We are running a 300GB Progress database on our AIX 5.2 P670.
The DB started as one volume group (b_dbvg) with three file systems /prod/db1, /prod/db2, /prod/db3. As the database grew we added a new VG called b_dbexvg to expand the database. It is made up of filesystems /prod/db4, /prod/db5 and /prod/db6. Each vg is made up of about twenty five 5.5 gb luns. Because we are using EMC DMX1000 with Power Path they show up as disks hdiskpower1 thru hdiskpower25. The second vg has hdiskpower26 thru hdiskpower51. Each AIX logical volume and file systems has several of these hdiskpower devices.
I'm looking for a fast way to make a snapshot of the DB. I was thinking of using mirrorvg and splitvg. Can I use extendvg to create 5 new larger luns for each vg and then mirrorvg the 25 luns to the new 5 larger luns? I can only have a max of 32 disks in my jfs2 vg. I would then use splitvg to split off the 5 new luns to a different vg called snapdbvg.
splitvg -y snapdbvg -c 2 b_dbvg.
I was hoping that I could mount the file systems for the new snapdbvg and have both dbs up at the same time.
I'm guessing that I would somehow need to change the names for the new snapdbvg file systems so they would not use duplicate names. Does this make sense?
Another option would be to gain more info on the DMX array and use something like EMC'c SRDF or Time Finder to clone the database. At this point our company has very little knowledge of the DMX although we are very familiar with the EMC Clariion CX models.
The DB started as one volume group (b_dbvg) with three file systems /prod/db1, /prod/db2, /prod/db3. As the database grew we added a new VG called b_dbexvg to expand the database. It is made up of filesystems /prod/db4, /prod/db5 and /prod/db6. Each vg is made up of about twenty five 5.5 gb luns. Because we are using EMC DMX1000 with Power Path they show up as disks hdiskpower1 thru hdiskpower25. The second vg has hdiskpower26 thru hdiskpower51. Each AIX logical volume and file systems has several of these hdiskpower devices.
I'm looking for a fast way to make a snapshot of the DB. I was thinking of using mirrorvg and splitvg. Can I use extendvg to create 5 new larger luns for each vg and then mirrorvg the 25 luns to the new 5 larger luns? I can only have a max of 32 disks in my jfs2 vg. I would then use splitvg to split off the 5 new luns to a different vg called snapdbvg.
splitvg -y snapdbvg -c 2 b_dbvg.
I was hoping that I could mount the file systems for the new snapdbvg and have both dbs up at the same time.
I'm guessing that I would somehow need to change the names for the new snapdbvg file systems so they would not use duplicate names. Does this make sense?
Another option would be to gain more info on the DMX array and use something like EMC'c SRDF or Time Finder to clone the database. At this point our company has very little knowledge of the DMX although we are very familiar with the EMC Clariion CX models.