CLOSED Build 7904

[stapp 130]

Win 7 64 bit via manual updates ( with EAM service stopped in Services and program shutdown)

Update finally installed after 12 minutes  (I almost fell asleep )

I know it was a big download and instal,l but 12 minutes is a long time. I don't know a lot about page faults but they went up to a million during the install.

Had a look around and all seemed ok, so I asked to do a malware scan. It asked to restart GUI to load things as usual.

Page faults went up to 2 million during scan (hope I'm not getting to get obsessed with them!!)

After it found my test file it brought up the new window to see if I would like to use Emsisoft as ''we found malware that managed to get through your current defence''  (much better wording now)

I clicked on the option to see what would happen  and it completely started up EAM again, even restarting the service !! At the same time EEK shut down and disappeared.

Just in case they are of any interest I attach the debug logs.

 

 

a2emergencykit_20170825191141(456).zip

[Frank H 91]

hi stapp,

did you ever read :https://en.wikipedia.org/wiki/Page_fault ?

I'll ask Fabian for feedback on this, maybe it's normal maybe it's not.

have a nice weekend

[Frank H 91]

The pagefaults (which are no faults in fact)  appear to be high, but related to some other common processes, you will notice that actually 1 million is normal.

monitored during a few days:

• EAM: 20 million
• WMI: over a billion
• each Chrome instance has around 15 - 20 Million
• Skype is close to 2 billion

 

[JeremyNicoll 57]

Some OSes' management of virtual memory is better than others' - though I suppose it depends on what the OS designers were trying to achieve.  Windows has - as long as I've been aware of it - not been all that good (in my biased opinion).  Windows will use increasing amounts of virtual memory when there's still available/unused free real memory.  Both linux and (in my experience) z/OS don't use (much) vm until there's a dire need for it.  It's possible to run a z/OS system with no (or hardly any) page faults at all, even though paging datasets/files will be defined.   Anyone used to the latter sort of behaviour will be mystified as to why a Windows system, even with vast amounts of real RAM, will still take page faults.   If you google appropriately you can find reams of discussions about this...

[Christian Mairoll 228]

I'd guess that the number of page faults gets smaller when you disable the memory optimization feature (that effectively pushes the signatures and other stuff to the pagefile on the disk).