mcafee autoupdate fails for ver 5100

[Glenn Steen]

On 26/09/06, Phil Udel <Phil.Udel at salemcorp.com> wrote:
> About 5 years ago I wrote a script that tries multiple mcafee ftp servers
> until it can find a good file to download.  It works about 99% of the time.
> One of the nice options is it will email you when the update happens.
> Normally I ran it from Cron but recently I retrofitted it for MS.  I could
> make it available if there is any interest.
>

I too have typed up my own mirroring script (well, it used to use wget
for the actual mirroring, now use lftp...) to mirror their
CommonUpdater thing for our EPO (secondary repository type of thing).
I've learnt quite a bit about their infrastructure by this
excersise:-).
For one thing... Their main ftp server address seem to be a
round-robin or loadbalancer type of thing, so you needn't check
several ftp hosts, since they will send you to different ones (perhaps
they've outsorced that to akamai too, as with the HTTP related ones).
The update time for them to mirror everything out is very often
extremely long. This, coupled with the windoze ftp servers and the
fact that they change case on the relevant files (like Replica.log,
Current etc) after all files have been uploaded, make mirroring their
stuff pretty interresting:-).

The reason to use CommonUpdater (ftp://ftp.mcafee.com/CommonUpdater)
is because then I can use the information in Replica.log
(one/directory, as opposed to update.ini which is more limited in
scope) to ascertain that the mirror action is complete and successful
(another script that check sizes, SHA1 sums etc) before moving it into
place and mailing me etc.

But for the purpose we use it in MailScanner, I think that method
would be overkill, seeing as we only need one file, and them supplying
http access as they do.
What would be interresting is perhaps to use the information in
update.ini to check the supplied md5 sum for the tar-ball.

-- 
-- Glenn
email: glenn < dot > steen < at > gmail < dot > com
work: glenn < dot > steen < at > ap1 < dot > se