[EMAIL PROTECTED] writes:
> >
> >
> > Anyone here know if redhat linux updates can be rsynced?
> >
> > If so, is it necessary to have rsh installed.
> >
> > I guess what I really need is to see the commands necessary to connect
> > to a redhat `updates' ftp site with rsync. If it is even possible.
> >
>
> I have a script that I run with cron each day to get updates from a
> redhat mirror site. (It only gets the new files) Almost as good as rsync
> and probably less computation intensive for the mirroring server. The email
> goes to an address which I filter with procmail, take the redhat update messages
> and put them into a hypermail archive, so I have a daily web-based report
> regarding if any redhat updates were recently made available.
>
> [root@pour /root]# more /usr/local/sbin/mirrorupdates
> #!/bin/bash
> ncftpget -d /logs/mu ftp.valinux.org /rsync/updates/5.2
>/pub/mirrors/redhat/redhat/updates/5.2/i386/*.rpm
> ncftpget -d /logs/mu ftp.valinux.org /rsync/updates/6.0
>/pub/mirrors/redhat/redhat/updates/6.0/i386/*.rpm
> ncftpget -d /logs/mu ftp.valinux.org /rsync/updates/6.1
>/pub/mirrors/redhat/redhat/updates/6.1/i386/*.rpm
> ncftpget -d /logs/mu ftp.valinux.org /rsync/updates/6.2
>/pub/mirrors/redhat/redhat/updates/6.2/i386/*.rpm
> ncftpget -d /logs/mu ftp.valinux.org /rsync/updates/7.0
>/pub/mirrors/redhat/redhat/updates/7.0/i386/*.rpm
> cat /logs/mu |grep "Cmd\: RETR" > /logs/mu2
> elm [EMAIL PROTECTED] -s "redhat updates" < /logs/mu2
> rm -rf /logs/mu /logs/mu2
[...]
Since your post misses the subject of the thread widely, I'll explain
why I want rsync to do the initial procural. Simply put, It can be
made to delete files on the local dir that are no longer present on
the parent. True enough that other packages can also do this. `Mirror'
and `mirrordir' to mention a few. But I like other aspects of rsync
as well so prefer to use it for this task.
In your scheme:
Over time your ncftpget commands will have downloaded current
packages. Once those (if any) become outdated and are removed your
ncftpget script downloads the new version, repeat as it occurs.
Over time, without outside intervention you will have accumulated a
number of useless packages (some quite large such as glibc or
kernel-source).
Now your network of machines rsync in and pick up the current packages
and then later the newer packages, so that you eventually accumulate
all the useless packages on every machine.
So compound the amount of useless data stored by 6 (roughly) will
shows the huge amount of useless data you are storing.
