|
High "Hash Fail" Value for AYIYA tunnel
![[de]](/s/countries/de.gif) Shadow Hawkins on Sunday, 30 September 2012 13:43:44
Hello,
I am operating two tunnels, one being a heartbeat and the other one being a AYIYA tunnel. While within the new "Live Tunnel Status" pages for these tunnels the error count for "Hash Fail" is "none" in case of my heartbeat tunnel, in contrast any IPv6/AYIYA/UDP/IPv4 packet seems to produce a "Hash Fail" resulting in a present error count value of more than 16 millions. Does "Hash Fail" mean, there is something wrong in the context of chapter 5 of the AYIYA specification (http://unfix.org/~jeroen/archive/drafts/draft-massar-v6ops-ayiya-02.html#anchor24)? But if so, my AYIYA tunnel shouldn't work, right? Fortunately I can use it without any problems instead. So maybe this error counter produces wrong results for AYIYA tunnels?
Andreas
High "Hash Fail" Value for AYIYA tunnel
The counters are correct, it just means that there are packets arriving which have a wrong hash.
Note that these counters can increase quickly when something is trying to send packets which are not valid, also note that the counters show since the PoP (or well sixxsd) started, thus that can be very long time spans.
High "Hash Fail" Value for AYIYA tunnel
![[es]](/s/countries/es.gif) Shadow Hawkins on Monday, 01 October 2012 09:37:26
Just for your information, I also have a high hash fail rate (using AYIYA):
Hash Fail 3431, last: ::94f2:ab6:68f2:ab6 2012-10-01 09:33:57 (1349084037; 0 days 00:00:17 ago)
The tunnel seems to be working OK.
Thank you.
High "Hash Fail" Value for AYIYA tunnel
3431 packets arrived over a period of almost a year that looked like destined for your tunnel.... nothing odd with that.
I think we might need to be adding a 'reset error count' option as it seems a lot of people are freaking about this.
High "Hash Fail" Value for AYIYA tunnel
Shadow Hawkins on Monday, 01 October 2012 10:01:40
Hi again,
this is the present situation for my AYIYA tunnel:
Hash Fail 16201324, last: :: 2012-10-01 09:53:33 (1349085213; 0 days 00:00:02 ago)
So 16201324 is quite high, isn't it? And if I sit at the end of this tunnel clicking "Refresh" at the tunnel's "Live Tunnel Status" page, the counter is raised by another 10-30 packets the last one always being 0 seconds away. So this is just about the traffic I produce to refresh this page and so I draw the conclusion that any packet I send has a wrong hash. This looks really odd to me.
Andreas
High "Hash Fail" Value for AYIYA tunnel
So 16201324 is quite high, isn't it?
Over a period of 9 months, maybe not. If those packets where all transmitted at the same time.
But it does seem you are getting more unhashed packets every once in a while.
This looks really odd to me.
Packets that are reported there are dropped, as such they are not forwarded anywhere or processed.
Note also that the Refreshing of that page is limited on a daily basis.
These numbers are indicators, but there are a lot of factors that can be causing them.
Eg, you might have a second instance running somewhere which is still generating packets but with a wrong password and thus producing wrong hashes.
It might also be that some network component is echoing out broken packets, eg half the packet.
Unfortunately I have little time to start looking into every error packet.
High "Hash Fail" Value for AYIYA tunnel
![[dk]](/s/countries/dk.gif) Shadow Hawkins on Tuesday, 23 October 2012 06:15:31
I can beat that:
24265495, last: :: 2012-10-23 06:11:37 (1350972697; 0 days 00:00:00 ago)
And my tunnel seems to be working fine...
Leif
High "Hash Fail" Value for AYIYA tunnel
Shadow Hawkins on Wednesday, 24 October 2012 15:40:03
Today I changed my tunnel from type AYIYA to type 'heartbeat'. Since then the 'Hash Fail' value stopped to increase and is frozen to its last AYIYA-value respectively:
17597010, last: :: 2012-10-24 09:16:47 (1351070207; 0 days 06:11:14 ago)
But of course, 'Hash Fail' might have no meaning for heartbeat tunnels(?) According to http://tools.ietf.org/id/draft-massar-v6ops-heartbeat-01.txt 'hash' can only refer to the md5-signatures of heartbeat packets which are only sent once a minute whereas with AYIYA a hash happens in every IPv6/AYIYA packet sent through the UDP/IPv4 tunnel (in addition to the AYIYA heartbeats).
So if at all the 'Hash Fail' value can only raise with a maximum rate of 1/minute for heartbeat tunnels.
High "Hash Fail" Value for AYIYA tunnel
But of course, 'Hash Fail' might have no meaning for heartbeat tunnels(?)
As you have noted the hash fails are for both AYIYA and heartbeat, and they indeed can only happen when a packet arrives which is a lot more often for AYIYA (per packet) and heartbeat (only one heartbeat per minute).
The counter is actually split in newer version of the daemon which we are gradually pushing out to all the PoPs.
|
![[ch]](/s/countries/ch.gif)
on Sunday, 30 September 2012 14:07:41
Posting is only allowed when you are