I run a couple of workstations and a server, and I’d like to set YaST to autoupdate, so that the machines get updated when I’m on holiday etc. It’s not clear to me, however, how YaST handles updates which require a reboot or users to log in again. Obviously I don’t want the server to just go down without warning after a kernel patch, but equally obviously, I do want to reboot it as soon as possible after it needs it.
Does YaST send root an email or write to the console when a reboot is needed, or am I stuck with just monitoring the install logs and guessing which patches require a reboot?
As I never use automatic updates my answer is not the ultimate one.
I assume you mean the update from the update repos (security and recommended).
For most programs this normaly not a problem. Executables in use will not be deleted until the use-count of the files gets to zero. Thus users calling a program will get the new one loaded while others (or the same user) are still using the old one. I think it is possible to create combinations of old and new that do not function, but I also believe that that is not often the case. When your users are tidy people and log out at last at the end of the day, I do not see much of problem here.
A kernel update is different. Conflicts between the old kernel and new modules might occur. Also when there is no reboot, the security vulnarabilaty is still there. I do those updates using the Appllet, but like YaST > Online update it will show a window when a reboot is recommended. Thus the principle is there, but I do not know how this works out when you run this from a CLI and also not when this CLI command is run in the background. Testing is a bit difficult because security updates for the kernel are not that abundant.
I manage a couple of servers, for various purposes, and never auto update. I don’t even update things in the week before my holidays.
In general I suggest you do some thorough reading on openSUSE package management. That would answer all your questions. Don’t expect multiple reboots during one update session, there’ll be a warning if an update requires a reboot, at the end of the update session.
But first, sit next to the machine and see what happens when you update it, that will make it less misterious, give you something to relate to when you read.
>Thus the principle is there, but I do not know how this works out >when you run this from a CLI and also not when this CLI command >is run in the background. Testing is a bit difficult because >security updates for the kernel are not that abundant.
Indeed, that’s why I was wondering whether anybody knew. It’s not something that I can easily test. Of course Knurpht’s point that kernel updates can easily break things perhaps explains why I can’t find a description of what happens. Knurpht, can you point me to the precise documentation which explains how the package management system deals with this? None of my searches have thrown up anything other than “it is possible to autoupdate; here are the steps you need to take in YaST to make it happen.” I haven’t succeeded in finding a description of how the system behaves when it autoupdates.