2016-09-06 06:10 AM
Have been doing a series of NDU upgrades > 824P4 7-mode, culminating soon on a heavily used CIFS vfilers with thousands of shares.
System is a FAS 3240 HA pair 8.1.2P4
CIFS is not configured on physcial filers....so whilst the UA procedure discusses running "cifs terminate" this is not appropriate as cifs only
configured on vfilers.
But the specific Upgrade Advisor (UA) doesn't advocate running a cifs terminate on each of the cifs vfilers .....and ....
Moreover, something like "vfiler run * cifs terminate" has caused one of my colleagues problems in past where the command hung up.
So should i expect cf takeover and cf giveback to deal with all of these CIFS sessions ?
So please what is advise of operation experience on doing NDU ONTAP upgrades where lots of CIFS usage...please noteam familiar with the general Upgrade Advisor procedure , which works well, its just this system has vfilers with loads of shares that am concerned with ...... any thoughts appreciated ... john
2016-09-06 12:35 PM
When doing a NDU you should terminate cifs on physical as well as all vfilers (cifs terminate -t 0).
In huge environments with plenty of open shares, files, locks etc. you might need to run a lock break -p cifs to get rid of oplocks, then you can use cifs terminate -t 0 and command wont get stuck.