Koozali.org: home of the SME Server
Obsolete Releases => SME Server 7.x => Topic started by: portedaix on September 11, 2010, 02:03:22 PM
-
Hello,
I changed my servername to a new one, from the admin console login. Even after reboot, it happens that I still connect to https://oldservername/server-manager. https://newservername/server-manager does not work. So reboot is done as I mentionned. I tried to pass the instrunction "signal-event domain-modify" without success.
vpn is not working anymore, I suspect this servername mismatch to be resposible of it.
Any hint ?
Thanks
-
portedaix
signal-event post-upgrade
signal-event reboot
If that does not fix your issues, please report a bug.
-
Hello,
Thanks for the answer. But I tried it already. Did not work.
-
so please report a bug, thank you
-
Hello,
Thanks for the answer. But I tried it already. Did not work.
Before rise a new bug try change it on server-manager, configuration, workgroup
I think is from there the https/smb name cames from!
-
Hello,
I fixed it by removing vpn rpms. "yum remove pptpd"
Then of course I had "Removing for dependencies: e-smith-pptpd noarch 2.0.0-1.el4.sme installed 28 k"
And I finished by
"signal-event post-upgrade && signal-event reboot"
I reinstalled pptpd after that. I had a problem with it. It was somethnig else. Thunder recently -> wan card dead -> new eth1 nic -> new mac address -> new sme wan wan address because behind my internet box -> vpn redirection not functionning.
I did not see it immediatly. At least, by luck, I fixed this servername thing.
I am note ure at all it a bug. I do not file it.
Salut
-
I fixed it by removing vpn rpms. "yum remove pptpd"
That is very bad practice and will remove other packages unnecessarily. You may also find problems upgrading in future using the above procedure.
Then of course I had "Removing for dependencies: e-smith-pptpd noarch 2.0.0-1.el4.sme installed 28 k"
And I finished by
"signal-event post-upgrade && signal-event reboot"
I reinstalled pptpd after that. I had a problem with it. It was somethnig else. Thunder recently -> wan card dead -> new eth1 nic -> new mac address -> new sme wan wan address because behind my internet box -> vpn redirection not functionning.
I did not see it immediatly. At least, by luck, I fixed this servername thing.
I am note ure at all it a bug. I do not file it.
You have probably removed any trace of the potential bug by doing what you did and have probably left your system in an unknown state, you would have helped us more by reporting (as mentioned earlier) a bug.
-
Before rise a new bug try change it on server-manager, configuration, workgroup
jader, please don't make such a suggestion in the future.. everything that is not working out of the box is a bug and should be posted in bugzilla
I think is from there the https/smb name cames from!
sorry, but you are wrong :-)
-
Stefano
Thank by your tip... I promise do not say "do not open a bug" anymore
Could someone explain what is the source of name as seen by SMB and HTTP: the host name or workgroup name ?
I use to name host after an animal (to name the hardware and be able to find out any docs/drivers about it) and a logical name (server) to workgroup... but I admit to use allways the hostname to map SMB drives... and now I'd like to change that behaviour because this way I can change HW without change anything on client side!
-
That is very bad practice and will remove other packages unnecessarily. You may also find problems upgrading in future using the above procedure.
Only uninstalled "pptpd" and "e-smith-pptpd". No other packages were uninstalled. I reinstalled those two afterwards.
You have probably removed any trace of the potential bug by doing what you did and have probably left your system in an unknown state, you would have helped us more by reporting (as mentioned earlier) a bug.
I cannot afford to wait on my production server for a solution to come. Hence a question I raised previously : is it possible to have a clone sme server, to test anychanges before applying to the production server ?
-
Yes, perhaps Affa might be worth investigating, have a search for it in the wiki. You can also try and use VMWare or any other virtualisation software and experiment with SME Server in there, the latter also facilitates in making snapshots (restore points), making it easier to revert to a previous state while experimenting.
-
Only uninstalled "pptpd" and "e-smith-pptpd". No other packages were uninstalled. I reinstalled those two afterwards.
as Byte said, this is definitely a bad practice
I cannot afford to wait on my production server for a solution to come. Hence a question I raised previously : is it possible to have a clone sme server, to test anychanges before applying to the production server ?
I usually clone my production servers (to make some tests) in this way:
- setup a real or virtual machine.. virtual one is a better choice
- if necessary, update it
- install all the contribs
- create a backup on the original server with a custom procedure:
+ create a file include.txt in /root/ with this content:
home/e-smith/
etc/e-smith/
etc/group
etc/gshadow
etc/passwd
etc/samba/secrets.tdb
etc/samba/smbpasswd
etc/shadow
etc/smbpasswd
etc/ssh
etc/sudoers
+ create a file exclude.txt in /root/ with this content:
home/e-smith/files/ibays/*/files/*
home/e-smith/files/users/*/home/*
+ launch the pre-backup event
signal-event pre-backup
+ create the smeserver.tgz backup with
tar -cz --totals -vvp --directory / --exclude-from=./exclude.txt --files-from=./include.txt -f /tmp/smeserver.tgz
note: the --totals directive is not necessary
- copy the smeserver.tgz file on the "new" server or to a usb disk and follow this (http://wiki.contribs.org/Backup_server_config#Command_line_manual_Restore_from_USB_drive_or_memory_stick) wiki page
in this way I have a clone of the server without user data and files in ibays.. I still have users' emails
NOTE: this is only for testing purpose, use it at your own risk