| 1 |
Mark Rudholm wrote: |
| 2 |
> The result, however, is that Gentoo becomes an inappropriate |
| 3 |
> choice for a production server deployment. I haven't suggested |
| 4 |
> Gentoo for production servers to anyone (especially my employers) |
| 5 |
> since somewhere around 2003 for this reason. |
| 6 |
|
| 7 |
Strange as that's the year I started pushing Gentoo into production. |
| 8 |
One of the reasons was it allowed us to keep Apache 1.3 when Redhat, |
| 9 |
Suse, etc were forcing 2.0 upon us. We weren't ready at the time, had |
| 10 |
some custom Apache 1.3 modules we needed to update, and frankly 2.0 |
| 11 |
hadn't been around long enough for anyone to trust it. |
| 12 |
However we did transition to Apache 2.0 in late 2004 because Apache 1.3 |
| 13 |
was 10-15% slower, module support was starting to go downhill, and |
| 14 |
eventually you do need to update your systems. Being able to run the |
| 15 |
content servers multi-threaded was a nice gain as well. Overall the |
| 16 |
transition allowed us to grow with 20-25% less hardware than Apache 1.3 |
| 17 |
would have required. In a system of one hundred servers that's an |
| 18 |
elimination of twenty servers, one switch, a ton or so of air |
| 19 |
conditioning, half a rack, 40-60A of electric, one KVM, and so on. |
| 20 |
|
| 21 |
You can't run 1.3 forever and IMHO four years has been more than enough |
| 22 |
time to decide to update a web server even if you have thousands of them. |
| 23 |
|
| 24 |
kashani |
| 25 |
|
| 26 |
|
| 27 |
-- |
| 28 |
gentoo-server@g.o mailing list |
Archives