Forum - Policy for aiccu reconnects? :: SixXS - IPv6 Deployment & Tunnel Broker

Policy for aiccu reconnects?

Shadow Hawkins on Thursday, 13 September 2012 10:59:23

One of my tunnels died when the cable-ISP changed IPv4 on me (again). I'd like to script restarting aiccu when this happens. sixxs, understandably, frowns very heavy on automatic reconnects because of the load repeated (and, presumably, failed) connect attempts give on the infrastructure. At the same time, manual restarts partially defeat the purpose of IPv6 tunnels: I really don't want to go into the mess of figuring out the dynamic-IP-of-the-day to remotely log in, and manually restart. I would like sixxs to define a policy saying 'xxx connect attempts, interspersed over yyy minutes, with a minimum of zzz minutes between connects, are not considered to be abuse'. Would such a policy be possible and what would the parameters be? Just a few more days of IPv4-space in the RIPE-region, Geert Jan

Policy for aiccu reconnects?

Jeroen Massar SixXS Staff

on Thursday, 13 September 2012 11:15:36

One of my tunnels died when the cable-ISP changed IPv4 on me (again).

What do you mean with "died" ? Please provide technical details so that we can work on a proper solution.

I'd like to script restarting aiccu when this happens.

Explicitly NOT allowed. You should solve the cause of the problem instead. One does not have to restart Windows anymore either just because there is something wrong, thus resolve your problem.

I would like sixxs to define a policy saying 'xxx connect attempts, interspersed over yyy minutes, with a minimum of zzz minutes between connects, are not considered to be abuse'.

We are not going to define such a policy or disclose when our system determines that the limits in place are being hit as you should resolve the real problem.

Policy for aiccu reconnects?

Shadow Hawkins on Friday, 14 September 2012 20:08:54

We are not going to define such a policy or disclose when our system determines that the limits in place are being hit as you should resolve the real problem.

Understand, which is why I'm asking..

What do you mean with "died" ? Please provide technical details so that we can work on a proper solution.

The cable-isp is apparently doing works on the RF plant for a few weeks now (my RF spectrum analyzer confirms this). There are several symptoms: a. sudden, pathetically low DHCP lease times (1h, 30 mins) instead of the "usual" 3 days, b. nightly RF outages lasting several hours (i.e. past the DHCP lease time), c. a cable modem starting handing out 90 second leases for 192.168.100.0/2, apparently when the RF link is down (you may know that on many docsys modems, http://192.168.100.1 gives the modem's status page), d. a confused border router box when things (partially) come up. Currently I power-cycle the kit (which is the ISP's "recommended fix") when I have physical access. Sometimes, I do not. Again, I'm not trying to play the system but trying to find an acceptable workaround for a problem neither you nor I have caused. That is all. Geert Jan

Policy for aiccu reconnects?

Shadow Hawkins on Friday, 14 September 2012 20:10:51

I forgot, but relevant: changed public-IP after these works. Cold-starting all the kit does give me the "old" IP back though. No, I don't fully understand this either.

Policy for aiccu reconnects?

Jeroen Massar SixXS Staff

on Sunday, 16 September 2012 12:12:11

None of these issues should have a requirement for AICCU to restart, AYIYA should work fine in such an environment except that you have broken connectivity once in a while because IPv4 does not properly work.

Please note Posting is only allowed when you are logged in.