I am using the DB Server and have noticed recently that the date windows reports for a file last modified date may not match up with what the database says until a few minutes (roughly) after the last session has closed down in the database.

I am not sure if this is an issue with our server (not DB Server)  or the fact I have recently changed the server to allows SMB2 so it will work with Windows 10.

I think this might be the windows file cache not updating the file date time until all file handles have been released but just wanting to check if there is something odd going on that I should be concerned about.

I only noticed this because at the end of the day I copy using robocopy the database tables and on one day some tables had updated but others hadn't.  This was because there was a disconnected session that did not expire until after the robocopy.  I have never noticed this before.

Anyone else seen this activity?

David,

<< I am using the DB Server and have noticed recently that the date windows reports for a file last modified date may not match up with what the database says until a few minutes (roughly) after the last session has closed down in the database.

I am not sure if this is an issue with our server (not DB Server)  or the fact I have recently changed the server to allows SMB2 so it will work with Windows 10.

I think this might be the windows file cache not updating the file date time until all file handles have been released but just wanting to check if there is something odd going on that I should be concerned about. >>

You're correct:

https://blogs.technet.microsoft.com/asiasupp/2010/12/14/file-date-modified-property-are-not-updating-while-modifying-a-file-without-closing-it/

Tim Young
Elevate Software
www.elevatesoft.com

Thanks Tim.

I note from the manual I can use table.flushbuffers to force the OS to write the file to disk.  I was thinking I might put this into a few select places and forcing this on every write will be a performance hit.  

Is there any possible issues with multiple clients calling flush buffers in a C/S enviroment on the same table? or is this a safe thing to impiment for DI?

Regards
David.

Tim Young [Elevate Software] wrote:

I think this might be the windows file cache not updating the file date time until all file handles have been released but just wanting to check if there is something odd going on that I should be concerned about. >>

You're correct:

https://blogs.technet.microsoft.com/asiasupp/2010/12/14/file-date-modified-property-are-not-updating-while-modifying-a-file-without-closing-it/

Tim Young
Elevate Software
www.elevatesoft.com

David,

<< I note from the manual I can use table.flushbuffers to force the OS to write the file to disk.  I was thinking I might put this into a few select places and forcing this on every write will be a performance hit.  

Is there any possible issues with multiple clients calling flush buffers in a C/S enviroment on the same table? or is this a safe thing to impiment for DI? >>

It's safe, but, as you say, you do need to be careful so as to prevent any performance issues.  The flushing can cause the disk operations to start queueing if you hit it with a lot of writes, although an SSD can certainly help quite a bit in such cases.

Tim Young
Elevate Software
www.elevatesoft.com