seth vidal wrote: >>>not really true. >>>yum installs kernels, always. It will never update a kernel. So you'd >>>never be in that state with yum. >> >>is that documented? or I just don't read something again? > > > I'm fairly sure it is mentioned in the docs - I could be wrong - if not > it should be. not in: /usr/share/doc/yum-1.95/README man yum man yum.conf man yum-arch anyway I'm not realy like when I program would like to more clever then me. if I'd like to install the latest kernel and run yum update kernel, then probably I'll reboot the machine nad if I just install than use yum install. although it a question whether you suppose to use yum in batch mode or in an interactive mode (to replace rpm -Uvh, -Fvh, -ivh...) it the first case you may got right, but in the second case IMHO it's better not make any exceptions (why just kernel and what about x and y ...) >>>ie: the modules for old kernels are still there. >>> >>>It'd be unsafe to do otherwise. >> >>why? in daemon mode you've got right, but whem I use it manually...? >>it's the same as if you said rpm -Uvh kernel-2* would be unsafe. > > > when you use it manually why should the behavior change - think about it > like this - if you update the kernel and replace the old one and the new > one breaks your system you're out of luck, but if you just install the > new one, not updating the old one then you can always select the old one > from grub/lilo. but if I has 12 servers and I already try the new kernel on one and I just would like to do the same on all of the other... anyway in worst case I always has a boot floppy but this is again just a matter of tast. are you also prevent from yum remove kernel?:-) or yum remove glibc (I do it once 10 years ago:-() IMHO consistence is more important than prepare for stupid users. >>actualy I don't realy understand it. could you give me an example? > > > Let's take an example: > let's say this fictional config file says this: > samba*:/etc/init.d/smb restart > > which we could take to mean - if something happens to a package matching > 'samba*' then run the command '/etc/init.d/smb restart' > > ok - so then your system updates samba > > and that command is run. > > this is a simple example and clearly more things would need to be > implemented but do you get the idea now? not the bast example since smb restart should have to go to postinstall script but I got the idea:-) >>could you give me an explanation what the update and install exacly >>means in yum? it seems to me some kind of rpm -F -U -i mix... > > > > yum install pkgs: > will install any pkg in the list > will update any pkg in the list if there is an update available > will exit if any pkg in the list is already installed and fully updated > > yum update pkgs: > will update all pkgs in the list > will NOT install a pkg that is uninstalled but in the list > will exit if any pkg in the list is fully updated. > > The above 'will exit' behaviors may change - I've not yet made up my > mind. aha. in this case yum install is equal with rpm -U yum update is equal with rpm -F but in the case of kernel yum install is equal with rpm -i yum install is equal with ? this would be useful to documnet... but as I wrote don't handle kernel differently and curerntly there is no way to keep to different version of package x since both yum install do nothing in case of latest is installed yum update just install... huum I'm getting more confused... -- Levente "Si vis pacem para bellum!"
