[clamav-users] ClamAV mirrors have gotten worse!

[Paul Kosinski]

"But the OP has *refused* to consider trying that."

Nonsense!

I did, of course, do that for a while, in a desperate attempt to get any
updates at all. But it seemed like an incredibly crude way to try to
get back to the pre-Cloudflare behavior when updates didn't regularly
fail (and eventually block all updates).

Merely getting updates was not the problem, since that could be "fixed"
by deleting mirrors.dat every so often, but I wanted to get them when
they were advertised (via the DNS TXT) as being available. (On occasion,
a delay in getting virus updates may be dangerous.)

That's when I started to do something more sophisticated (the curl read
of the cvd header), to gather some data as to exactly what was going on.

Even if it's only our installation that has this problem, I'd still like
to fully solve it.


On Fri, 23 Nov 2018 23:00:20 +1100
"Gary R. Schmidt" <grschmidt at acm.org> wrote:

> On 23/11/2018 22:45, Gene Heskett wrote:
> > On Friday 23 November 2018 03:43:40 Dennis Peterson wrote:
> > 
> >> On 11/22/18 8:51 PM, Paul Kosinski wrote:
> >>> I wonder how many users of ClamAV actually log their freshclam
> >>> updates. Those who don't likely won't notice freshclam temporary
> >>> failures due to an out-of-sync condition.
> >>
> >> I just checked logs on two systems dating from July 1 and see no
> >> failures. I isolated the signature serial numbers and time tags and
> >> all were received with clock like precision. Freshclam is launched
> >> every three hours from cron.d and incorporates a randomizer to
> >> create a delay to help avoid pileups on common cardinal clock
> >> positions. No serial numbers were missed within the time slot.
> >>
> >> dp
> > 
> > I did keep a tail on mine, but it Just Works(TM)  That is what its
> > supposed to do I believe. :)
> > 
> As does mine, as it has since I don't know when, and my
> freshclam.conf has logging turned on.
> 
> It *didn't* work during the switch over to Cloudflare, but has been
> fine since that stabilised, but when it wasn't working I had a cron
> job that deleted "mirrors.dat" every hour, which reduced the failure
> rate.
> 
> But the OP has refused to consider trying that.
> 
> 	Cheers,
> 		Gary	B-)