On 03/24/2016 02:07 PM, Chris Murphy wrote:
In my first post on this thread I didn't suggest a Btrfs /home for example. My two recommendations included retaining XFS for /home.
Every analysis I've done of the advantages of BtrFS for myself lead me to think that when its finished, when its done all the deduplication bits and more, it will be of more use for /home than for the RootFS. I do have mainframe experience and do recall the idea of 'rollback' of updates and patches. BTDT. But a few decades of observing corporate use of PC based production systems is that PC are cheaper than mainframes so it was always economically feasible and possible to duplicate a test environment for releases before putting them into production. BTDT with banks, telcos, insurance and more. This framework is established. But end users are another matter. End users make mistakes. End users might delete or overwrite a file and want to get back the previous version. Snappshoting of the user space is a major issue. That's why we have http://snapper.io/manpages/pam_snapper.html I have a similar argument for "shared services", ISPs, that it is better to put user files on a SSD that the core binaries of {/usr/,}/bin and {/usr,}/lib since the shared binaries are going to be loaded once then resident but the real 'churn"' is with the users' application data and they want that loaded fast. Most people seen to disagree and get a glazed look when when I discuss this with them. All to often the responses are of the class "proof by assertion" or "proof by the lemmings principle". -- A: Yes. > Q: Are you sure? >> A: Because it reverses the logical flow of conversation. >>> Q: Why is top posting frowned upon? -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org