SixXS::Sunset 2017-06-06

Ticket ID: SIXXS #13507837
Ticket Status: User

PoP: decgn01 - NetCologne Gesellschaft fur Telekommunikation mbH (Cologne)

aiccu (linux) instable
[de] Shadow Hawkins on Friday, 17 April 2015 11:46:43
Running aiccu 2007.01.15 in debian 7.8 x64 environment at vanilla kernel 4.0/3.19.4 with the problem tunnel stops working after about 20 hours. If I run 'ifconfig', there is no sixxs interface, but if I run 'ifconfig sixxs' it shows interface information. I can't stop aiccu deamon because it's not running, I can't start aiccu (no error message) perhaps because interface sixxs is half on life. I can't 'ifdown sixxs' because interface is not configured. But I'm able to run aiccu using an equal config with interface sixxs2. This way 'ifconfig sixxs' is empty, I can stop aiccu and restart using interface sixxs again. So something seems going down but not removing clean sixxs interface. Any hints? best regards, Ronny aiccu.conf : username RSD9-SIXXS password ***** protocol tic server tic.sixxs.net ipv6_interface sixxs tunnel_id T75690 verbose true daemonize true automatic true requiretls false defaultroute true routing : 0.0.0.0 0.0.0.0 0.0.0.0 U 0 0 0 ppp0 192.168.99.0 0.0.0.0 255.255.255.0 U 0 0 0 eth0 192.168.100.0 0.0.0.0 255.255.255.0 U 0 0 0 eth0 217.0.116.25 0.0.0.0 255.255.255.255 UH 0 0 0 ppp0 firewall : problem is not firewall related interface : in not working moments its-gw:~# ifconfig sixxs sixxs Link encap:UNSPEC Hardware Adresse 00-00-00-00-00-00-00-00-00-00-00-00-00-00-00-00 PUNKTZUPUNKT NOARP MULTICAST MTU:1280 Metrik:1 RX packets:347612 errors:0 dropped:0 overruns:0 frame:0 TX packets:341520 errors:0 dropped:0 overruns:0 carrier:0 Kollisionen:0 Sendewarteschlangenlnge:500 RX bytes:219489370 (209.3 MiB) TX bytes:48766141 (46.5 MiB) its-gw:~# ifconfig eth0 Link encap:Ethernet Hardware Adresse e0:3f:49:eb:29:b5 inet Adresse:192.168.100.1 Bcast:192.168.100.255 Maske:255.255.255.0 inet6-Adresse: fe80::e23f:49ff:feeb:29b5/64 Gltigkeitsbereich:Verbindung inet6-Adresse: 2001:4dd0:ff00:8abd::2/64 Gltigkeitsbereich:Global UP BROADCAST RUNNING MULTICAST MTU:1500 Metrik:1 RX packets:416804892 errors:0 dropped:0 overruns:0 frame:0 TX packets:189266795 errors:0 dropped:0 overruns:0 carrier:0 Kollisionen:0 Sendewarteschlangenlnge:1000 RX bytes:550630216826 (512.8 GiB) TX bytes:206450665753 (192.2 GiB) lo Link encap:Lokale Schleife inet Adresse:127.0.0.1 Maske:255.0.0.0 inet6-Adresse: ::1/128 Gltigkeitsbereich:Maschine UP LOOPBACK RUNNING MTU:65536 Metrik:1 RX packets:454766 errors:0 dropped:0 overruns:0 frame:0 TX packets:454766 errors:0 dropped:0 overruns:0 carrier:0 Kollisionen:0 Sendewarteschlangenlnge:0 RX bytes:644232712 (614.3 MiB) TX bytes:644232712 (614.3 MiB) ppp0 Link encap:Punkt-zu-Punkt-Verbindung inet Adresse:84.179.162.117 P-z-P:217.0.116.25 Maske:255.255.255.255 UP PUNKTZUPUNKT RUNNING NOARP MULTICAST MTU:1492 Metrik:1 RX packets:1752739 errors:0 dropped:0 overruns:0 frame:0 TX packets:1343870 errors:0 dropped:7 overruns:0 carrier:0 Kollisionen:0 Sendewarteschlangenlnge:3 RX bytes:1846991984 (1.7 GiB) TX bytes:162657892 (155.1 MiB)
State change: user Locked
[ch] Jeroen Massar SixXS Staff on Friday, 17 April 2015 11:48:46
Message is Locked
The state of this ticket has been changed to user
aiccu (linux) instable
[ch] Jeroen Massar SixXS Staff on Friday, 17 April 2015 11:52:58
Running aiccu 2007.01.15 in debian 7.8 x64 environment at vanilla kernel 4.0/3.19.4
Why are you running a Debian host with a "vanilla kernel"? Are you aware that user space and kernel have to match up? If you are wanting to run 4.0 you might consider running at minimum Debian unstable for things to aline a little bit.
If I run 'ifconfig', there is no sixxs interface, but if I run 'ifconfig sixxs' it shows interface information.
Sounds like something is removing the IP address and/or configuring it as 'down'. You also might want to check the 'ip' command which shows a lot more details.
I can't stop aiccu deamon because it's not running,
What do the logs tell? Was it killed?
So something seems going down but not removing clean sixxs interface
When the aiccu process exits for an AYIYA tunnel the related interface is closed automatically as there is no process associated to it anymore. Hence, if it is only half there and the process is not running anymore, your system is odd. Log files and more details are needed to even remotely look into this.
aiccu (linux) instable
[de] Shadow Hawkins on Tuesday, 21 April 2015 09:51:39
Why are you running a Debian host with a "vanilla kernel"?
I think the answer does not matter.
Are you aware that user space and kernel have to match up?
Yes, I think so. For years there were no problems using my own builds of vanilla kernels with debain user space. Back on "my" 3.19.3, running for weeks stable, I can confirm same behavior. I now can focus the times of losing connection to times of ppp reconnect (WAN). But since years we life with at least one ppp reconnect daily and sixxs/aiccu was self-healing. As a possible solution I'm able to force this reconnect. I tried to stop aiccu, wait a bit, reconnect ppp, wait another while and restart aiccu - but same behavior (aiccu is not starting again because interface is half-online).
You also might want to check the 'ip' command which shows a lot more details.
I'll use next time
What do the logs tell? Was it killed?
Apr 21 00:00:02 its-gw aiccu[2083]: Already running instance signaled, exiting Apr 21 00:00:47 its-gw aiccu[2241]: Successfully retrieved tunnel information for T75690 Apr 21 00:00:59 its-gw aiccu[2318]: Successfully retrieved tunnel information for T75690 Apr 21 00:00:59 its-gw aiccu[2323]: AICCU running as PID 2323 Apr 21 00:00:59 its-gw aiccu[2323]: [tun-start] Couldn't set interface name to sixxs: Device or resource busy (16)
aiccu (linux) instable
[ch] Jeroen Massar SixXS Staff on Tuesday, 21 April 2015 10:02:20
Apr 21 00:00:02 its-gw aiccu[2083]: Already running instance signaled, exiting
Apr 21 00:00:47 its-gw aiccu[2241]: Successfully retrieved tunnel information for T75690
Apr 21 00:00:59 its-gw aiccu[2318]: Successfully retrieved tunnel information for T75690
Obviously something is already restarting AICCU, with 12 seconds a part of that, seems you are trying to run it a lot of times in parallel. I can only suggest checking all the magic customisations you are doing to your host. Nothing we can do about.
aiccu (linux) instable
[de] Shadow Hawkins on Wednesday, 29 April 2015 07:55:29
I can only suggest checking all the magic customisations you are doing to your host.
I've disabled ANY magic, also the debian one. Now aiccu starts only once with booting the linux host, reaching runlevel 2. Around midnight I force ppp refesh using 'poff', 'sleep 10' an 'pon'. That's the aiccu relevant log now: Apr 28 10:28:25 its-gw aiccu[28202]: Response not accepted: Database maintainance is in progress. Apr 28 10:28:25 its-gw aiccu[28202]: Couldn't retrieve first tunnel for the above reason, aborting Apr 28 10:28:44 its-gw aiccu[28226]: Successfully retrieved tunnel information for T75690 Apr 28 10:28:44 its-gw aiccu[28231]: AICCU running as PID 28231 Apr 28 10:28:44 its-gw aiccu[28231]: [AYIYA-start] : Anything in Anything (draft-02) Apr 28 10:28:44 its-gw aiccu[28231]: [AYIYA-tun->tundev] : (Socket to TUN) started Apr 28 10:28:44 its-gw aiccu[28231]: [AYIYA-tundev->tun] : Error (-1) while sending 120 bytes to network: Operation not permitted (1) Apr 29 00:00:06 its-gw aiccu[28231]: [AYIYA-tundev->tun] : Error (-1) while sending 146 bytes to network: Invalid argument (22) As you can see, in night from 28th to 29th there were no more aiccu restarts. But last line repeats 4 times a second from this moment until I manually stopped aiccu. I also manually forced this behavior, grepping aiccu log entries: 1) running/working ppp and aiccu - nothing logged 2) poff - "aiccu[19955]: [AYIYA-tundev->tun] : Error (-1) while sending 124 bytes to network: Invalid argument (22)" 4 times a second, amount of bytes differing 3) pon dsl-provider - further this log entries 4) aiccu stop - "aiccu[26466]: Already running instance signaled, exiting" 5) aiccu start - "aiccu[26488]: Successfully retrieved tunnel information for T75690", "aiccu[26493]: AICCU running as PID 26493", "aiccu[26493]: [tun-start] Couldn't set interface name to sixxs: Device or resource busy (16)" Something is keeping his hand over the "sixxs" interface, I believe. Next I'll try: aiccu stop -> poff -> pon -> aiccu start

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

Static Sunset Edition of SixXS
©2001-2017 SixXS - IPv6 Deployment & Tunnel Broker