On Wed, Jun 7, 2017 at 5:29 PM, Japheth Cleaver <cleaver at terabithia.org> wrote:

On 6/7/2017 1:29 PM, Asif Iqbal wrote:

On Wed, Jun 7, 2017 at 4:26 PM, Ryan Novosielski <novosirj at rutgers.edu> wrote:

...

I stop the MTA sometimes when I know this is about to happen.

In our case, we do not see a pattern when the firewall is crashing.

...

---

From: Xymon <xymon-bounces at xymon.com> on behalf of Asif Iqbal < vadud3 at gmail.com> Sent: Wednesday, June 7, 2017 4:14:43 PM To: xymon at xymon.com Subject: [Xymon] Avoid email storm

xymon is behind a firewall and lately firewall is dying a lot and at restore we are getting email storm about devices that are on other sites.

firewall team will have a replacement "soon".

In the meantime I guess I should add a depend (first hop) for all 2000+ devices in hosts.cfg file.

Is there another option with less editing?

Thanks for any suggestion!

Flap detection and delaying the alert might help, but delayred= (hosts.cfg(5)) might be a better option for handling this, depending on the number of network tests you have.

If the entire server is behind the FW or you have some other programmatic way of determining what will be on the other side of this FW, I'd suggest adding (either via route or depends) data in regardless (you might be able to get creative with sed or perl to make this addition easier). I've found the more flexibility you can give yourself in describing what you're testing, the more options you'll have for coping with situations like this.

HTH, -jc

yes the entire server is behind FW. When the FW fails, conn on remotehost definitely fails along with conn of thousands of other hosts.

So for all hosts in hosts.cfg I should append a depends like below ?

1.2.3.4 foo # conn bar qaz depends=(conn:remotehost/conn)

-- Asif Iqbal PGP Key: 0xE62693C5 KeyServer: pgp.mit.edu A: Because it messes up the order in which people normally read text. Q: Why is top-posting such a bad thing?