Hi Siju , Great explanation man ! regards Praseed Pai
--- On Fri, 9/3/10, Siju George <[email protected]> wrote: > From: Siju George <[email protected]> > Subject: Re: [ILUG-Cochin.org] Linux machines with no rebooting…? Is this > what we want? > To: "This List discusses GNU/Linux & GNU, GPL Software" > <[email protected]> > Date: Friday, September 3, 2010, 1:20 AM > On Fri, Sep 3, 2010 at 9:43 AM, > Ganesan Venkata Subramanian > <[email protected]> > wrote: > > HI ALL > > > > IS THIS WHAT WE WANT ? > > > > http://www.standalone-sysadmin.com/blog/2010/09/linux-machines-with-no-rebooting-is-this-what-we-want/ > > > > :-) > > Just entering the world of Mainframes for a few seconds > > "Four nines availability" and "Five nines availability" are > terms > unfortunately unfamiliar to PC users. Yet it is these > figures - 99.99% > and 99.999% availability - that are used to rate the > reliability of > mainframes. Such figures equate to between 5 and 53 minutes > of > downtime a year. In fact, for System/390 mainframes - the > average time > between failures that force a reboot and an initial program > load - is > 20 to 30 years. Such reliability is truly stunning from > the > perspective of a PC user, yet this kind of performance is > crucial to > businesses where a crash could incur losses of millions of > dollars for > every hour of downtime. " > > Comming back to our world of FOSS OSes;-) > > Today, as of now, when somebody brags about uptime of a > server ( not a > service ) he betrays an insecure server most of the time. > > Most OSes have security updates to the kernel which will > force a reboot. > So uptime is an indicator of how secure the OS on the > Server is. > > Secondly ( even when there is not kernel update ) from an > OS > developer's perspective. > > " a libc fix would require rebuilding libc, plus relinking > statically linked programs. Followed by a restart of all > programs. > That includes init(8), so you're probably better of just > rebooting > after all." > > http://marc.info/?l=openbsd-misc&m=125035083804327&w=2 > > Some OSes don't support kernel modules in the name of > security as you > saw in the above thread. > > The arguments in that URL posted are also valid concerns. > > But again why I am not keen to have a > *dontneedtoreboot-server* is > because we have technologies that can be used to if we > really want a > 99.x uptime. Clustering, Loadbalancing, Replication, etc. > There are > challenges ofcourse :-) > > Even Firewall services can be run without down time ( or > losing states > ) when one of the firewalls in the cluster goes down for > upgrade or > reboot or hardware maintenence. > > http://www.openbsd.org/lyrics.html#35 > http://en.wikipedia.org/wiki/Common_Address_Redundancy_Protocol > http://www.openbsd.org/faq/pf/example1.html > > --Siju > > _______________________________________________ > Indian Libre User Group Cochin Mailing List > http://www.ilug-cochin.org/mailing-list/ > http://mail.ilug-cochin.org/mailman/listinfo/mailinglist_ilug-cochin.org > #[email protected] > _______________________________________________ Indian Libre User Group Cochin Mailing List http://www.ilug-cochin.org/mailing-list/ http://mail.ilug-cochin.org/mailman/listinfo/mailinglist_ilug-cochin.org #[email protected]
