Conditions for delta resync operations
When you perform the delta resync, it is required that the the 3DC multi-target configuration using the delta resync is established by mapping remote command devices to the each site and assigning the mirror IDs to the remote command devices. For the delta resync to finish successfully, you must satisfy the following conditions for performing delta resync:
If a journal necessary for the delta resync does not exist, you cannot perform delta resync.
The status of the UR delta resync pair's P-VOL is HOLD.
TC volumes are used as UR P-VOL.
Required pair and volume statuses:
When sharing TC P-VOLs, the UR pair status must be PAIR.
When sharing TC S-VOLs, the UR pair status must be SSWS. Check this status after performing the CCI horctakeover command.
When UR S-VOL is using two mirror IDs and a mirror ID's S-VOL pair status is HOLD, and the other mirror ID's S-VOL pair status is PAIR, PSUS, or PSUE.
In a 3DC multi-target configuration, the number of the regular UR S-VOL and the UR delta resync S-VOL is the same.
The paths between the UR delta resync pair sites must be valid.
If even one of the above conditions is not satisfied, an error occurs and delta resync will fail. In the following cases, the journal necessary for delta resync does not exist and delta resync will not be performed.
After creating UR delta resync pair, UR delta resync P-VOL is updated without updating the production volume shared between UR P-VOL and TC P-VOL after creating UR delta resync pair.
S-VOL is updated after UR pair split.
After UR pair split, journal data exceeds 70% of the journal volume capacity in the TC secondary site a result of updating P-VOL.
As a result of updating UR delta resync P-VOL, the journal data exceeds 70% of the journal volume capacity at the UR delta resync primary site.
When a UR pair has not been suspended and resynchronized for a long time, the data in the restore journal might exceed 70% of capacity. If this happens, old journal data is automatically deleted. In this case, the P-VOL and S-VOL are not synchronized completely by just copying the data, and delta resync will fail. In case of delta resync failure, resynchronize the UR pair.
After creating the UR delta resync pair, if the production volume at the primary site is not updated (including the cases when the production volume is not updated after failover or failback), there will be no journal data in the TC secondary site.
Even if there is no journal data for delta resync in the TC secondary site, you can perform delta resync, but the following two conditions must be met:
- No data update at all from the host to the primary site, or data update is stopped.
- The data in the TC secondary site and the UR secondary site is the same. You must resync all of the TC and UR pairs that belong to the target journal in the primary site and change the pair status to PAIR.
In addition, delta resync might not be performed in the following cases because the necessary journal data might get lost:
Cache memory or control memory is added or removed by the maintenance operation.
The storage system is powered off by the maintenance operation.
TC pair or UR pair is split and then resynchronized.
The UR pairs in the primary system and the secondary system are resynchronized, and then the P-VOL is updated.
Disaster or failure occurred in the TC secondary site, and access to the journal volume is impossible.
Retry operation is caused because the P-VOL update is delayed.
TC S-VOL update is delayed.
UR delta resync pair status is changed to HLDE.
In this case, even if the status is changed from HLDE to HOLD, you cannot copy only the differential data by the later delta resync operation. However, you can copy the entire P-VOL data to the S-VOL.
To accumulate journal data after correcting the above conditions, update the primary site data from the host. Even if the status meets all the conditions to perform delta resync without journal data as described in the above tip, the pair status will be changed to HOLDING regardless of the journal existence once the pair status falls in the possibility of no journal status. If the access to the remote command device in the TC secondary site is impossible by failure or disaster, the pair status will not be changed from HOLDING to HOLD.