This post is more than 5 years old
8 Posts
0
1354
TimeFinder
In the -nocopy option, as per the product guide, it says that the copy happens when there is host I/O on target device or if there is a change in a track on the source device.
Can someone please clarify, in a TimeFinder/Clone operation, when and in what situations is the target device available for host I/O.
Thanks!
pablosanzoni_bn
22 Posts
0
November 13th, 2013 11:00
Hi Narohna,
You have made such a good question, let me answer you.
1 - In normal R1->R2 situation, and when you perform a timefinder snap or clone operation, from Host standpoint R1 is always RW for host and R2 is WD - Write Disable or what is the same Read Only to the host, since both vmax or dmx can't be RW at the same time to the host,
2 - Under normal conditions, I don't know if you deploy your timefinder operations thru SYMCLI or you have mainframe in your installation, anyway when you begin whith your copy or replica, it starts at local level(R1) and it propagates thru SRDF link to R2 Side, so your info is consistent at all times, but host has RW access only in R1 and RO to R2.
3 - In such special conditions, you can achieve to have both sides RW at the same time, for instance when you perform a "split" operation over the link or over a special pair of devices, in that particular case from Host Standpoint R1 and R2 will be seen as RW, but you will be able to copy only in one site, and it will accumulate owe tracks to the other side, and both sides will be sychronized when a establish operation takes place,
Does it make sense to you?
Regards, and I hope it helps you
Pablo
Fenglin1
2.1K Posts
0
September 27th, 2013 00:00
Yes, -nocopy option will enabler the CopyOnAccess. After activating the copy session, only those tracks that have been written to the source or written/read from the target, will be copied to the target device. A full data copy to the target device will not occur unless all of the device tracks are accessed or written to while participating in the active session.
Kleanthis76
26 Posts
0
September 30th, 2013 23:00
First of all I agree with Fenglin Li and your clone will be immediately available for host I/O once you activate it.
LBM99
119 Posts
0
October 1st, 2013 06:00
You can also change the mode of -nocopy from CopyOnAccess to CopyOnWrite, which means a read to a track on target will not cause the track to be copied from the source. The data that will be presented to the host is the data that is on the source since that is the point-in-time. The copy will only take place after a write operation to source or target.
You can change the mode of -nocopy by editing the options file in the EMC\SYMAPI\config directly. It needs to be changed before the session is created, cannot change the type of -nocopy on existing sessions. The entry is below. To enable CopyOnWrite simply uncomment that line.
#------------------------------------------------------------------------------
#
# Parameter: SYMAPI_CLONE_COPY_ON_WRITE
#
# Allowed values: ENABLE | DISABLE
#
# Default value: DISABLE
#
# Description: Sets clone nocopy mode to copy-on-write (ENABLE) or
# copy-on-access (DISABLE).
#
#SYMAPI_CLONE_COPY_ON_WRITE = ENABLE
#------------------------------------------------------------------------------
satheeshnp
11 Posts
0
October 1st, 2013 07:00
After creating the relationship , u can activate the same .
Once it gives message that " Activation completed successfully" , You can map that lun to host for I/O.
Time finder Clone is a point in Time copy . While in Activation mode , the Target Lun will be Not ready . After Activation , the LUN will become ready and you can access via host.
Regards
Satheesh
narohna
8 Posts
0
November 13th, 2013 07:00
Thank you for your replies.
narohna
8 Posts
0
November 14th, 2013 01:00
Thank you Pablo, that makes sense. Its easy to understand with an example :-)