Vmin update to 3.40 on suse 10

hi @ all ,

i think, its time for me, to update, to the new virtualmin…
the last update produced a lot of errors and i needed to restart webmin and vmin by hand…
as a result of all the errors, i dont have any virtualmin updater , so i need to do it by hand
a few questions .

  • is it possible, to just install the new version , without to lose current settings and v-servers ? ( made an backup with all settings and all servers already )

  • is it ok, to make a complete suse update ( incl. kernel ) before updating ( reinstalling ) virtualmin ?

  • is there a way ( in the new version ), to set mail quotas again, from virtualmin interface ? e.g. user1.vdom has 100mb space and 50mb attachments per mail ?

ok, i made an os update without any errors … virtualmin 3.39 works fine so far…

tried to make an update via yast -i wbm-virtualmin-server

ups , got some not so nice errors

yast told me, that virtualmin-base needs 2 more files :
glibc xxx.xxx.xxx ( 64 bit ) but i have a 32 bit system

how can i fix that ?

suse 10.0 32 bit on a 64bit cpu

virtualmin 3.39 and webmin 1.341

problem not solved, but eliminated … reinstall box with plesk and run virtualmin only on privat box until it will be stable.

<i>yast told me, that virtualmin-base needs 2 more files :
glibc xxx.xxx.xxx ( 64 bit ) but i have a 32 bit system

how can i fix that ?

suse 10.0 32 bit on a 64bit cpu

virtualmin 3.39 and webmin 1.341[/i>

Scratching my head :frowning:

I avoid SUSE like the Plague so I’m no expert. But obviously SUSE isn’t fully disclosing that it’s dumbed down to 32bit. Then again, is that AMD or Intel 64 bit? I don’t trust the Intel any further than I can spit for being backward compatible.

Don’t put the blame on VM for not being stable…

k, i ll give it another try …

using ubuntu 6.06LTS is impossible, cause the install.sh will abort with the message, that ruby1.8 is nowhere available
… i searched the package via google and supprisingly not one single link works anymore , seems all ruby1.8xxxx packages are gone …

now i ll try it with debian etch as the recomended os …

The 32 bit on x86_64 platform confused the installer, so yast was looking at our 64 bit repository. That won’t work, and it would need to be modified to to point to the 32 bit repository.

Sorry about that…I didn’t think about folks running a 32 bit OS on a 64 bit system. I’ll figure out a way to detect the actual architecture of the OS rather than the architecture of the underlying system.

I do believe Debian Etch is a great OS choice, though. I’ve been very pleased with how easy it is to support.

If you run into problems, let us know, we’ll get them sorted out. Etch support is still quite new, so I’m sure there are still issues. It’ll probably take a few days to sort them out.

I believe Debian and Ubuntu systems will be immune to the architecture problem you saw, even if you’re still running a 32 bit version on a 64 bit platform, as apt-get makes its own decisions about architecture. With yast, we have to explicitly set it–yum also uses a variable for architecture.

Oh, yeah, I’ll be spending time with SUSE this week, once CentOS/RHEL5 go out the door, to bring our SUSE support up to 10.2, and it’ll get some more testing, as well. We just have so many fewer SUSE users than Debian/Ubuntu and CentOS/RHEL that is doesn’t get as much attention. I’m still streamlining our build and test infrastructure to allow me to better support the less popular systems (Mandriva and SUSE) as well as the more popular ones. We’re still not quite there.

using ubuntu 6.06LTS is impossible, cause the install.sh will abort with the message, that ruby1.8 is nowhere available
… i searched the package via google and supprisingly not one single link works anymore , seems all ruby1.8xxxx packages are gone …

Hmm…That’s strange, but it’ll be fixed shortly (I’m not sure why ruby would disappear from the Ubuntu repository–we don’t care what version it is, we just want ruby to be installed…I find it hard to believe ruby would be completely removed from Ubuntu).

Oh, I bet the universe repository wasn’t enabled on your Ubuntu installation. That’s a quirk of Ubuntu that I haven’t yet dealt with in install.sh. So, right now, one has to manually enable the universe repository on Ubuntu.

hi @ all :slight_smile:

thx for the fast answers

i installed debian etch on my box , after some small problems with dovecot ( disallow plain-test in nonssl …) it works more or less fine
without one thing , think something is wrong in my postfix configuration, but i cant figure it out .

i cant send mails from ( e.g. outlook/express) from any of my virtual servers …
i allways get the error msg : 554.7.1. relaying denied
anyone an idea, what is wrong ?

Hey Richard,

Are your Outlook clients configured to authenticate? SMTP authentication is required to send through a Virtualmin server. Username and password is the same as for POP3/IMAP.

hi joe,
yes outlook is configured right … smtp auth is same as po auth.

i relly think, i am missing something in postfix conf … but i cant find it …:frowning:

maybe, u could have a look at my settings , i m sure, u will find the failure really fast :wink:

Hey Richard,

Sure, post the results of this:

grep sasl /etc/postfix/main.cf

That will tell us if something went wrong during the Postfix configuration step (we’ve had so many little bugs in that lately, I wouldn’t be surprised).

It ought to look about like this:

smtpd_sasl_auth_enable = yes
broken_sasl_auth_clients = yes
smtpd_recipient_restrictions = permit_mynetworks permit_sasl_authenticated reject_unauth_destination

If it does, it’ll be something else, and we’ll want to see the /var/log/mail.log during a failed relay attempt.

hmm, looks not so bad …
i have 1 more entry there :
smtp_sasl_security_options = noanonymous

That extra option is fine.

What’s in /var/log/mail.log during a failed relay attempt?

sorry, cant send u the /var/log/mail.log cause of forbidden "names" in it ( erika )

lol, i cant post one single line from the logfile …

For security reasons we only accept the submission of HTML containing the following tags: B I P A LI TT IMG OL UL EM BR TT STRONG BLOCKQUOTE CODE PRE FIRST_NAMES LAST_NAME EMAIL GROUP_NAME DL DT DD FONT. You have a INFO tag in there.

i sent u the logfile by mail to joe@virtualmin.com

Hey Richard,

Just pull out the > and < symbols when posting.

The HTML parser in OACS thinks anything in angle brackets is an HTML tag and it filters down to an accepted subset. This is an anti-spam and anti-cross-site-scripting feature. The new forums package we’re using on the new website approaches it from the other direction (using BBCode instead of HTML to markup things). Neither is a particularly wonderful solution, but both work for most situations.

Hey Richard,

I don’t see any relay denied messages there. I’m not sure what’s happening, but it doesn’t look like the client is even trying to send anything. We ought to be seeing errors related to the actual attempted sends…but we’re not. That’s all just crap related to incoming mail, as far as I can tell.

Maybe Debian writes out is SMTP auth related stuff somewhere else…or maybe the client isn’t actually trying to talk to us at all. I’ll have to dig a bit in the Debian docs.

is there anywhere an option, where i can tell postfix, what virtual servers and users it should relay ?