Forum Discussion

Marius_Gordon's avatar
Marius_Gordon
Level 4
11 years ago
Solved

VVR

Hi

I am new to VVR and need assistance please.

Replication works fine until I create the “replication service group” in VCS.

VCS initiates a disk group deport on secondary node 2 and then replication stops.

If you freeze the “replication service group” the disk group on the secondary node stays imported and replication works.

I obviously missing something. Shouldnt the disk groups be in an imported state on all node for VVR replication to work?

 

Thanks

Marius

business continuity
Replicator Option
Training
Windows Server (2003-2008)
  • mikebounds's avatar
    mikebounds
    11 years ago

    It appears you are trying to replicate between your 2 nodes in your cluster SRV003457 and SRV003458 - this is called a Replicated Data Cluster (RDC) which requires that you set up the replication service group as a hybrid service group.  If you are using SQL, there are instructions on how to do this in the SFW_HA_DR_SQL2008_Solutions_6.0.1.pdf and SFW_HA_DR_SQL2012_Solutions_6.0.1.pdf guides or equivalent for your versions of SFWHA and SQL.

    Your  PRDdatdg-RVG should look like:

     

    group PRDdatdg-RVG (
        SystemList = { SRV003457 = 0, SRV003458 = 1 }
        Parallel = 2
        AutoStartList = { SRV003457, SRV003458 }
        SystemZones = { SRV003457 = 0, SRV003458 = 1 }
        )

        IP PRDdatdg-RVG-IP (
            Address@SRV003457 = IP_for_SRV003457
            Address@SRV003458 = IP_for_SRV003458
            SubNetMask = "255.255.255.0"
            MACAddress @SRV003457 = 00-0A-F7-20-29-0A
            MACAddress @SRV003458 = 00-0A-F7-20-29-14
            )

        NIC PRDdatdg-RVG-NIC (
            MACAddress @SRV003457 = 00-0A-F7-20-29-0A
            MACAddress @SRV003458 = 00-0A-F7-20-29-14
            )

        VMDg CLS10-SQL-PROD-SG-VMDg (
            DiskGroupName @SRV003457 = PRDdatdg
            DiskGroupName @SRV003458 = PRDdatdg
            )

        VvrRvg PRDdatdg-RVG-VvrRvg (
            RVG = PRDdatdg-RDS
            VMDgResName = CLS10-SQL-PROD-SG-VMDg
            IPResName = PRDdatdg-RVG-IP
            )

        PRDdatdg-RVG-VvrRvg requires CLS10-SQL-PROD-SG-VMDg
        PRDdatdg-RVG-VvrRvg requires PRDdatdg-RVG-IP
        PRDdatdg-RVG-IP requires PRDdatdg-RVG-NIC

     

    You do not have "Parallel = 2" which means your PRDdatdg-RVG is a failover group only allowed online on one node and therefore you get errors:
    2014/04/28 08:09:26 VCS ERROR V-16-1-10214 Concurrency Violation:CurrentCount increased above 1 for failover group PRD_MSDTC-RVG
    2014/04/28 08:09:26 VCS WARNING V-16-6-15034 (SRV003458) violation:Offlining group PRD_MSDTC-RVG on system SRV003458

     

    If you stop VCS ("hastop -all" from one node or "net stop had" on both nodes) and then change main.cf as above and then restart VCS ("hastart" or "net start had" from both nodes), then it should work, but if you are using SQL, then usually you should be using the SQL agents set up using the SQL wizards as in guides mentioned above which will also configure RDC for you.

    Ideally you should be using virtual IPs for replication (these should be separate from application IPS) as this means you can add more nodes to your cluster in the furture (so for instance you could add node SRV003456 and then you could replicate from SRV003456 to SRV003458 or replicate from SRV003457 to SRV003458 so the virtual replication IP would fail between SRV003456 and SRV003457) , but if you only ever have 2 nodes, you don't technically need to use virtual IPs and if this is the case then you should remove the IP (PRDdatdg-RVG-IP) resource.

    Mike

3 Replies

Sort By
  • Gaurav_S's avatar
    Gaurav_S
    Moderator
    11 years ago

    Hello,

    Yes, diskgroups will be in imported state on the remote site for replication to work. VVR works at block level replication so volumes have to be visible on remote site for replication to work. Volumes will appear only if diskgroup is imported. This is also proved once you keep group frozen, replication works. The problem here is why VCS deports the diskgroup ?

    I  am wondering if your systemlist is correct on remote site ? can you attach main.cf for review & also attach engine_A.log which carries entries of VCS taking diskgroup offline.

     

    G

  • mikebounds's avatar
    mikebounds
    Level 6
    11 years ago

    It appears you are trying to replicate between your 2 nodes in your cluster SRV003457 and SRV003458 - this is called a Replicated Data Cluster (RDC) which requires that you set up the replication service group as a hybrid service group.  If you are using SQL, there are instructions on how to do this in the SFW_HA_DR_SQL2008_Solutions_6.0.1.pdf and SFW_HA_DR_SQL2012_Solutions_6.0.1.pdf guides or equivalent for your versions of SFWHA and SQL.

    Your  PRDdatdg-RVG should look like:

     

    group PRDdatdg-RVG (
        SystemList = { SRV003457 = 0, SRV003458 = 1 }
        Parallel = 2
        AutoStartList = { SRV003457, SRV003458 }
        SystemZones = { SRV003457 = 0, SRV003458 = 1 }
        )

        IP PRDdatdg-RVG-IP (
            Address@SRV003457 = IP_for_SRV003457
            Address@SRV003458 = IP_for_SRV003458
            SubNetMask = "255.255.255.0"
            MACAddress @SRV003457 = 00-0A-F7-20-29-0A
            MACAddress @SRV003458 = 00-0A-F7-20-29-14
            )

        NIC PRDdatdg-RVG-NIC (
            MACAddress @SRV003457 = 00-0A-F7-20-29-0A
            MACAddress @SRV003458 = 00-0A-F7-20-29-14
            )

        VMDg CLS10-SQL-PROD-SG-VMDg (
            DiskGroupName @SRV003457 = PRDdatdg
            DiskGroupName @SRV003458 = PRDdatdg
            )

        VvrRvg PRDdatdg-RVG-VvrRvg (
            RVG = PRDdatdg-RDS
            VMDgResName = CLS10-SQL-PROD-SG-VMDg
            IPResName = PRDdatdg-RVG-IP
            )

        PRDdatdg-RVG-VvrRvg requires CLS10-SQL-PROD-SG-VMDg
        PRDdatdg-RVG-VvrRvg requires PRDdatdg-RVG-IP
        PRDdatdg-RVG-IP requires PRDdatdg-RVG-NIC

     

    You do not have "Parallel = 2" which means your PRDdatdg-RVG is a failover group only allowed online on one node and therefore you get errors:
    2014/04/28 08:09:26 VCS ERROR V-16-1-10214 Concurrency Violation:CurrentCount increased above 1 for failover group PRD_MSDTC-RVG
    2014/04/28 08:09:26 VCS WARNING V-16-6-15034 (SRV003458) violation:Offlining group PRD_MSDTC-RVG on system SRV003458

     

    If you stop VCS ("hastop -all" from one node or "net stop had" on both nodes) and then change main.cf as above and then restart VCS ("hastart" or "net start had" from both nodes), then it should work, but if you are using SQL, then usually you should be using the SQL agents set up using the SQL wizards as in guides mentioned above which will also configure RDC for you.

    Ideally you should be using virtual IPs for replication (these should be separate from application IPS) as this means you can add more nodes to your cluster in the furture (so for instance you could add node SRV003456 and then you could replicate from SRV003456 to SRV003458 or replicate from SRV003457 to SRV003458 so the virtual replication IP would fail between SRV003456 and SRV003457) , but if you only ever have 2 nodes, you don't technically need to use virtual IPs and if this is the case then you should remove the IP (PRDdatdg-RVG-IP) resource.

    Mike