rpmsave's ?

[dgs]

Yes I have. As it was shipped by dell, I just pushed the CD in and installed. At the time it was 7.1 or 7.2, and now ... hold on. Could it be that my install was still at 7.2, since I did burn a 7.3 CD after my action that initiated this post?

Where/How can I determine when 7.3 was installed on the server? Would the logs show that? And where do I need to search?

I've had no luck at with the onboard NIC, I don't think I have tried a SC440 with an SME version prior to 7.3. Whenever I have had problem I've just added a NIC and got on with things.

What I've observed could have no relation to your problem whatsoever, and I do not wish to waste anyones time, but if you are still at an impass and wish to bang your head on the monitor maybe try another NIC and a quick re-run of the setup. Just a thought.

[rm]

... maybe try another NIC and a quick re-run of the setup.

Bright idea, so I pulled a sturdy 3com 3c900 combo 10Mbit of the shelve. This however doesn't change anything, still can't login as admin, neither eth0 nor eth1 are configured.

[rm]

So where am I now:
- resetting password of admin/root doesn't change anything, no eth0/eth1, no login as admin possible
- no backup to external USB device (via admin screen) possible
- so, disk content copied to a (excuse me) WinXP machine (thx explore2fs)

Unless someone has a bright idea to get the configuration back to live, I'll reinstall a fresh version of SME (hoping to get the inboard NIC operational). Then I'll be occupied copying and chowning a lot.

[pfloor]

Have you tried:

Log in as root then "su admin"
Log in as root then "/sbin/e-smith/console"

[rm]

Log in as root then "su admin"

This causes a screen full of code, possible something is on top of this, as the line below is at the top of the screen, with the cursor right at the very bottom.

"Can't locate esmith/Backup.pm in @INC (@INC contains: [... and then a lot of paths, half a screen full of them...]) at /usr/lib/perl5/site_perl/esmith/console/perform_backup.pm line 8.
BEGIN failed--compilation aborted at /usr/lib/perl5/site_perl/esmith/console/perform_backup.pm line 8.
Compilation failed in require at /sbin/e-smith/console-menu-items/usbBackup.pl line 2.
BEGIN failed--compilation aborted at /sbin/e-smith/console-menu-items/usbBackup.pl line 2.
Compilation failed in require at /sbin/e-smith/console line 61."

I remain logged in as root. Does this ring a bell?

[cactus]

This causes a screen full of code, possible something is on top of this, as the line below is at the top of the screen, with the cursor right at the very bottom.

"Can't locate esmith/Backup.pm in @INC (@INC contains: [... and then a lot of paths, half a screen full of them...]) at /usr/lib/perl5/site_perl/esmith/console/perform_backup.pm line 8.
BEGIN failed--compilation aborted at /usr/lib/perl5/site_perl/esmith/console/perform_backup.pm line 8.
Compilation failed in require at /sbin/e-smith/console-menu-items/usbBackup.pl line 2.
BEGIN failed--compilation aborted at /sbin/e-smith/console-menu-items/usbBackup.pl line 2.
Compilation failed in require at /sbin/e-smith/console line 61."

I remain logged in as root. Does this ring a bell?

Ohoh, that does not look very good. That file should be provided by e-smith-backup, one of the core packages of SME Server. Is it installed on your server?

Code: [Select]

rpm -q e-smith-backup

[rm]

Is it installed on your server?

Code: [Select]

rpm -q e-smith-backup

Yes, "e-smith-base-4.18.0-81.el4.sme"

[cactus]

Yes, "e-smith-base-4.18.0.81.el4.sme"

I made a typo, you might have not seen in your quick reply:

Code: [Select]

rpm -q e-smith-backup

[rm]

Code: [Select]

rpm -q e-smith-backup

package e-smith-backup is not installed

And by the way, just checked your website, and I have/had your subversion running on the server as well, nice job! thanks for providing that!

[cactus]

package e-smith-backup is not installed

Strange, does this do anything for you:

Code: [Select]

yum update

And by the way, just checked your website, and I have/had your subversion running on the server as well, nice job! thanks for providing that!

Your welcome, you are aware that the latest version can be downloaded from the smecontribs repository?

[rm]

Strange, does this do anything for you:

Code: [Select]

yum update

ifconfig only shows a loopback, both eth0 and eth1 aren't configured. Rerunning the CD doesn't change that.

Your welcome, you are aware that the latest version can be downloaded from the smecontribs repository?

I'm not sure. Installing is several months ago, however, during one update it got through via contribs iirc.

[pfloor]

In a opportunistic mode I uninstalled (individual packages) several ispell languages... That didn't work out too well.

How did you "uninstall (individual packages)"? Via software panel, via command line with yum, or via CL with rpm -e?

Which packages did you uninstall?  If you used yum, look at your yum logs.

Note: Yum has no mercy and will remove dozens of packages that you don't ask it to remove due to dependencies.  I suspect you used the software panel or yum and accidentally removed some core packages, your yum logs will tell you what was removed, let's start there.  Post the last part of your yum logs showing what was removed/added from the time the problem started on forward.

[rm]

Packages were removed using admin panel. Disk was running full, so projects were moved to DVD. It should have stopped there.
However, next I had a look at installed modules, the option 'selected individual modules/packages' was checked. I noticed the ispell for languages I didn't need (wandered how much MB of space that would offer) so selected several language files (except 'en' and possibly 'nl') and pressed submit. That's it.

I do remember being surprised, because the web based server-manager wasn't accessible after that. Being on a WinXP logged into the PrimaryDomainController the box provides, all shares were still available. Also the SSH terminal seemed to work just fine for root.

At the next reboot (of WinXP) the local profile was loaded (happens every now and then, but goes away after a day or so) instead of loading the profile of the server. Next the shares weren't available. That's when I started to get sweaty palms

[Stefano]

Hi

I suspect that the OP has done something like

Code: [Select]

yum remove rpmname

and because of dependencies, many of e-smith-* and smeserver-* rpms are gone..

it's not the first time I see this situation.

my 2c
ciao
Stefano

I quote myself here because I've already seen this situation..

solved grepping removed rpms from yum log, then downloading them from another pc, installing on the server again..

configuration dbs are there, in *.rpmsave files in /home/e-smith/db..

so, these are the steps:
- search in yum.log for removed rpms
- download them from contribs.org into, for example, a usb key
- move rpms to sme server, then install them with "yum localinstall"
- rename all /home/e-smith/db/*.rpmsave to original filename
- signal-event post-upgrade; signal-event reboot
- check if everything is working, then do a yum update

HTH
Ciao

Stefano

p.s. the problem is smeserver-support rpm that is "bond" to every e-smith-*/smeserver-* rpm.. if you remove it, because of dependencies, almost every e-smith-*/smeserver-* rpm will be removed

[pfloor]

Packages were removed using admin panel. Disk was running full, so projects were moved to DVD. It should have stopped there.
However, next I had a look at installed modules, the option 'selected individual modules/packages' was checked. I noticed the ispell for languages I didn't need (wandered how much MB of space that would offer) so selected several language files (except 'en' and possibly 'nl') and pressed submit. That's it.

I do remember being surprised, because the web based server-manager wasn't accessible after that. Being on a WinXP logged into the PrimaryDomainController the box provides, all shares were still available. Also the SSH terminal seemed to work just fine for root.

At the next reboot (of WinXP) the local profile was loaded (happens every now and then, but goes away after a day or so) instead of loading the profile of the server. Next the shares weren't available. That's when I started to get sweaty palms

OK, enough explaining yourself.  2 of us have suspected that you removed vital packages and need to restore them.  Let's find out if Stefano and I are correct.

Do the following and post the results:

Code: [Select]

yum list e-smith* smeserver*

[rm]

so, these are the steps:
- search in yum.log for removed rpms
- download them from contribs.org into, for example, a usb key
- move rpms to sme server, then install them with "yum localinstall"
- rename all /home/e-smith/db/*.rpmsave to original filename
- signal-event post-upgrade; signal-event reboot
- check if everything is working, then do a yum update

SERVER IS BACK ONLINE..!!!

Just grant, it's solved by stepping through the list by stefano.

A bit from the yum logs:

Dependencies Resolved

=============================================================================
 Package                 Arch       Version          Repository        Size
=============================================================================
Removing:
 aspell-da               i386       50:0.50-12.2.2.el4.sme  installed          12 M
 aspell-de               i386       50:0.50-11.2.2.el4.sme  installed         9.4 M
 aspell-el               i386       50:0.50-4.2.2.el4.sme  installed          12 M
 aspell-es               i386       50:0.50-13.2.2.el4.sme  installed          11 M
 aspell-fr               i386       50:0.50-9.2.2.el4.sme  installed          17 M
 aspell-id               i386       50:0.50.1-4.2.2.el4.sme  installed         396 k
 aspell-it               i386       50:0.53-4.2.1.el4.sme  installed          39 M
 aspell-pt               i386       50:0.50-10.2.2.el4.sme  installed         8.2 M
 aspell-sl               i386       50:0.50-1.2.1.el4.sme  installed          22 M
 aspell-sv               i386       50:0.51-1.2.1.el4.sme  installed         3.4 M
Removing for dependencies:
 e-smith-LPRng           noarch     1.14.0-8.el4.sme  installed          50 k
 e-smith-apache          noarch     1.2.0-16.el4.sme  installed          44 k
 e-smith-backup          noarch     1.14.0-22.el4.sme  installed         107 k
 e-smith-base            i386       4.18.1-10.el4.sme  installed         601 k
 e-smith-domains         noarch     1.4.0-10.el4.sme  installed          41 k
 e-smith-dynamicdns-dyndns  noarch     1.4.0-3.el4.sme  installed          19 k
 e-smith-dynamicdns-dyndns.org  noarch     1.4.0-4.el4.sme  installed          19 k
 e-smith-dynamicdns-tzo  noarch     1.4.0-3.el4.sme  installed          18 k
 e-smith-dynamicdns-yi   noarch     1.4.0-3.el4.sme  installed          18 k
 e-smith-email           noarch     4.18.0-2.el4.sme  installed         214 k
 e-smith-flexbackup      noarch     1.10.0-3.el4.sme  installed          31 k
 e-smith-horde           noarch     1.13.0-24.el4.sme  installed          86 k
 e-smith-hosts           noarch     1.14.0-12.el4.sme  installed          70 k
 e-smith-ibays           noarch     1.2.0-6.el4.sme  installed          68 k
 e-smith-imp             noarch     1.13.0-11.el4.sme  installed          64 k
 e-smith-ingo            noarch     1.1-6.el4.sme    installed          31 k
 e-smith-ldap            noarch     4.12.0-9.el4.sme  installed          49 k
 e-smith-mysql           noarch     1.12.0-15.el4.sme  installed          29 k
 e-smith-netatalk        noarch     1.14.0-6.el4.sme  installed          23 k
 e-smith-ntp             noarch     1.16.0-16.el4.sme  installed          57 k
 e-smith-oidentd         noarch     1.2.0-3.el4.sme  installed          20 k
 e-smith-packetfilter    noarch     1.18.0-6.el4.sme  installed          46 k
 e-smith-php             noarch     1.12.0-4.el4.sme  installed          24 k
 e-smith-pop3            noarch     1.2.0-3.el4.sme  installed          23 k
 e-smith-portforwarding  noarch     1.2.0-8.el4.sme  installed          29 k
 e-smith-pptpd           noarch     1.12.0-11.el4.sme  installed          28 k
 e-smith-proftpd         noarch     1.12.0-11.el4.sme  installed          32 k
 e-smith-proxy           noarch     4.14.0-13.el4.sme  installed          32 k
 e-smith-qmail           noarch     1.10.0-16.el4.sme  installed          45 k
 e-smith-qmailanalog     noarch     1.12.0-3.el4.sme  installed          27 k
 e-smith-quota           noarch     1.10.0-10.el4.sme  installed          80 k
 e-smith-radiusd         noarch     1.0.0-16.el4.sme  installed          67 k
 e-smith-starterwebsite  noarch     1.0.0-3.el4.sme  installed          30 k
 e-smith-tinydns         noarch     1.0.0-7.el4.sme  installed          40 k
 e-smith-turba           noarch     1.7.0-17.el4.sme  installed          38 k
 e-smith-viewlogfiles    noarch     1.8.0-10.el4.sme  installed          17 k
 smeserver-locale-da     noarch     1.4.0-8.el4.sme  installed         201 k
 smeserver-locale-de     noarch     1.4.0-8.el4.sme  installed         221 k
 smeserver-locale-el     noarch     1.4.0-8.el4.sme  installed         217 k
 smeserver-locale-es     noarch     1.4.0-8.el4.sme  installed         218 k
 smeserver-locale-fr     noarch     1.4.0-8.el4.sme  installed         226 k
 smeserver-locale-id     noarch     1.4.0-8.el4.sme  installed         205 k
 smeserver-locale-it     noarch     1.4.0-8.el4.sme  installed         204 k
 smeserver-locale-pt_BR  noarch     1.4.0-8.el4.sme  installed         211 k
 smeserver-locale-sl     noarch     1.4.0-8.el4.sme  installed         193 k
 smeserver-locale-sv     noarch     1.4.0-8.el4.sme  installed         207 k
 smeserver-qpsmtpd       noarch     1.2.1-57.el4.sme  installed          52 k
 smeserver-spamassassin  noarch     1.4.0-18.el4.sme  installed         7.2 k
 smeserver-support       noarch     1.6.0-50.el4.sme  installed          73 k

So the rpms were installed from CD, the config files renamed, signal-event post-upgrade; signal-event reboot, and all worked again. then the yum update, a boot, and I am as happy as anything.

Thank you all for your contributions in solving my "experiment".

[Stefano]

SERVER IS BACK ONLINE..!!!

well done

Thank you all for your contributions in solving my "experiment".

next time, for your experiment... use a test machine (a virtual one for example) before trying something "unusual"

Ciao

Stefano

[rm]

*I stand corrected*

Well, I did learn more about the internals of SME, that's for sure.

Anyone in for a drink? I'm heading over to the local bar, it's on me!

[Stefano]

Anyone in for a drink? I'm heading over to the local bar, it's on me!

a bit late at night (1.00 am here) and maybe too far from here (treviso, italy)