I highly doubt that this is possible while upgrading from Webmin 1.994 directly to 2.001. This bug pops when installing Webmin 1.999 only.
I would assume that you installed Webmin 1.999 and simply didn’t notice this bug?
Nope. We certainly check for such errors after every webmin update such as the server status page.
During the update to 2.001 which resulted the error of BIND as stated before, our monitor status also report:
Monitor on xxxx.com for 'BIND DNS Server' has detected that the service is uninstalled at 25/Sep/2022 15:40
also it produce error message on virtual server daily backup:
Copying records in DNS domain ..
.. domain not found!
There is no way we did not notice this kind of error if it happens before the update.
So should we at this stage hold off on this update? My config is as follows
| SYSTEM INFORMATION | |
|---|---|
| OS type and version | AlmaLinux Linux 8.6 |
| Webmin version | 1.999 |
| Usermin version | 1.854 |
| Virtualmin version | 7.1 Pro |
| Theme version | 19.99 |
| Package updates | 3 package updates are available |
I currently have no issue with BIND and am running the supposed buggy 1.999?
AlmaLinux was never affected by this bug.
Can I login at your system and have a look?
Nevermind. I think I still see an issue here for users who has never saved Webmin Bind module config.
Anyone with CentOS 7 needs to make sure that /etc/webmin/bind8/config file has the following directives added at the end of the file before installing Webmin 2.001 upgrade:
auto_chroot=
chroot=
We will try to fix that issue for 2.002 release.
1 Like
This also happened to my CentOS 7 after I installed the updates. I immediately noticed that BIND wasn’t running and all dns records were gone for each domain hosted. So I freaked out and restored the previous day back up of my VPS server on Vultr. Now my dashboard appears to be messed up and is missing a bunch of stuff.
Any idea how to fix this dashboard issue now?
You should check on /etc/webmin/virtual-server/config and /etc/webmin/system-status/config files and that they are in place and look correct.
Moreover, using UI you could check Webmin ⇾ Webmin Users: Edit Webmin User / Available Webmin modules / System Status / Module Access Control and see if all system information to show is selected.
Although, I don’t understand how you’re getting this issue if you’re restored from a snapshot. I would suggest to simply revert to the state before restoring from a snapshot and fix a Bind module issue as described on my previous comment.
All the config files are there, don’t know if everything inside them is correct.
Everything is selected under Available Webmin Modules. Unfortunately I can’t go back to the state before I restored from backup because I didn’t save that state. Should it be safe now to install the updates again after I applied your fix?
Yes, sure! That bug is simple to fix as all one should do is to replace those two lines in Webmin Bind config file.
I’m sorry but replace what with what? Based on your post it says to add those 2 lines to the end of the file.
Yes, correct – add in case those two lines are missing or replace existing.
In the other thread, with linked solution, you mentioned if users initially installed 1.999-2 they wouldn’t see it. One would presume if they installed 1.999-1 then 1.999-2 or reinstalled 1.999-2 then it would appear. Perhaps those affected now, simply did the 1.999-2 update initially.
Anyway, glad you’ve figured out the issue regarding the two additional missing directives in the Bind config. Many thanks.
Hi Ilia,
I applied your fix and installed the latest webmin, virtualmin and usermin.
My Dashboard now looks right but some things are still missing.
My biggest concern now is missing mail aliases, as you can see on the screenshot it says 0 mail aliases. I also no longer have mail aliases tab under virtualmin. Please help.
Can confirm, all previous server that was
Operating system CentOS Linux 7.9.2009
Webmin version 1.999
Virtualmin version 7.1
Usermin version 1.854
Authentic theme version 19.99
were missing that chroot and auto_chroot on /etc/webmin/bind8/config
After we add that two lines on the config, and continue to do webmin update on one of our small server.
The previous error was not seen this time, BIND was able to start correctly.
Will be rolling this fix and doing update for the rest of the server after we renew our snapshots.
Thanks.
You need to make sure that Mail for domain feature is enabled on Edit Virtual Server page.
Thanks for the feedback.
This is completely unrelated to the BIND issue that started this thread. Please start a new topic for new problems.
Edit: I have deleted your other posts about the unrelated problem to keep this thread on-topic. Feel free to re-post to a new topic about the problem.
Hi @Ilia,
My server details are the same as the topic creator. I was able to upgrade to 2.001 with no Bind issues thus far. In my config file, for chroot= is blank but for auto_chroot I have the following:
auto_chroot=sh -c '. /etc/sysconfig/named && echo "$ROOTDIR"'
Since I’m experiencing no Bind issues, should I leave this as is or should I set auto_chroot to blank?
Thank you!
Set it to blank.
1 Like