Unreliable wireless connections
Posted: Sun Feb 19, 2012 6:11 pm
Hello,
I am the Assistant Clerk in the stake and also work closely with the STS group since I am a "computer guy" by trade. We have a situation in one of our buildings where we are seeing strange problems with the wireless and I hope someone here might have an idea of what we can try to fix it...
We have the 881 firewall connected to Centurylink DSL and the wired access seems to be very stable. In order to add wireless access (at the stake president's request), we purchased a couple of Cisco WAP4410N devices and have them connected to a netgear switch which provide PoE to them. The behavior that we see is that a device can connect to the wireless and will seem to be fine. Then if we try to stream a video or download a large file, the download will often stop and the device will go into a "searching for Wifi" mode and then reconnect to the wifi. It gets the same IP address and the transfer may or may not resume automatically. This makes it so none of our instructors can use the wireless in their classes and keeps me constantly making excuses for our internet... I have a Mac and have looked at the console logs and this is the common info that I see:
1/22/12 2:14:23 PM kernel Auth result for: d0:d0:fd:66:57:2f MAC AUTH succeeded
1/22/12 2:14:23 PM kernel AirPort: Link Up on en1
1/22/12 2:14:23 PM kernel com.deterministicnetworks.driver.dne: dne_event: Proto 0 interface en, event KEV_DL_LINK_ON
1/22/12 2:14:23 PM kernel com.deterministicnetworks.driver.dne: socket_event_callback: en1 link on
1/22/12 2:14:23 PM kernel AirPort: RSN handshake complete on en1
1/22/12 2:14:23 PM kernel com.deterministicnetworks.driver.dne: socket_event_callback: en1 change in IFF flags
1/22/12 2:14:23 PM kernel com.deterministicnetworks.driver.dne: socket_event_callback: en1 change in IFF flags
1/22/12 2:14:23 PM configd[14] network configuration changed.
1/22/12 2:14:23 PM mDNSResponder[9978] RegisterInterface: Frequent transitions for interface en1 (10.135.20.163)
1/22/12 2:14:24 PM ntpd[38] bind() fd 25, family 30, port 123, scope 6, addr fe80::6233:4bff:fe29:f919, in6_is_addr_multicast=0 flags=0x11 fails: Can't assign requested address
1/22/12 2:14:24 PM ntpd[38] unable to create socket on en1 (213) for fe80::6233:4bff:fe29:f919#123
1/22/12 2:14:26 PM mDNSResponder[9978] RegisterInterface: Frequent transitions for interface en1 (FE80:0000:0000:0000:6233:4BFF:FE29:F919)
1/22/12 2:14:54 PM kernel AirPort: Link Down on en1. Reason 4 (Disassociated due to inactivity).
1/22/12 2:14:54 PM kernel com.deterministicnetworks.driver.dne: dne_event: Proto 0 interface en, event KEV_DL_LINK_OFF
1/22/12 2:14:54 PM kernel com.deterministicnetworks.driver.dne: socket_event_callback: en1 link off
1/22/12 2:14:55 PM configd[14] network configuration changed.
1/22/12 2:14:55 PM mDNSResponder[9978] DeregisterInterface: Frequent transitions for interface en1 (FE80:0000:0000:0000:6233:4BFF:FE29:F919)
It kind of looks like there is something happening with the DNS, but I am not an expert in networking.
The WAPs are not set to serve DHCP addresses and the assigned addresses are all in the pool from the 881, so it seems that the basic addressing is not a problem. We have also looked at the number of devices that are active and we are not out of addresses. We have shut down one of the access points and just left a single one running (thinking maybe they were interfering with each other) and the single device has the same problem. We don't see any "outside" devices interfering (this building is in a rural location), so we are pretty much stumped. We have considered removing the 881W from the path as a test to see if somehow it is causing the problem, but we haven't done that yet (and would need to do it when the building is mostly empty and we would have minimal risk of access to inappropriate sites). Does anyone have any suggestions as to what might be the problem and how we can get things working reliably?
We are also considering whether to move to the 1041n access points but are a bit concerned about the cost and timing (since they have to come from the FM group). The access via LDS account login sounds good, but I am also concerned about not being able to shut off the wireless access during a stake conference webcast. I am reading on the forum about the 1041n WAP's but any additional thoughts on this system would be appreciated.
Thanks,
Rick Klaus
I am the Assistant Clerk in the stake and also work closely with the STS group since I am a "computer guy" by trade. We have a situation in one of our buildings where we are seeing strange problems with the wireless and I hope someone here might have an idea of what we can try to fix it...
We have the 881 firewall connected to Centurylink DSL and the wired access seems to be very stable. In order to add wireless access (at the stake president's request), we purchased a couple of Cisco WAP4410N devices and have them connected to a netgear switch which provide PoE to them. The behavior that we see is that a device can connect to the wireless and will seem to be fine. Then if we try to stream a video or download a large file, the download will often stop and the device will go into a "searching for Wifi" mode and then reconnect to the wifi. It gets the same IP address and the transfer may or may not resume automatically. This makes it so none of our instructors can use the wireless in their classes and keeps me constantly making excuses for our internet... I have a Mac and have looked at the console logs and this is the common info that I see:
1/22/12 2:14:23 PM kernel Auth result for: d0:d0:fd:66:57:2f MAC AUTH succeeded
1/22/12 2:14:23 PM kernel AirPort: Link Up on en1
1/22/12 2:14:23 PM kernel com.deterministicnetworks.driver.dne: dne_event: Proto 0 interface en, event KEV_DL_LINK_ON
1/22/12 2:14:23 PM kernel com.deterministicnetworks.driver.dne: socket_event_callback: en1 link on
1/22/12 2:14:23 PM kernel AirPort: RSN handshake complete on en1
1/22/12 2:14:23 PM kernel com.deterministicnetworks.driver.dne: socket_event_callback: en1 change in IFF flags
1/22/12 2:14:23 PM kernel com.deterministicnetworks.driver.dne: socket_event_callback: en1 change in IFF flags
1/22/12 2:14:23 PM configd[14] network configuration changed.
1/22/12 2:14:23 PM mDNSResponder[9978] RegisterInterface: Frequent transitions for interface en1 (10.135.20.163)
1/22/12 2:14:24 PM ntpd[38] bind() fd 25, family 30, port 123, scope 6, addr fe80::6233:4bff:fe29:f919, in6_is_addr_multicast=0 flags=0x11 fails: Can't assign requested address
1/22/12 2:14:24 PM ntpd[38] unable to create socket on en1 (213) for fe80::6233:4bff:fe29:f919#123
1/22/12 2:14:26 PM mDNSResponder[9978] RegisterInterface: Frequent transitions for interface en1 (FE80:0000:0000:0000:6233:4BFF:FE29:F919)
1/22/12 2:14:54 PM kernel AirPort: Link Down on en1. Reason 4 (Disassociated due to inactivity).
1/22/12 2:14:54 PM kernel com.deterministicnetworks.driver.dne: dne_event: Proto 0 interface en, event KEV_DL_LINK_OFF
1/22/12 2:14:54 PM kernel com.deterministicnetworks.driver.dne: socket_event_callback: en1 link off
1/22/12 2:14:55 PM configd[14] network configuration changed.
1/22/12 2:14:55 PM mDNSResponder[9978] DeregisterInterface: Frequent transitions for interface en1 (FE80:0000:0000:0000:6233:4BFF:FE29:F919)
It kind of looks like there is something happening with the DNS, but I am not an expert in networking.
The WAPs are not set to serve DHCP addresses and the assigned addresses are all in the pool from the 881, so it seems that the basic addressing is not a problem. We have also looked at the number of devices that are active and we are not out of addresses. We have shut down one of the access points and just left a single one running (thinking maybe they were interfering with each other) and the single device has the same problem. We don't see any "outside" devices interfering (this building is in a rural location), so we are pretty much stumped. We have considered removing the 881W from the path as a test to see if somehow it is causing the problem, but we haven't done that yet (and would need to do it when the building is mostly empty and we would have minimal risk of access to inappropriate sites). Does anyone have any suggestions as to what might be the problem and how we can get things working reliably?
We are also considering whether to move to the 1041n access points but are a bit concerned about the cost and timing (since they have to come from the FM group). The access via LDS account login sounds good, but I am also concerned about not being able to shut off the wireless access during a stake conference webcast. I am reading on the forum about the 1041n WAP's but any additional thoughts on this system would be appreciated.
Thanks,
Rick Klaus