Skynot wrote: ↑January 24, 2018 - 10:59 AM
Hello,
I have several questions that I need answered
It's best to ask your question in a discussion thread. It's easier to follow that way. And please specify which WAPT version you're using.
Skynot wrote: ↑Jan 24, 2018 - 10:59
First, if I launch the installation of a Windows patch that automatically triggers a reboot, will the installation resume on its own after the reboot?
If the installation of a package has not completed, the package will be marked as "to upgrade" or "missing" on the machine, and it will be installed at the next shutdown of the machine if WAPTExit is enabled, or before if the admin or the user requests the installation immediately.
Skynot wrote: ↑Jan 24, 2018 - 10:59 AM
Secondly, my infrastructure has a production site and a backup site. Is it possible to cluster the WAPT server? For now, I've only managed to set up a remote repository. But if the server goes down, I have nothing to replace it with.
There are two important data "sources" in WAPT:
- the machine, software and group packages in /var/www
- the postgresql (wapt 1.5) or mongodb (1.3) database which contains the inventory.
In your case, for a disaster recovery plan (DRP), I doubt it's useful to cluster the PostgreSQL database and use a clustered filesystem for the packages. The simplest solution would be to use the virtualization layer for the DRP component, as Simon suggested. If that's not possible, you'll need to regularly replicate the database and the /var/www files to the second site.
From an architecture and disaster recovery point of view, a WAPT server is not the most urgent; it is easy to restore from backups or snapshots, and it is not a critical server in terms of minutes: if the update takes an extra hour to come down because the main server room is burning down, it remains acceptable (unless you need WAPT to push configuration changes and change IP addresses urgently, etc.).
Denis