Brief Outages...

[Just Marc]

Hi
> Hello
>
> On 29/06/2006 16:49, Just Marc wrote:
>> Just out of curiosity guys, why are you using ancient kernels? 2.6 is 
>> stable by now :)
>
> "Guy", if anything - my posts referred to earlier were from this time 
> last year :)
>
Guy!
>> If you insist, upgrade to the latest 2.4 kernel which is probably a 
>> year or two NEWER than what you are running.
>>
>> Time to make that long overdue upgrade,
>
> That's not always possible. At the time I wrote that problem up last 
> year, and the helpful chaps figured out the actual problem with the 
> FTP helper breaking persistence, I had about 10,000 websites in that 
> cluster and couldn't afford downtime. Although saying that I had 
> mitigated it by using keepalived to failover/back between two 
> directors, so I was in a position to reboot the live director at will.
>
Not always possible indeed, but if you can't afford a downtime of ONE 
machine, something is wrong and that something needs to be fixed first.
> Not everyone has that luxury - and a lot of people seem to be using 
> LVS "out of the box" with whatever distribution they currently use (as 
> seems to be the case with the OP here). Several distributions (which 
> shall remain nameless!) keep their STABLE (oh, there I go) branches 
> aeons behind the bleeding edge for the simple reason that more often 
> than not there just isn't any reason to go there.

Debatable. People should know how to upgrade their kernels and should 
keep current when the situation permits -- when their ancident 
rh-as-if-enterprise kernel doesn't have dependencies on ancient 
binary-only provided drivers or what not.

> The adage "if it ain't broke, don't fix it" applies. Although not in 
> this case, clearly :)
Oh well!