Koozali.org: home of the SME Server
Contribs.org Forums => Koozali SME Server 10.x => Topic started by: hjnvejlen on March 23, 2021, 09:32:44 PM
-
:-(
I have now tried to install sme-10 RC 1, both with netinstall and with normal install.
And every time it fails. And when it fails, it offers to create a bug report.... to Red Hat :-(
I select keyboard, (Danish) and with netinstall I select network.
And when I try to set region / timezone the installation crashes.
With Beta 1 I have seen the same error once or twice, but I was always able to restart the installation and then it would normally wotk fine.
What is wrong? Or.. am I doing something wrong?
I have my Beta 1 installation upgradet to RC 1 via
yum update --skip-broken
But it puzzles me that the installation don't work.. But if I'm the only one observing this issue please ignore.
Best regards Henning
-
I wouldn't ever do a skip-broken from experience on any centos box.
However I've done over 20 installs of UK without that issue.
Is it the keyboard language or the time zone that causes this issue. As I'm assuming English is understood what heppens then? Just trying to help
James
-
I believe the idea of sme is it just works, I doubt anything wants to be ignored
-
I believe the idea of sme is it just works, I doubt anything wants to be ignored
In this case I had to do it, because of an error in the contribs list, caused by the restore of a 9.2 backup.
Without that skip-broken no updates would install.
And with the skip-broken only one element was skipped, but sems to be fixed by the other updates, since now, after the upgrade, yum update reports no problem :-)
But, you are right in your previous e-mail, I should have tried using english keyboard, even when it's a pain in the.xxx to try to remember/figure out where more special keys are located on en english keyboard.
I will try that, maybe tomorrow on an other installation.Now it's too late for tonight.. in my timezone of course ;-)
But thanks for you input.
br
Henning N
-
for the update of beta 1 to RC1 start with
yum update smeserver-yum --disablerepo=* --enablerepo=smeos,smeupdates
a last resort if this one is not working is to move you yum_repositories db and recreate one by doing initialize-default-databases:
mv /home/e-smith/db/yum repositories /home/e-smith/db/yum repositories.ori
/etc/e-smith/events/actions/initialize-default-databases
For the installer, usually this issue arise a lot on vm when the host is quite loaded.
the keyboard is not an issue, but rushing in setting all you want could be. When you lend on a page, just wait everything settle down.
-
Jpp I assume skip-broken should be avoided at all costs?
-
Jpp I assume skip-broken should be avoided at all costs?
indeed
the issue you have is probably centos 6 repo remaining from restore. Doing the yum update of smeserver-yum shall fix that.
you do not want to install any centos6 / el6 packages.
-
For the installer, usually this issue arise a lot on vm when the host is quite loaded.
the keyboard is not an issue, but rushing in setting all you want could be. When you lend on a page, just wait everything settle down.
Absolutely, a not uncommon event on the likes of Proxmox where you have 1 or more VMs already running then run the installer for sme10, patience is more than a virtue, particularly in the summary form. slower and less is better :-)
-
for the update of beta 1 to RC1 start with
yum update smeserver-yum --disablerepo=* --enablerepo=smeos,smeupdates
a last resort if this one is not working is to move you yum_repositories db and recreate one by doing initialize-default-databases:
mv /home/e-smith/db/yum repositories /home/e-smith/db/yum repositories.ori
/etc/e-smith/events/actions/initialize-default-databases
For the installer, usually this issue arise a lot on vm when the host is quite loaded.
the keyboard is not an issue, but rushing in setting all you want could be. When you lend on a page, just wait everything settle down.
Actually this was NOT on a VM, but with a different HDD than the one I normally use for my SME-server..which now IS upgraded to RC 1..
So this is why I ask here, in case others have seen it and it may be a real problem.. But as I said. I will try tomorrow on a different PC .. and perhaps with an empty hdd.
-
indeed
the issue you have is probably centos 6 repo remaining from restore. Doing the yum update of smeserver-yum shall fix that.
you do not want to install any centos6 / el6 packages.
No no no not at all, it was just, as you described, something left over from the restore of a backup from 9.2..
As far as I remember, you found the bug yourself a few days ago, when I reported the issue.. (sorry if my spelling is bad, but I'm close to my bedtime and tomorrow is another working day ;-)
br
Henning
-
Have seen the odd issue when using used drives, sme9 was same in early days, if I recall wiki has a note about nuking the drive with a dd before using it..
So region/timezone selection, which one .. will try a test VM...The Great Siouthern Land / Melbourne has always been good :-)
and sleep well.. :-)
-
It's never an issue here they just want to make sure it's working well.
-
No no no not at all, it was just, as you described, something left over from the restore of a backup from 9.2..
As far as I remember, you found the bug yourself a few days ago, when I reported the issue.. (sorry if my spelling is bad, but I'm close to my bedtime and tomorrow is another working day ;-)
br
Henning
that is exactly what i said. when migrating an eol sme9 db got hard encoded the centos 6 vault repo.
the last smeserver-yum fix that.
for the rc1 installer this is the exact same as beta1. no upgrade of version. i just copied beta1 one to rc1. so likely hardware, or metaphysics ;)
-
:-(
I have now tried to install sme-10 RC 1, both with netinstall and with normal install.
And every time it fails. And when it fails, it offers to create a bug report.... to Red Hat :-(
I select keyboard, (Danish) and with netinstall I select network.
And when I try to set region / timezone the installation crashes.
With Beta 1 I have seen the same error once or twice, but I was always able to restart the installation and then it would normally wotk fine.
What is wrong? Or.. am I doing something wrong?
I have my Beta 1 installation upgradet to RC 1 via
yum update --skip-broken
But it puzzles me that the installation don't work.. But if I'm the only one observing this issue please ignore.
Best regards Henning
I've seen similar.
Created a USB install from the ISO, installing on bare metal, 2 x HDD
First attempt died after selecting "I will configure" on install destination, an aparrant time out error.
Second attempt died on Selecting English UK, an aparrant time out error.
Third attempt died when reclaiming disc space.
Fourth attempt, I avoided the "Install To" option and the install completed.
But... no sign of the RAID configuration from the console...
-
2 x HDs and just accepting the summary defaults should deliver a Raid 1 install, only thing on summary page is to confirm Local/timezone, and keyboard and I also always turn on the network setting, thats it..
Only proviso I can think of is HDs need to be same size, I think there is some logic setip in the installer the raid Guru, Chris, may read this, at worst he is in Rocket, second used HDs have often been an issue in the past, better to have dd them prior. Having said that I have now performed multiple installs to used SSds that had had sme9 and sme10 alphas and betas on..
-
also as soon as you enter in the menu for disk setup and confirm the « default automatic partitioning » then you just erase the one we put and accepted the default RH/CentOS = one boot and one huge LVM with all the disks merged.
-
Yes, should have added that, let the custom install setup do its thing, Raid 1 - 6 depending number of disks and sizes, if you go into the partitioning it resets all that and you are back to the standard setup, NO raid, unless you manually set it all up, having played with it no simple thing, and why? all the good work by Chris is wasted :-) of course if what you want is as JP wrote above, that is how.
-
So is the bare metal install as was, 2 disks raid 1 4 disks raid 5 etc Im glad if so plus vote here
-
Chris updated wiki https://wiki.contribs.org/Raid
-
Terry is spot on as usual. By default we'll configure the appropriate RAID setup based on the number of drives you have within 100MB of the first detected disk. If you go into partitioning you should be able to see the custom setup which has been defined, but there shouldn't be any need to change anything.
If you're having issues with the graphical installer, try the text installer. It's not super friendly but you only have to set time zone and then hit 'b' to begin.
Once installed, if you have any issues with RAID, share the contents of /root/sme-partitioning.log to help us debug.
-
:-(
I have now tried to install sme-10 RC 1, both with netinstall and with normal install
Hello, I feel a bug here : https://bugs.koozali.org/show_bug.cgi?id=11547
Issue explain in the report, wait on dev analyze but follow my comment if you want to acheive the installation