I was performing some server maintaince earlier, and I was shocked to find the size of the 32bit repo was riviling that of the mainline repo, containing both 32bit and 64bit packages. I had thought that the mirror would be about half the size, given that there are about half as many packages: tyler@cdn:/var/www$ du -sh 32.arlm.tyzoid.com arlm.tyzoid.com 54G 32.arlm.tyzoid.com 58G arlm.tyzoid.com tyler@cdn:/var/www/32.arlm.tyzoid.com$ find . -name '*.sig' | wc -l 14026 tyler@cdn:/var/www/arlm.tyzoid.com$ find . -name '*.sig' | wc -l 34090 While not a huge deal, I am a bit surprised by this amount of disk space used by packages. Just to confirm: tyler@cdn:/var/www/32.arlm.tyzoid.com/web$ du -sh * 2.7G archisos 51G i686 856K irc-logs Anyone have any ideas what might be going on here? Is this to be expected? or is this an oversight/bug? Thanks, Tyler
Hi, On 08.06.2017 15:19, Tyzoid D wrote:
I was performing some server maintaince earlier, and I was shocked to find the size of the 32bit repo was riviling that of the mainline repo, containing both 32bit and 64bit packages. I had thought that the mirror would be about half the size, given that there are about half as many packages:
tyler@cdn:/var/www$ du -sh 32.arlm.tyzoid.com arlm.tyzoid.com 54G 32.arlm.tyzoid.com 58G arlm.tyzoid.com
tyler@cdn:/var/www/32.arlm.tyzoid.com$ find . -name '*.sig' | wc -l 14026
tyler@cdn:/var/www/arlm.tyzoid.com$ find . -name '*.sig' | wc -l 34090
While not a huge deal, I am a bit surprised by this amount of disk space used by packages.
Just to confirm: tyler@cdn:/var/www/32.arlm.tyzoid.com/web$ du -sh * 2.7G archisos 51G i686 856K irc-logs
Anyone have any ideas what might be going on here? Is this to be expected? or is this an oversight/bug?
well, the issue currently is/was: I'm currently trying to rebuild all packages available - so far ~60% success - these reside in "staging" -> factor of 1.6 2nd: My script was somewhat buggy, not removing older versions from "staging" (should be fixed, now) -> even bigger factor Also note, that "any" packages are not duplicated on a mulit-arch mirror, so your estimate of 50% is too low.
Thanks, Tyler
cheers, Erich
On Thu, Jun 8, 2017 at 11:53 AM, Erich Eckner <arch@eckner.net> wrote:
well, the issue currently is/was:
I'm currently trying to rebuild all packages available - so far ~60% success - these reside in "staging" -> factor of 1.6
2nd: My script was somewhat buggy, not removing older versions from "staging" (should be fixed, now) -> even bigger factor
Sounds good. I had suspected something like that, but couldn't confirm it without further investigation.
Also note, that "any" packages are not duplicated on a mulit-arch mirror, so your estimate of 50% is too low.
Alright, thanks for the clarification.
cheers, Erich
Thanks, Tyler
participants (2)
-
Erich Eckner
-
Tyzoid D