apt-smart
apt-smart copied to clipboard
martin68
maybe a smarter choice needed?
For me the output of the command apt-smart --list-mirrors is:
------------------------------------------------------------------------------------------------------
| Rank | Mirror URL | Available? | Updating? | Last updated | Bandwidth |
------------------------------------------------------------------------------------------------------
| 1 | http://archive.ubuntu.com/ubuntu | Yes | No | Up to date | 335.33 KB/s |
| 2 | http://ubuntu.turhost.com/ubuntu | Yes | No | 1 hour behind | 708.1 KB/s |
| 3 | http://ubuntu.saglayici.com/ubuntu | Yes | No | 3 hours behind | 430.89 KB/s |
| 4 | http://ubuntu.vargonen.com/ubuntu | Yes | No | 3 hours behind | 122.09 KB/s |
| 5 | http://mirror.muvhost.com/ubuntu | Yes | No | 4 hours behind | 666.03 KB/s |
| 6 | http://mirror.ni.net.tr/ubuntu | Yes | No | 12 hours behind | 588.75 KB/s |
| 7 | https://mirror.sh.com.tr/ubuntu | Yes | No | 1 week behind | 801.89 KB/s |
| 8 | http://ftp.linux.org.tr/ubuntu | Yes | Yes | 4 hours behind | 422.29 KB/s |
------------------------------------------------------------------------------------------------------
And the output of the command apt-smart --find-best-mirror is http://archive.ubuntu.com.
It seems to me that apt-smart is not taking into account the bandwidth (server speed) enough.
I fully agree that, in general, server speed is less important than the time since it was updated.
For instance, the server ranking 7 in the list could even be at the very bottom.
However, I think being merely “1 hour behind” should be considered less important than having less than half the speed.
Therefore, I think the 2nd server in the list should actually be selected as the best mirror.
It seems to me that apt-smart is not taking into account the bandwidth (server speed) enough.
I completely agree with this. Having a mirror with faster bandwidth is the main reason of why I decided to start using apt-smart.
Unfortunately, I can't use this otherwise very useful utility until this issue is fixed, and I hope evcerything is OK with Martin68 of Hangzhou, China, and he or or somebody else (maybe @xolox ?) can take care of this issue.