Unsolved
This post is more than 5 years old
9 Posts
0
2941
Are my clone fracture/mount/unmount/resynch cmds correct?
Do my commands below look correct? Anything incorrect or missing? THANKS.
--SW
/*****
pseudo-code
Using: EMC Clariion CX4-120, Flare 29; SnapView Clone; naviseccli & admsnap.
1 Source Host Server (Windows 2003; Sybase Advantage 8)
1 Target Host Server (Windows 2003; CommVault Simpana 8 Media Server)
1 Source Host LUN
1 Clone Target of Source LUN
Summary: Fully synched Cloned LUN is fractured, then mounted to Target Host Server, which
then FileSystem BUs the Clone. After BU, Clone is then unmounted and re-synched with Source.
*****/
/***
1. (Source Host Server) -
Verify that Clone is "normal", "consistent", "100" %synchronized
Notes
[-CloneCondition] (returns "normal")
[-cloneid id] (insert id; returns id)
[-CloneState] (returns "consistent")
[-PercentSynced] (run before fracture; returns "100")
IF CloneCondition <> "normal", then escalate to Support. Process must abort and
cannot continue.
IF PercentSynced <> "100", then wait till value = "100". Clone must be 100% synched
to continue.
IF CloneState <> "consistent", then wait till value = "consistent". Clone must be in
a "consistent" state to continue. If consistent state not acheived within an as-yet
unspecified period, then escalate to Support. Process must abort and cannot continue.
***/
naviseccli -address 192.168.1.100 Clone -CloneCondition -cloneid -CloneState -PercentSynced
/***
2. (Source Host Server) - Flush buffers before fracture
Notes
= drive letter of Source LUN
Example: admsnap flush -o E:
***/
admsnap flush -o
/***
3. (Source Host Server) - Administratively fracture
***/
naviseccli -address 192.168.1.100 Clone -fractureclone -cloneid -o
/***
4. (Source Host Server or Target Host Server) -
Assign the clone to a storage group.
You must assign the clone LUN to the Secondary Server's storage
group, which must be other than the storage group that holds
the source LUN.
-hlu hlunumber
Specifies the host LUN number for the snapshot, which is the
unique number that the storage-system server will see.
hlunumber is the virtual LUN number.
-alu alunumber
Specifies the storage-system LUN number, the actual LUN
number internal to the storage system. This is the number
assigned when the LUN was bound. alunumber specifies the
logical unit number. This switch is required along with -hlu
and either the -gname or -uid switch.
-o the CLI prompts for confirmation unless you use the -o
switch.
See EMC Navisphere Command Line Interface (CLI) Reference R29 p.300
***/
naviseccli -address 192.168.1.100 storagegroup -addhlu -gname -hlu -alu -o
/***
5. (Target Host Server) -
Mount the Clone to the BU Host Server
clone_activate -- Scans for new devices and assigns a drive letter
to every new device it finds.
Q: how does the LUN get assigned to the Drive letter that CommVault is
configuered to BU? (i.e. the subclient)
***/
admsnap clone_activate
/***
6. (Target Host Server) -
unMount the Clone from the BU Host Server
Once the clone is mounted, Commvault Filesystem backs up the LUN.
Upon CV BU completion, unmount the LUN
Example using Drive E: -- admsnap clone_deactivate -o E:
***/
admsnap clone_deactivate -o
/***
7. (Target Host Server) -
Remove the clone from the storage group.
***/
naviseccli -address 192.168.1.100 storagegroup -removehlu -gname -hlu -o
/***
8. (Target Host Server) -
Synchronize the clone
Use the CLI command -syncclone for this (refer to clone |
snapview -syncclone on page 80).
NOTE: since we are NOT WRITING to the clone LUN, we do not need to
flush target host server buffers.
***/
naviseccli -address 192.168.1.100 Clone -syncclone -cloneid -o
/***
END - Clone is synchronizing and ready for next cycle.
***/
kelleg
4.5K Posts
0
October 15th, 2010 08:00
I don't believe you need to remove the clone from the storage group - the amdsnap deactivate should remove the drive and flush the data from Windows.
glen
sw7104
9 Posts
0
October 15th, 2010 09:00
I'm onsite w/EMC implementer now and Yes, that is what he recommends.
Am removing those two lines and correcting the Listclone cmd where I left out the actual -listclone cmd! (and clone group name/id).
Will post final tested script when complete.
Thanks.
--SW
kelleg
4.5K Posts
0
October 15th, 2010 10:00
Just found some old notes:
sw7104
9 Posts
0
November 3rd, 2010 08:00
I promised to include my final script. It is genericized to remove IDs and is specific to business requirements. Full script is located at:
https://docs.google.com/leaf?id=0B2_-C0y1VI5kZTI1ZDU0YjctMzljZi00ZmYxLWI3ZjItZmFmMTBkNGM3ZjEw&hl=en&authkey=COzD7b8E
A summary -- it's a VBS that semi-automates a 9-step process that begins with unmounting the clone from the target backup host, processes through its synchronization back to the source, and ends with flushing, fracturing & mounting the clone back to the target host.
The script includes thorough logging and basic exception handling (any returned error aborts the script).
The script can be started (i.e. manually resumed) at any of the 9 steps; default is Step 1.
Here is the generic command set that is embedded in the script, in run order, minus a couple of non-SnapView cmds: