Hyper-V WIndows Server 2012 R2 event ID 157 Disk x has been surprise removed using windows backup

i've found few postings on event id 157 windows 2012 r2, no answers

we've upgraded server 2012 2012 r2. hyper-v role. event id 157 occurs when windows backup running backup hyper-v machines. people have reported same issue having. have id 157 occurring only during windows backup process or if use vss snapshots backup hyper-v virtual machines.

i've found event id 157 designed 2012 r2 storage groups when disk removed, seems bug showing during windows backup.

one time (out of 10 - once each day) that event logged, software reset sent lsi mega raid controller half of array , consistency checks done. hope not happen again.

everything worked fine before moved 2012 2012 r2. i've done clean install , still happens.

i've looked through event logs on virtual machines , have 1 windows 2008 server running sql 2008 r2 , chokes through entire backup process hundreds of errors, starting sql vss writer receiving command not understand. onward says tables , data corrupt, normal sql maintenance job have running sql server backs databases , performs consistenacy checks without issues.

i've remove sql server windows backup , see how goes tonight.

in case related sql server, here's error 1st logged windows backup on host begins:

event id 513, capi2, error, cruptographic services failed while processing onidentity() call in system writer object.

details: addcorecsifiles: beginfileenumeratrion() failed.

system error: access denied.

after freezes databases, gives dozens of errors each table, , restores each table normal operation.

the guest sql server appears 1st vm backed up, doubt has it, thought should provide many details possible, when else has discovered vss related , guest reporting vss errors.

what think?
thanks
john
p.s. there no hard drive errors, running check disk not help.

i looked sql errors , found fix set permissions, solved small portion of vss errors on sql virtual machine. did notice in same area changing settings work 4k sector disk.... ok, used power shell query vm , physical drives on older hardware has 4k sector drives , database vm imaged , turned vm. new server has 512k older\normal sector drives.

when remove vm windows 2008\sql 2008r2 there no errors or bad events in event log on host , no event id 157 or "disk surprise removed".

my gut feeling because 2008 vm created physical machine had 4k sector drives, vss not able function properly. base on 4k event messages , event log complains database tables being corrupted during host windows backup.

oddly, either went unnoticed or did not happen on windows 2012 hyper-v host, on same server hardware. noticed time, because threw port2 reset lsi megaraid controller reset half of hard drives , sent email alerts me telling me drives had been removed while still in use. originally, thought lsi email alert\lsi events same thing 2008 server reporting "surprise removed", not, @ least after 1st day, there has not been software reset sata raid controller.

i'm summing windows 2008 server being run on new vhd has both logical , physical 512k sector hard drives, because vm imaged (sysinternals disk2vhd or whatever called) thinks physical hard drives still 4k sectored!!! , vss shadow\snapshot\backup thing trying read 4k aligned files , failing saying corrupted. oh, mentioned in event log on sql server - adjust databases work on 4k aligned sector disk.

Windows Server  >  Hyper-V