Disk Problem?

[Nobody's Hero]

Running 11.11 on an Nclass. Storage is a VA7100.

When I use SAM to look at my disk structure, it tells me that /u02 is 69 Gbyte is size.

When I issue a bdf, it tells me that /u02 is 57 Gbytes is size.
Here is what I am planning to do to correct the problem. Please Advise.

Also, will I have to restore my oracle data on this volume after I do this or will the data still reside there, and the import will tie everything together?

1. unmount all the file systems resides on /dev/vg01
i.e. umount /u02
2. vgchange -a n /dev/vg01 -> deactivate vg01
3. cp /etc/lvmtab /etc/lvmtab.backup
4. vgexport -m /tmp/vg01.map -s /dev/vg01
5. mkdir /dev/vg01
6. mknod /dev/vg01/group c 64 0x010000
7. vgimport -m /tmp/vg01.map -s /dev/vg01
8. vgdisplay /dev/vg01

[Cheryl Griffin]

bdf is showing the filesystem size. Maybe you need to lvextend and fsadm/extendfs for the space to show up in bdf.

[Torsten.]

What does armdsp say about the LUN size?

[Nobody's Hero]

I never used armdsp. What kind of syntax should I use?

[Torsten.]

armdsp -a

is showing all information about your VA. Run from the host where Commandview SDM software is installed.

[Mridul Shrivastava]

Find array id from Alias of armdsp -i

[Nobody's Hero]

I will repost output of command in 15 minutes.

[Nobody's Hero]

Also, If I do my choice to resole this, will I have to restore all of my data or will it still be there?

[Torsten.]

Cheryls posting sounds very likely.

Are you sure your LVOL is using the whole LUN? Have a look at your LVM configuration.

Is the LUN the only physical volume in the VG?

Is the LVOL using all space available in the VG?

Did you extend the LVOL without extending the filesystem?

[Nobody's Hero]

I'll check that..

Let me state this. Before I check that.

Act PV and cur PV show different numbers. Shouldnt they be the same?

[Torsten.]

"Act PV and cur PV show different numbers"

Some more explicit, please.

[Nobody's Hero]

I dont know what else to tell you except, when I do a vgdisplay on /dev/vg01

Cur PV = 13
and Act PV = 10

I did delete 3 luns a while back that were in vg01, maybe this is where things went goofy. THe Act and Cur PV should = 10.

[Deoncia Grayson_1]

Your active and current physical volumes in a volume group should be the same. There should not be any discrepencies between the two.

[Nobody's Hero]

OK thanks everyone, but could someone please try to answer this?

If I do my steps above in my original post, will my data still be there or will I have to restore it?

Thanks.

[Patrick Wallek]

By just exporting and importing, your data SHOULD still be there.

That being said, make SURE you have a backup or two. If you have it, you won't need it.

[Deoncia Grayson_1]

In theory since you're not removing any physical data and you're not experiencing any disk read errors, then your data should be there once you reactive your volume group and mount it back, since you're not replacing any disk. I don't see why your data should disappear but just to be safe, make sure you have a good backup in case you need to restore. Also I would do a vgscan to repopulate the lvmtab and make sure your vgdisplay come back clean.

[Torsten.]

Dear Robert,

we came from

filesystem size != LVOL size

to

current PV != actual PV

Since the crystal ball is in maintenance we could try to use "The Force" (to quote some well-known persons) to solve your real problem, but I'm not experienced and it would be more helpful to provide some more information.

Why not post something like "strings /etc/lvmtab", "vgdisplay -v ", "ioscan -fn" and the mentioned armdsp to give us a little chance?

Have fun ;-)