hvserv
asked on
Windows 7 connectivity problems
Our small business network has been having reproducible but unexplained connectivity problems lately. Router is Linksys WRT54G (DHCP), machines on network run several flavors of XP, a couple Vista, and a handful Win7. A Macbook also connects from time to time. All, with the exception of the Mac are in the same workgroup. We are not using homegroup. I have disabled IPv6 on all Win7 and Vista machines by unchecking it in adapter properties. I have made no registry changes in this regard although I have read about some that may be necessary. Everything has been running smoothly until recently when an new Win7 Pro x64 machine was added. There was already a similar machine on the network, but the new one replaced our Quickbooks (2008 Pro) server which had been XP. Then things started to happen. Other machines would lose internet. Often they would be able to see local shares and interact with them but no internet. Sometimes neither. Tried reboots, ipconfig /renew, router power cycling, manually resetting winsock settings and combinations of the above.
eventually, the affected machine/s would reconnect. Winsock changes/restart/router cycle seemed to be most effective. What is odd, is that not all machines would be affected at the same time. One might go down. Or two or three, but never all and never the Quckbooks server. In other words, the router and the new machine were always happy together. Only one other machine accesses the Quickbooks data, but is on all the time. It loses connectivity.
This is beginning to sound like a logic puzzle, sorry. Conicidentally, another local business with a similar but smaller network, also running Quickbooks (2010 Pro) on a win7 machine is experiencing identical problems. They are also using a Linksys router (wired only). IPv6 is configured but not needed. Both offices receive internet through the same local Wifi provider.
Often, they will lose internet on their server but retain local share access including QB access.
Hope you can offer some insight - I am out of ideas.
eventually, the affected machine/s would reconnect. Winsock changes/restart/router cycle seemed to be most effective. What is odd, is that not all machines would be affected at the same time. One might go down. Or two or three, but never all and never the Quckbooks server. In other words, the router and the new machine were always happy together. Only one other machine accesses the Quickbooks data, but is on all the time. It loses connectivity.
This is beginning to sound like a logic puzzle, sorry. Conicidentally, another local business with a similar but smaller network, also running Quickbooks (2010 Pro) on a win7 machine is experiencing identical problems. They are also using a Linksys router (wired only). IPv6 is configured but not needed. Both offices receive internet through the same local Wifi provider.
Often, they will lose internet on their server but retain local share access including QB access.
Hope you can offer some insight - I am out of ideas.
ASKER
I should have added that one aspect of this problem, in most instances but not all, is loss of DHCP for some reason. ie a computer is turned on, but fails to get an address. Other machines, which had been on all along are still enjoying full connectivity. A router restart does not fix the confused machine. However, a winsock reset, router reset, and machine reboot does. If the Win7 QB server is off, problem
does not occur with newly booted machines. Here is a chronological outline of one instance:
Day 1 - server on, client machine on -everything fine
Day 1 - (later) client machine off, server happy
Day 1 - (later still) server shut off for the night, client machine still off
Day 2 - Server turned on, gets address, surfs, happy.
Day 2 - (a little later) client turned on, does not get address.
Day 2 - reset client winsock settings, restart, no DHCP, restart router, client gets address.
server has connectivity the whole time both before and after router restart.
I will implement the DNS recommendation - good idea.
does not occur with newly booted machines. Here is a chronological outline of one instance:
Day 1 - server on, client machine on -everything fine
Day 1 - (later) client machine off, server happy
Day 1 - (later still) server shut off for the night, client machine still off
Day 2 - Server turned on, gets address, surfs, happy.
Day 2 - (a little later) client turned on, does not get address.
Day 2 - reset client winsock settings, restart, no DHCP, restart router, client gets address.
server has connectivity the whole time both before and after router restart.
I will implement the DNS recommendation - good idea.
Are the clients that get no DHCP parameters connecting via wireless?
If so, you may have too much interference or too many Wi-Fi networks on the same channel.
Use inSSIDer 2 ( http://metageek.net/products/inssider ) to see which channel is not being used or has the fewest access points, and configure your WRT54G to use that channel.
If so, you may have too much interference or too many Wi-Fi networks on the same channel.
Use inSSIDer 2 ( http://metageek.net/products/inssider ) to see which channel is not being used or has the fewest access points, and configure your WRT54G to use that channel.
ASKER
all but the Mac are wired connections.
added google DNS and that's nice but a side issue as DHCP
loss is big part of puzzle and not related to DNS issues.
Did a minor firmware upgrade to router Linksys WRT54G hardware v6.0
went from v1.02.6 to v1.02.8 firmware. Network has been stable since (48 hours)
support notes for firmware are as follow:
Firmware 1.02.8
- Resolves issue with UPnP denial of service security vulnerability.
Firmware 1.02.7
- Resolves issue with not being able to block IPSec pass-through.
- Resolves issue with router disconnecting from some particular online game servers.
- Resolves issue with user not being able to get incoming video and audio when video phone is connected to the LAN hosted by the router.
- Resolves issue with WEP 64-bit key not being properly validated when user enters the key.
- Resolves issue with disabling UPnP not being effective until the router is rebooted.
- Resolves issue with Wireless MAC Filter not working properly when there are more than 32 entries in filter list.
- Resolves issue with DNS vulnerability.
- Better handling of TCP packet with invalid option field.
- Updated TZO.com DDNS service support.
- Adds PPPoE service name support.
- HNAP 1.2 support.
not sure what might have been relevant here, will watch it for another day or so and report back.
If this is the fix, and someone can identify the firmware element/elements that made a
difference I would love to know what it was.
added google DNS and that's nice but a side issue as DHCP
loss is big part of puzzle and not related to DNS issues.
Did a minor firmware upgrade to router Linksys WRT54G hardware v6.0
went from v1.02.6 to v1.02.8 firmware. Network has been stable since (48 hours)
support notes for firmware are as follow:
Firmware 1.02.8
- Resolves issue with UPnP denial of service security vulnerability.
Firmware 1.02.7
- Resolves issue with not being able to block IPSec pass-through.
- Resolves issue with router disconnecting from some particular online game servers.
- Resolves issue with user not being able to get incoming video and audio when video phone is connected to the LAN hosted by the router.
- Resolves issue with WEP 64-bit key not being properly validated when user enters the key.
- Resolves issue with disabling UPnP not being effective until the router is rebooted.
- Resolves issue with Wireless MAC Filter not working properly when there are more than 32 entries in filter list.
- Resolves issue with DNS vulnerability.
- Better handling of TCP packet with invalid option field.
- Updated TZO.com DDNS service support.
- Adds PPPoE service name support.
- HNAP 1.2 support.
not sure what might have been relevant here, will watch it for another day or so and report back.
If this is the fix, and someone can identify the firmware element/elements that made a
difference I would love to know what it was.
I see you have dhcp setup. Is everything on dchp? Do you have any static IP addresses? You might want to setup static IP's on your server and any permanent workstations then change the range of DHCP address leases so that it doesn't include that range. Example. Static range of 10.0.0.1 - 10.0.0.20 for servers, default gateway, network printers and workstations with shares. Dhcp lease range 10.0.0.21-10.0.0.50 or whatever will work for the size of your network. Do you have a DNS server for internal clients?
When a computer loses connectivity find out what it's IP address is then do an nslookup on that IP to see if there is another computer with that address. If it is a 169. address then it is not getting an address from the server.
When a computer loses connectivity find out what it's IP address is then do an nslookup on that IP to see if there is another computer with that address. If it is a 169. address then it is not getting an address from the server.
Another problem may be a rouge wireless access point or DCHP server. You may be getting a bad address that is outside your network. Checking IP's will help rule this out.
ASKER
router gets DHCP from Wifi provider's hardware outside of building on antenna.
machines on network get DHCP, within a specified range, from router. 192.168.199.xxx
If one or more machines lose their address (I don't know how) they show a 169.... number.
there is no rogue access point or DNS server downstream of the router.
machines on network get DHCP, within a specified range, from router. 192.168.199.xxx
If one or more machines lose their address (I don't know how) they show a 169.... number.
there is no rogue access point or DNS server downstream of the router.
Have you tried entering a static IP when they lose their address?
Also have you looked at the logs on the router to see if you are getting dropped packets? If you are getting a large about of tx/rx errors on port, the router will turn off the port to keep from shutting the whole system down.
How are they wired? Home run to the Linksys or switches that connect to a linksys? You could try switching ports when a computer loses an address to see if that port is being shutdown.
Also have you looked at the logs on the router to see if you are getting dropped packets? If you are getting a large about of tx/rx errors on port, the router will turn off the port to keep from shutting the whole system down.
How are they wired? Home run to the Linksys or switches that connect to a linksys? You could try switching ports when a computer loses an address to see if that port is being shutdown.
ASKER
stations are wired home run to a single 24-port switch that
connects to the router. have enabled logging - thanks for the
heads-up on that. When a given machine loses its address, it
will not be assigned a new address unless I reset winsock settings,
restart machine, and restart router. Other machines maintain
connectivity up until I restart the router. Have tried switching
network drops for failed machine to give it a shot at a different
port in the switch, but this does not help. will try static IP next time
one fails.
connects to the router. have enabled logging - thanks for the
heads-up on that. When a given machine loses its address, it
will not be assigned a new address unless I reset winsock settings,
restart machine, and restart router. Other machines maintain
connectivity up until I restart the router. Have tried switching
network drops for failed machine to give it a shot at a different
port in the switch, but this does not help. will try static IP next time
one fails.
ASKER CERTIFIED SOLUTION
membership
Create a free account to see this answer
Signing up is free and takes 30 seconds. No credit card required.
ASKER
solved the problem - all my earlier assumptions were wrong.
Regards,
Mandar Waman