Some years ago I installed FreeBSD on a couple of systems. The version installed at the time were 4.X. Over the years, the systems ticked happily and I've been using the FreeBSD handbook procedures to do source code-based upgrades.
When freebsd-update was introduced I became a bit confused/jealous: I wanted to be able to download only binary packages of applications and also do binary upgrades to new FreeBSD versions. I didn't, I was and still am anxious of something that might go wrong: both systems are heavily used in production and I simply can not face the possibility of downtime due to something going seriously wrong (I am no FreeBSD expert, just a user with some sysadmin knowledge).
At the moment, my systems have got 8.2 installed. My questions are the following, any help will be appreciated:
1) Can I still use the same source code-based procedure (cvsup, make buildworld/buildkernel/installkernel, mergemaster -p, installworld, mergemaster) ?
2) I am using custom kernels on both systems, mainly to reduce kernel size, leaving only needed code like the only net cards installed, removing kernel debugging code etc. I can see from the 9.0 release notes that a couple of kernel options have been introduced, like CAPABILITIES, CAPABILITY_MODE. Are there any other options I should add?
3) Finally, is it possible to switch to a binary-based update from now on? If so, is it a safe or a dangerous procedure? Links?
Please accept my apologies for my numerous questions.
When freebsd-update was introduced I became a bit confused/jealous: I wanted to be able to download only binary packages of applications and also do binary upgrades to new FreeBSD versions. I didn't, I was and still am anxious of something that might go wrong: both systems are heavily used in production and I simply can not face the possibility of downtime due to something going seriously wrong (I am no FreeBSD expert, just a user with some sysadmin knowledge).
At the moment, my systems have got 8.2 installed. My questions are the following, any help will be appreciated:
1) Can I still use the same source code-based procedure (cvsup, make buildworld/buildkernel/installkernel, mergemaster -p, installworld, mergemaster) ?
2) I am using custom kernels on both systems, mainly to reduce kernel size, leaving only needed code like the only net cards installed, removing kernel debugging code etc. I can see from the 9.0 release notes that a couple of kernel options have been introduced, like CAPABILITIES, CAPABILITY_MODE. Are there any other options I should add?
3) Finally, is it possible to switch to a binary-based update from now on? If so, is it a safe or a dangerous procedure? Links?
Please accept my apologies for my numerous questions.