On Wed, Mar 19, 2008 at 12:40:37PM +0100, Petter Reinholdtsen wrote:
> +# Select a random day to submit on, to spread the load over time, unless it
> is already set.
> +select_random_day() {
> + RANDOM=$(dd if=/dev/urandom count=1 2> /dev/null | cksum | cut
> -c"1-5")
> +
I cannot find any reference to it now, but somebody suggested the
following solution to spread the load more evenly:
1. Move the cron job to cron.hourly.
2. Let the cron job keep track of when statistics were last reported.
3. If more than a week has elapsed since the last time statistics were
repo
[Bill Allombert]
> We move the popcon cronjob to cron.dayly, and pick a random weekday
> for each submitter. The dayly cronjob check if it is running on this
> weekday before reporting, else it abort immediatly.
Here is a draft patch to implement this. It does not address the
renaming/moving of t
[Bill Allombert]
> I asked for feedback from Holger about the DOS issue and got nothing.
I assume you are talking about the "Is it possible to get more data on
this issue, like the relevant apache logs ?" question in bug #459910.
As far as I know, Holger is not a Debian sysadmin, so I doubt he got
On Wed, Jan 23, 2008 at 04:44:59PM +0100, Bill Allombert wrote:
> It won't be removed from testing. Instead the new version will not
> propagate to testing and testing will stay with 1.42 and sid will have
> 1.43 so it will be easy to discriminate systems running testing from
> system running sid,
On Wed, Jan 23, 2008 at 04:23:22PM +0100, Enrico Zini wrote:
> On Wed, Jan 23, 2008 at 03:49:39PM +0100, Bill Allombert wrote:
>
> > I would like to keep popularity-contest out of testing for
> > 1) See how many people use sid versus testing
> > 2) give us a chance to address the DOS issue in a b
6 matches
Mail list logo