The ISO is ready! You can download it from .torrent (490MB) before the official announce (/cc @alefattorini @quality_team ).
2 Likes
Good news! I need to work on it, give me a few days 
1 Like
stuck at 99.8%…
The seed machine seems ok 
Do you have any hint? What could I check?
Do you want to try stopping+starting again the download?
You can also check the sourceforge link: https://sourceforge.net/projects/nethserver/files/nethserver-6.8-beta1-x86_64.iso/download
2 Likes
A lot of updates/upgrades (changelog stats. 153 pkgs, 98 source pkgs, 98 changelogs) in Software center (NS 6.7 final), only 2 for NethServer.
Between all these, there is also the upgrade to CentOS 6.8 from RedHat.
Because still we cannot select which update/upgrades to be made, it’s ok to make all of them? There are known issues for NS 6.7 (final) to work on CentOS 6.8?
I saw one issue ( Hylafax configuration bug ).
My NS 6.7 (final) is in production as email server and I don’t want to upgrade to NS 6.8 beta1 yet, but I want to keep it up-to-date.
TIA,
Gabriel
1 Like
We updated yesterday (all packages) and after two days of full use we had no major problems. (Just a small issue with Hylafax configs).
Samba seems to be fixed as we had no login issues.
Regards
Massimo
2 Likes
Thank you Massimo!
I will proceed to updates!
Gabriel
Hello Massimo,
Thank you!
Me too!
After over 20 years with Windows it’s hard not to do this!
BR,
Gabriel
EDIT:
No issues till now.
After upgrade I have installed Fail2Ban by @stephdl and works fine.
1 Like
While I rarely do it, I’d like to reboot after a full system upgrade.
The most important reason is to check that the system boots regularly, avoiding the risk of discovering issues in an emergency after a crash.
Honestly, I don’t recall experiencing such a problem in many years, that’s why I usually avoid the post upgrade reboot.
Moreover, I usually examine the kernel changelog and reboot if the new kernel fixes something really important.
Having a predefined window for maintenance helps a bit.
What about updates like the latest glibc ( I think that was it) that demand a reboot? I’ve never been clear on why a process restart isn’t sufficient.
Updating glibc requires restarting pretty everything, so a reboot is simpler.
1 Like
Just for Info: Updated everything on Wednesday including kernel update and reboot.
No Problems. Everything works fine.
Have a nice weekend all together. 
2 Likes
What, no shiny new badge?
Hi @fasttech,
First, as the title of this post said, it’s about CentOS 6.8 not about NethServer 6.8 beta1.
Unfortunately from Software center cannot be made upgrade to the newer version of NS, but only for CentOS. Also from here can be made updates for current version of NS.
(There are some discutions about upgrade from Software center on this forum. One of them:
Which way do you prefer for the upgrade to 6.7? )
To upgrade NS 6.7 final to NS 6.8beta1, you must follow the procedure as @apradoc said, from here:
NethServer 6.8 beta1 Released!
or from here:
http://docs.nethserver.org/en/latest/release_notes.html#manual-upgrade-from-6-7
Generally, if you want to upgrade NS to a newer version, you must follow the procedure described in the announcements for a new version, or from
http://docs.nethserver.org/en/latest/index.html
After the upgrade to the newer version of NS, the updates for that version will be available from Software center.
Gabriel
5 Likes
Ahhh… @GG_jr I saw this;
Jun 03 12:53:43 Updated: centos-release-6-8.el6.centos.12.3.x86_64
…amongst the huge update and mis-assumed, thanks to your post I’m clear now on what’s what now, thank you.
3 Likes
You’re welcome!
I’m glad when I can help!
1 Like
Some minor things about NethServer 6.8 beta1 (and 6.7) that might be improved. Excuse me if some of them have been reported already, or if there is any wrong assumption from my side.
Note this was just for testing and breaking things on a VM while gathering information for wiki userguides.
NethServer 6.7
Configuration → Windows Network
-
takes 16 first characters from domain name (shouldn’t be 15, as the 16th is reserved?)
-
Workgroup, Domain and Realm input fields do not validate input data (case-type, length, allowed characters…) from the client-side.
- server-side validator does limit it to 63.
-
Not able to guess the meaning of some error messages:
NethServer 6.8 beta1
( ^ some of the errors also happen on 6.7)
First Config Wizard ^
- Hostname and Domain input fields do not pre-validate input data (allowed characters, length…) from the client-side. This also happens with other Nethgui’s input fields.
- Validation is done only from the server-side (which is what really matters after all).
- Validator limits them to 63 characters. OK for hostname (system name) and domain labels, but a DNS domain name can be longer, and can have a TLD suffix.
- When an error occurs and a red warning banner is shown, the title for the next section is highlighted on the left-column.
- Also, some error messages are not very helpful (i.e. hostname > 63)
- No support for punycode/IDN domain names (reported already)
Ports assignment ^
- SSH port: can be assigned in the range of 1-65535, and ports in use are not excluded. Couldn’t this also conflict with well-known ports and with Dynamic ports?
Jun 11 18:25:30 localhost sshd[2108]: error: Bind to port 980 on 0.0.0.0 failed: Address already in use.
Jun 11 18:25:30 localhost sshd[2108]: fatal: Cannot bind any address.
Administration → Software center ^
- After installing a module the menu entries are not shown on the sidebar until the page is refreshed manually or any other menu is accessed.
- After removing a module the related menu entries are still shown on the sidebar until the page is refreshed manually or any other menu is accessed.
Configuration → Server name
- exit status 256 error when setting a long hostname + domain name. Then it is not shown correctly in the Nethgui header (root’s profile menu), instead php
gethostname()gets the old name (or none after a reboot), but it’s properly set on/etc/hostsand/etc/sysconfig/network
from /var/log/messages:
Jun 12 17:25:49 123456781012345678201234567830 /sbin/e-smith/db[7041]: /var/lib/nethserver/db/>configuration: OLD DomainName=a23000001012345678201234567830.co
Jun 12 17:25:49 123456781012345678201234567830 /sbin/e-smith/db[7041]: /var/lib/nethserver/db/configuration: NEW DomainName=a23000001012345678201234567830.com
Jun 12 17:25:50 123456781012345678201234567830 esmith::event[7046]: Event: hostname-modify
Jun 12 17:25:50 123456781012345678201234567830 esmith::event[7046]: hostname: name too long
Jun 12 17:25:50 123456781012345678201234567830 esmith::event[7046]: Action: /etc/e-smith/events/hostname-modify/S02interface-config-hostname FAILED: 1 [0.05714]
As the log reports, the name is too long. I think this happens when hostname.domain combined surpasses the 64 characters enforced by the HOST_NAME_MAX linux variable:
# getconf HOST_NAME_MAX
64
when the hostname command is called from:
# tail -n 1 /etc/e-smith/events/hostname-modify/S02interface-config-hostname
/bin/hostname ${system_name:-localhost}.${domain_name:-localdomain}
Dashboard ^
- Boxes: long text can overflow its container.
4 Likes