LVM Label error

[mdraja]

Hi

Our storage team dynamically copies a database (on SAN) from the hdisks assigned to the volume group of the database

However, everytime a backup is done of the hdisks, a /fs is appended to the lv label



root@servername# lslv db01_data01

LOGICAL VOLUME: db01_data01 VOLUME GROUP: vg_db01

LV IDENTIFIER: 00c8192e00004c0000000112d8828cea.1 PERMISSION: read/write

VG STATE: active/complete LV STATE: opened/syncd

TYPE: jfs2 WRITE VERIFY: off

MAX LPs: 10059 PP SIZE: 64 megabyte(s)

COPIES: 1 SCHED POLICY: parallel

LPs: 10059 PPs: 10059

STALE PPs: 0 BB POLICY: relocatable

INTER-POLICY: minimum RELOCATABLE: yes

INTRA-POLICY: middle UPPER BOUND: 1024

MOUNT POINT: /pctmsora01/001/oradata/PCTMS01 LABEL: /fs/fs/fs/fs/fs/fs/pctmsora01/001/oradata/PCTMS01

MIRROR WRITE CONSISTENCY: on/ACTIVE

EACH LP COPY ON A SEPARATE PV ?: yes

Serialize IO ?: NO

DEVICESUBTYPE : DS_LVZ



sure enough, when you look in /fs there is another fs/fs etc

Anyone ever seen this?

Any help greatly appreciated

 

[khalidaaa]

I'm still waiting for a solution to this interesting problem!

Did you find out why?

Regards,
Khalid

 

[RodKnowlton]

The Label field is just that, a label. When you made the JFS2 filesystem on the lv, it would have put the mount point in the label.

Since the Mount Point and Label fields have different values, I'd suspect that both the label and the directory structure are being created by the procedure or script being used for the dynamic copying.

I'm guessing the filesystem is actually mounted on /pctmsora01/001/oradata/PCTMS01, or this would have been a more urgent problem.

- Rod


IBM Certified Advanced Technical Expert pSeries and AIX 5L
CompTIA Linux+
CompTIA Security+

A Simple Code for Posting on the Web

 

[p5wizard]

Most splitcopy/instantcopy/timefinder/whatever... backup solutions allow/require a "prefix" when importvg-ing a copied set of LUNs so as not to conflict with the original VG.

I'm guessing this copied VG is being imported, redefined, ... for a number of times, and hence the LV label info being prefixed again and again.

Although if every time the original VG is being FlashCopied, the LV labels should start from their original value, so this still looks strange to me.

Either way, as Rod said, it is just a label, so you can ignore it. If need be, you can change the labels to their original value (chlv -L newlabel lvname) and it shouldn't be too hard to script this to automate the task.

Does this make sense? If not, kindly disregard.


HTH,

p5wizard

 

[RodKnowlton]

A preemptive resetting of the label might not be a bad idea.

The label is limited to 127 characters so, depending on the failure mode (I don't know if it's truncation or erroring out), going over that limit might break your process.


- Rod


IBM Certified Advanced Technical Expert pSeries and AIX 5L
CompTIA Linux+
CompTIA Security+

A Simple Code for Posting on the Web