Windows 7 - No error, slow scan

[rmeleiro]

Hi
We've upgraded most of our notebook and desktops to Windows 7 and a small number of them have a strange behaviour, which never happened before. For a handful of those W7 machines, located in a different network segment, only the machine gets recorded and no information appears for a reasonable amount of time. Several hours later information starts being shown - it may take a full working day for a single machine to be fully scanned.
We know Lansweeper activity is going on as tcp connections are established and, should the machine be shut down WMI errors will be present. This only happens with Windows 7 machines, Windows 2008 (not R2), Vista and XP on the same given segment work just fine. No other Windows 7 on the local segment experience this problem.
Lansweeper version is 3.5.2 freeware and it is running on a Windows 2008 Hyper-V guest server.
Any help appreciated.

Regards
Rui Meleiro

[rmeleiro]

For what is worth, I’ve narrowed down the problem to the well know Window scaling problem (RFC 1323 and http://technet.microsoft.com/pt-pt/library/bb878127(en-us).aspx ). In fact, once I’ve disabled autotuning for this particular segment, Lansweeper started acting accordingly. We had already done that in the past on Windows Vista machines on that location but on a different issue so correlation was not apparent. When upgrading to Windows 7, that setting was lost (it is now on GPO, to not be forgotten again). As a reference, disabling autotuning is made via netsh interface tcp set global autotuninglevel=disabled or Computer Configuration -> Windows Settings -> Policy Based QoS -> Right Click: Advanced Qos Settings -> Inbound TCP Traffic is a set of levels that map 0 -> disabled. This is a workaround, not a solution - the real alternative was to change the VPN router on the site - the Cisco IOS version we're using there has no provisioning for Windows scalling. But that's another story....

[rmeleiro]

No, that's not IPV6. We're using IPV6 (ISATAP and Teredo for DirectAccess) on all machines and have no problems with that. As I referred, a Windows 2008 server on that specific segment (with IPV6) has no problems also. For troubleshooting, I've disabled ISATAP and IPV6 all together on those Windows 7 machines with the exact same result. I've narrowed it down to WMI as wimmgmt.msc presents the very same delays. Mind you, it does not show any error, and no timeouts whatsoever. The machines get to be fully scanned in the end - they just take over 10 hours to do so.
Two of those W7 machines were upgraded in our HQ and then sent (without changes) to a different branch office 3100 kms away and only then the problems with Lansweeper were noticed. The Windows 2008 Server on that exact same location and other Windows XP machines are properly scanned and updated, so no firewall, routing and VPN encryption situations exist. I don't really think Teredo (IPV6 encapsulation on IPV4) is the culprit - even Windows XP machines are running Teredo.

[Hemoco]

Just a quick guess: IP6 problems in that network segment on the vlan.

Other OS don't have ip6

(just a guess)

[rmeleiro]

Update - it looks like a WMI problem (not Lansweeper’s) with those specific machines and locations. Testconnection.exe hangs indefinitely and the same happens with a remote access via WmiMgmt.msc. Oddly enough, Windows 7 machines on different segments are properly scanned, and non-Windows 7 machines on that same location are scanned without any problem. The exact same symptoms occur when using wmimgmt.msc in reverse order.