On Tue, 23 Feb 2016 15:35:36 +0100, Steve Lord <[email protected]>
wrote:
On 23 Feb 2016, at 16:29, Imre Kaloz <[email protected]> wrote:
On Tue, 23 Feb 2016 15:24:12 +0100, Steve Lord <[email protected]>
wrote:
And how do you imagine copying the contents when the IO is already
killing the box from time to time?
You don't store the mirror on the box. You store it elsewhere, and you
do it smartly. There's a difference between mirroring the site, wiki
and image/package downloads. The PHP project might be a good place to
look for inspiration. They have a mirroring programme.
As for I/o killing the box, maybe it would help to know the conditions
under which that happens and what solutions have already been tried.
The box is over 7 years old, the raid controller is the first part that
started to give up. I'm the only one who put money in the
infrastructure in the last 10 years - the solution is the replacement
of the box, which I've also taken care of, "just" need to move all data
over which isn't that easy given the random lockups and the public
services.
I'm sorry to hear that you've put so much money and time into it and are
still experiencing problems. Would it help if there was a publicly
announced window to temporarily shut down the site and move data off the
box? It might be feasible to do it bit by bit and move things off with
less disruption over time.
Absolutely, but I'll be back to home (and near the box) only on Saturday.
So I keep that option as a last resort if I can't move things there by the
time.
Imre
_______________________________________________
openwrt-users mailing list
[email protected]
https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-users
