Well primus does not mention about transIdle state and only talks about recovering from suspended state, So i will append one more step: If transIdle is enabled on the RA group then we dont need to perform the above steps.
Wait until the invalid track count is under 30,000 tracks
It depends on the overall number of tracks you are syncing... 30.000 tracks can be too many (in case you are syncing a single 18414cyl volume) or too few (in case you are syncing 20+Tb of data)... As a general rule of thumb, I'd switch from ACP to ASYNC when the number of tracks start to float around a value (when synced tracks equals newly written tracks from hosts).
Do u have composite groups? is it having MSC? how many invalid tracks are accumulated? do u have golden copies? are all the devices in precopy state? you cannot just issues the coomands to bring the SRDF back....there are lot of thinggs to be considered...
Amita2
93 Posts
1447
0
Posted February 3rd, 2012 02:00
You can use the following SYMCLI commands :
1. symrdf -g dgname disable (Disables consistency protection for SRDF/A pairs in device group)
2. symrdf -g dgname set mode acp_disk (This will turn on the adaptive copy write pending mode for the device group)
3. symrdf -g dgname est (This initiates an establish for all the SRDF pairs in the device group)
4. symrdf -g dgname query (To query a device Group)
Wait until the invalid track count is under 30,000 tracks
5. symrdf -g dgname set mode async (Change the Copy mode to Asyncronous)
6. symrdf -g dgname enable (Enables consistency protection for SRDF/A pairs in device group)
You can also refer to Primus article: emc174647 for the issue.