Windows Deployment Services (WDS) cache

We’ve had a constant frustration where clients would try to PXE boot more than once within an one hour period with the result being that the PXE boot fails. This problem occurs because of a PXE cache limitation (1 hour) set by default.

You can also confirm this configuration by looking in the smspxe.log file:

Loaded PXE settings from reg key HKLM\Software\Microsoft\SMS\PXE:
PXE Settings:
IsActive: Yes
SupportUnknownMachines: No
MACIgnoreListFile:
ResponseDelay: 0
CacheExpire: 3600
HTTP Port: 80
HTTPS Port: 443
IISSSLState: 0x0
BindPolicy: Exclude
TRK:
SiteSignCert:
Root CA Certs:
PXE GUID: fab8bfcc-3b15-431a-828d-fbfd184c3813
PXEPassword:

To fix this, reduce the cache size. Open regedit and navigate to HKEY_LOCAL_MACHINE\Software\Microsoft\SMS\PXE (or for 64-bit OS, head to HKEY_LOCAL_MACHINE\SOFTWARE\Wow6432Node\Microsoft\SMS\PXE). Modify the CacheExpire DWORD and set it to a lower value, in my case I set it to decimal value of 1, meaning that cache will only last for 1 second.

Restart the WDS service.

Further info here – http://support.microsoft.com/kb/2019640

Possibly related posts (auto generated):

1. Schedule SCCM reports to be emailed without SQL reporting services
2. Slow SCCM OSD TFTP / PXE
3. SuperFlow for SCCM Operating System Deployment via PXE