Re: Changing nsswitch.conf on a running system (was Re: /etc/nssswitch.conf is supposed to be a symlink now?)

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On 11/29/18 8:29 AM, Kevin Kofler wrote:
Ray Strode wrote:
(defer until next offline update?)
That would be never on many systems. Most users using command-line DNF and
all users using plasma-pk-updates or dnfdragora never do offline updates by
design.

I, being of the old school, still try to stick to online updates, but I came to believe that we just don't have the infinitely rolling update capability any more, so your online updates have to be punctuated by reboots anyway. This is true on several levels:

- kernel updates---the livepatch / ksplice capabilites are not mainstream enough

- basic session infrastructure does not support online updates, e.g. recent dbus discussion where people explicitly said that dbus cannot be restarted cleanly

- application-level issues (IPC protocol or file formats, etc)

Even though my updating method does not force it, I restart my system whenever I see that I am running a kernel that's more than one-two behind the latest update, or when I see instability in some userland processes I use.

I have given up trying for uptimes measured in months--in practice, my uptimes range between weeks and months. I don't even think infinite uptime is a reasonable goal---I'd rather work on setting my system up so that all long-term tasks restart automatically and correctly pick up where they left off --- I monitor/collect several extended datasets, such as weather/soil/environmental conditions, home automation, etc. Having said that, I definitely want control over offline/online updating---the choice between 'interruptions all the time' and 'interruptions only when it is most annoying' is not acceptable :)

I do think there should be a clear, established way to determine when the reboot is needed. There was "needs-restarting" from yum-utils, but it effectively disappeared because yum-utils conflict with dnf now, and it was a little flaky anyway.

There's a lot of gray area between the dnfdragora suggestion of rebooting for every update, and rebooting only for Fedora N-2->N version upgrades at the N-2 EOL. I don't know if it's practical or desirable to automate this, but maybe there should be a package boolean marking each upgrade as 'requiring reboot' or not; kernel and dbus upgrades would have it always as 'YES', and other packages would reflect the judgment of packagers.
_______________________________________________
devel mailing list -- devel@xxxxxxxxxxxxxxxxxxxxxxx
To unsubscribe send an email to devel-leave@xxxxxxxxxxxxxxxxxxxxxxx
Fedora Code of Conduct: https://getfedora.org/code-of-conduct.html
List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
List Archives: https://lists.fedoraproject.org/archives/list/devel@xxxxxxxxxxxxxxxxxxxxxxx




[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Index of Archives]     [Fedora Announce]     [Fedora Users]     [Fedora Kernel]     [Fedora Testing]     [Fedora Formulas]     [Fedora PHP Devel]     [Kernel Development]     [Fedora Legacy]     [Fedora Maintainers]     [Fedora Desktop]     [PAM]     [Red Hat Development]     [Gimp]     [Yosemite News]

  Powered by Linux