Staying with baselayout-1?

Message

z0ny · Post by **z0ny** » Mon May 09, 2011 9:37 am

Hello, is it possible to simply mask >baselayout-1 to continue with the old system or are there any downsides? When will support for baselayout-1 end? I successfully updated my desktop but I'm a bit scared of updating my servers as they're SSH-only.

Thanks in advance.

Sysa · Post by **Sysa** » Mon May 09, 2011 9:50 am

Of course, you can!

Any way, You'll better do it later when you'll be ready to visit your server in emergency but I do not think that your presence will be necessary - I migrated my 20+ hosts (both, physical and virtual) without any problem yet.

FYI: It is mandatory that You have to read migration manual first!

krinn · Post by **krinn** » Mon May 09, 2011 9:53 am

i suppose as long as baselayout-1 users use it and don't find a bug, devs might just forget and not remove it

but you should just consider baselayout2 stable = 1 dead, and if you have any trouble with baselayout-2 devs will activelly support it and try to correct the error, think about a dev seeing a baselayout-1 user bug ? For dev it mean still searching & correct it when others users have baselayout-2 install and not seeing the bug + fact that nearly all users are now using baselayout-2 and only few still with 1 = expect a "won't fix" answer.

Same for users, more you will wait, less users you will found with baselayout-1 to provide accurate help to you.

Might not be a proof of anything, but i'm using baselayout2 and openrc for so much time that i don't even remember since when i have it and how it was without it, and i don't see any troubles (not totally true, the shutdown is really weak and untrustable under certains conditions, but i suppose it's the same for baselayout1)

You might get troubles when migrating, shouldn't get any when migration is done well.

z0ny · Post by **z0ny** » Mon May 09, 2011 10:10 am

Ok, thanks. I'm going to upgrade the vservers for now but the "real" servers without remote console will continue to run baselayout-1.

z0ny · Post by **z0ny** » Thu May 12, 2011 6:48 am

Updating the vserver was quite a pita. After all the system is running again but some init-scripts have the following output:

Code: Select all

# /etc/init.d/postfix restart
 * Caching service dependencies ...
: command not found.d/net: line 3: 
: command not found.d/net: line 5: 
: command not found.d/net: line 3: 
: command not found.d/net: line 5:                                        [ ok ]
 * Stopping postfix (/etc/postfix) ...                                    [ ok ]
 * Starting postfix (/etc/postfix) ...                                    [ ok ]
#

What's causing this?

Sysa · Post by **Sysa** » Thu May 12, 2011 7:03 am

z0ny wrote:Updating the vserver was quite a pita. After all the system is running again but some init-scripts have the following output:

Code: Select all

# /etc/init.d/postfix restart
 * Caching service dependencies ...
: command not found.d/net: line 3: 
: command not found.d/net: line 5: 
: command not found.d/net: line 3: 
: command not found.d/net: line 5:                                        [ ok ]
 * Stopping postfix (/etc/postfix) ...                                    [ ok ]
 * Starting postfix (/etc/postfix) ...                                    [ ok ]
#

What's causing this?

Did you follow the migration manual? Read it again first: there are some changes in rc-scripts and network configuration.

z0ny · Post by **z0ny** » Thu May 12, 2011 7:51 am

If you mean the guide at http://www.gentoo.org/doc/en/openrc-migration.xml then that's a clear YES. Anything concrete?

Sysa · Post by **Sysa** » Thu May 12, 2011 2:45 pm

z0ny wrote:If you mean the guide at http://www.gentoo.org/doc/en/openrc-migration.xml then that's a clear YES. Anything concrete?

Yes.
e.g. network definitions are changed... runlevels also...
Did you make

Code: Select all

emerge --with-bdeps y -uDN @world
revdep-rebuild

? Did it finish normally?

z0ny · Post by **z0ny** » Fri May 13, 2011 10:46 am

Well, as the guide doesn't mention anything like this I didn't do it yet. But anyways, both commands just finished without errors but the strange messages still remain. Maybe the postfix init script needs updating? As I said, the messages do not appear with every init script...

Hwoarang · Post by **Hwoarang** » Fri May 13, 2011 11:53 am

I don't have problems with postfix. I'd assume that your conf.d/net needs fixing. Can you paste it?

z0ny · Post by **z0ny** » Fri May 13, 2011 1:53 pm

Sure thing, though it looks correct to me:

Code: Select all

# /etc/conf.d/net:
# $Header: /home/cvsroot/gentoo-src/rc-scripts/etc/conf.d/net,v 1.7 2002/11/18 19:39:22 azarah Exp $

# Global config file for net.* rc-scripts

config_eth0="1.2.3.4 netmask 255.255.255.0 broadcast 1.2.3.255"
routes_eth0="default via 1.2.3.1"

By the way, are you on a 64-bit system, too? Maybe only the 64-bit version of postfix has problems.

Hwoarang · Post by **Hwoarang** » Fri May 13, 2011 2:05 pm

The init script is the same for both 32/64 versions. I am not even sure if this is a postfix bug. You can always open a bug and see what happens

z0ny · Post by **z0ny** » Fri May 13, 2011 2:12 pm

Some other packages like portsentry show the same error message, so I guess it has something to do with the new baselayout. Nevertheless the services are starting/stopping as expected!

krinn · Post by **krinn** » Fri May 13, 2011 3:01 pm

if the error comes from the net file, many packages that have network as dependency will show the error

but after all, if it work, you might not care, but as it's not that hard to fix it, and it might fail later... better comply with new standard and remove the errors no?

Hwoarang · Post by **Hwoarang** » Fri May 13, 2011 3:03 pm

z0ny wrote:Updating the vserver was quite a pita. After all the system is running again but some init-scripts have the following output:

Code: Select all

# /etc/init.d/postfix restart
 * Caching service dependencies ...
: command not found.d/net: line 3: 
: command not found.d/net: line 5: 
: command not found.d/net: line 3: 
: command not found.d/net: line 5:                                        [ ok ]
 * Stopping postfix (/etc/postfix) ...                                    [ ok ]
 * Starting postfix (/etc/postfix) ...                                    [ ok ]
#

What's causing this?

What was your original update problem? What you think it was a PITA? Did you reboot after the openrc upgrade?

z0ny · Post by **z0ny** » Fri May 13, 2011 4:00 pm

The old net script also contained an IPv6 address and was booting fine. The new script refused to boot with the IPv6 address and the system was extremely inresponsive. So you can imagine what it was like tracing the error over a slow AJAX-based terminal which sometimes didn't show anything at all and where a reboot took 15 mins plus. Meanwhile I'm not quite sure if the master (not under my control) had problems at the same time.

As soon as I have some more time I will try to trace the error...

@krinn: Ignoring is not an option for me.

Ant P. · Post by **Ant P.** » Fri May 13, 2011 4:04 pm

/etc/conf.d/net and the init symlinks are legacy support only. The new way is to use /etc/conf.d/network.

Hwoarang · Post by **Hwoarang** » Fri May 13, 2011 4:05 pm

That is so NOT true

Post by **floppymaster** » Sun May 15, 2011 12:33 am

Some ml discussion on that topic:

http://archives.gentoo.org/gentoo-dev/m ... 69bd79.xml

Sysa · Post by **Sysa** » Sun May 15, 2011 9:35 am

z0ny wrote:Sure thing, though it looks correct to me:

Code: Select all

# /etc/conf.d/net:
# $Header: /home/cvsroot/gentoo-src/rc-scripts/etc/conf.d/net,v 1.7 2002/11/18 19:39:22 azarah Exp $

# Global config file for net.* rc-scripts

config_eth0="1.2.3.4 netmask 255.255.255.0 broadcast 1.2.3.255"
routes_eth0="default via 1.2.3.1"

...

Your error message said that errors are in lines 3 and 5 - here we see blank lines - maybe some illegal unreadable symbols are there.
Try to open this file with reliable text editor (e.g. vi) and remove these (3,5) blank lines.

z0ny · Post by **z0ny** » Mon May 16, 2011 6:47 am

The problem vanished on its own, I can't reproduce the error any more.

I didn't do anything except for the automated portage sync each night. WTF?
So, case closed, thanks for your help.