I currently use the following in our sql logging plugin to generate a unique
id in the database:
my @sname = split(/\./, $self->qp->config("me"));
my $sqlIdent = $sname[0].$$.'r'.int( (( time ^ $$ ) * rand($$)) /
rand(time/$$));
With each mx box running at ~30 cps we get dups often enough I'm considering
changing it.
On 8/23/07 1:34 PM, "David Sparks" <[EMAIL PROTECTED]> wrote:
> JT Moree wrote:
>> Hanno Hecker wrote:
>>>> Is there a message ID that is unique to each message?
>>> Not until the sending client has submitted a 'Message-ID' header ;-)
>>
>>> But it would be easy to add / generate a transaction id after every
>>> reset_transaction() call. This could be logged instead of (or as
>>> addition to) the PID.
>>
>>> + $self->{_transaction_id} = sprintf("%08X", rand(2**32 - 1));
>
> I calculate transaction ids like this (hires time):
>
> my $mid = sprintf("%.4f%d", time(), $self->{_id});
>
> $self->{_id} is the id of the server and currently calculated by using
> the last digit of the IP address. This allows transaction ids to be
> unique across 10 servers. When servers get fast enough this may end up
> causing dup ids. A scheme like this allows easy auditing of problems
> because the id contains the "when and where".
>
> Cheers,
>
> ds
>
>
>
>>
>> Is this uique enough? what is the chance of getting the same random
>> number again? should it be a combination of the PID + time + rand?
>
--
Ed McLain
Sr. Data Center Engineer
TekLinks, Inc.
205.314.6634
[EMAIL PROTECTED]
