| 1 |
Once, maybe twice a day to the admin responsible for that mirror seems |
| 2 |
fine to me. It doesn't look like you're hurting for rsync mirrors, |
| 3 |
personally I'd just toast any that have no admin contact information. Of |
| 4 |
the 13 with no email specified, 8 of them either have no timestamp or no |
| 5 |
response, so you'd lose 5 total. Then again, I like to cover my own ass. |
| 6 |
;) |
| 7 |
You could also script it up so that it instead of keeping track of the |
| 8 |
down notices including flapping, etc, it just accumulates a certain |
| 9 |
number, say after 5 attempts (depending on how often it's run) and no |
| 10 |
result, it fires off an email. |
| 11 |
|
| 12 |
burn |
| 13 |
|
| 14 |
Jeffrey Forman wrote: |
| 15 |
> On Tue, 2006-02-14 at 23:45 +0100, Michael Renner wrote: |
| 16 |
> |
| 17 |
>> Send it to the mirror contact. The mailing list should be kept |
| 18 |
>> "noise-free". I (and most likely other mirror-admins too) don't care |
| 19 |
>> when mirror $foo is broken. |
| 20 |
>> |
| 21 |
> |
| 22 |
> That's one side of the response that I can totally relate to. We'll see |
| 23 |
> how the rest of the admins think about this, and I'll go with majority |
| 24 |
> rules. |
| 25 |
> |
| 26 |
> /me scribbles down Michael as a "dont bother me unless I screw up" ;) |
| 27 |
> |
| 28 |
> |
| 29 |
>> Send it when the mirror goes "bad" and from that one once per day. As |
| 30 |
>> soon as the server gets "ok" again, reset the one-day-delay, so that |
| 31 |
>> "flapping" problems can get spotted. |
| 32 |
>> |
| 33 |
> |
| 34 |
> The flapping is something I thought about, but the hamsters in my head |
| 35 |
> kept burning out when I tried to think of a simple solution to keeping |
| 36 |
> track of it. Then again I hope with this script, I will be much more |
| 37 |
> on-top of removing mirrors that have gone by the way side. |
| 38 |
> |
| 39 |
> -Jeffrey |
| 40 |
> |
| 41 |
> |
| 42 |
> |
| 43 |
-- |
| 44 |
gentoo-mirrors@g.o mailing list |
Archives