tru64 lsm

[Arnulf Gercke]

Hello everybody,

where can i download technical documentation of the tru64 lsm, especially interfaces (cli/gui?), commands (volprint,...), examples and howtos (replacing a mirrored disk, ...).

[Venkatesh BL · Accepted solution]

For V5.1B, you can check out: http://h30097.www3.hp.com/docs/base_doc/DOCUMENTATION/V51B_HTML/ARH9BDTE/TITLE.HTM

The General V5.1B doc list is available at:
http://h30097.www3.hp.com/docs/pub_page/V51B_DOCS/V51B_DOCLIST.HTM

[Ross Minkov]

Make sure you read the right lvm manual for your server. Some things in lvm changed between Tru64 UNIX versions. Especially from V4 to V5.

What version of Tru64 UNIX do you have?

-Ross

[Gary Hansford]

To give you a head start, here is the command sequence I always use in replacing a LSM mirrored disk : -

1. Remove the failed disk from the disk group retaining the media records thus

# voldg â g mydg â k rmdisk media_name

2. Remove the disk access record

# voldisk rm disk_access_name

3. Ensure you have replaced the physical disk !

4. Restore the original disklabel or initialize a new one for the replaced disk.

#disklabel â wr dsk?? disktype

5. Initialise the new disk for use with LSM

if you have the original disk label: -
# voldisk â f init dsk?? offset=16

if you did not have the disk label: -
# voldisksetup â i dsk?? nconfig=? nlogs=? privlength=??
e.g.
# voldisksetup â i dsk? nconfig=2 nlogs=2 privlength=1024

6. Associate the new disk to a disk media record/disk group

# voldg â g mydg â k adddisk media_name=disk_access_name

7. Recover(synchronise) the plexes in the diskgroup

# volrecover â g mydg -sb

Enjoy

Gary

[Gary Hansford]

Weird... most of the minus signs got replaced with â  !!!...

-----------------------------------------
Here is the readable version !!!!!
-----------------------------------------

To give you a head start, here is the command sequence I always use in replacing a LSM mirrored disk : -

1. Remove the failed disk from the disk group retaining the media records thus

# voldg -g mydg -k rmdisk media_name

2. Remove the disk access record

# voldisk rm disk_access_name

3. Ensure you have replaced the physical disk !

4. Restore the original disklabel or initialize a new one for the replaced disk.

#disklabel -wr dsk?? disktype

5. Initialise the new disk for use with LSM

if you have the original disk label: -
# voldisk -f init dsk?? offset=16

if you did not have the disk label: -
# voldisksetup -i dsk?? nconfig=? nlogs=? privlength=??
e.g.
# voldisksetup -i dsk? nconfig=2 nlogs=2 privlength=1024

6. Associate the new disk to a disk media record/disk group

# voldg -g mydg -k adddisk media_name=disk_access_name

7. Recover(synchronise) the plexes in the diskgroup

# volrecover -g mydg -sb

Enjoy

Gary

[Arnulf Gercke]

@BL Venkatesh:
Seems to be a nice doku!

@Ross Minkov:
Tru64 is 5.1 (don't know whether it's a or b)

@Gary Hansford:
Thank you for these details.


Is there something in Tru64 like HP-UX OnlineJFS for easy expanding of mounted filesystems?

[Hein van den Heuvel]

Yes, the AdvFS file system has an 'addvol' option to transparently grow 'domains'.

Hein.

[Alexey Borchev]

>> Is there something in Tru64 like HP-UX OnlineJFS for easy expanding of mounted filesystems?

- Yes, if your storage is capable of on-line LUN expancion, the ADVFS can expand filedomain reziding on this LUN.

See man mount, around -extend option.

Unfortunately, AdvFs advansed utilities license required. And for 'addvol' too.

[Arnulf Gercke]

Thank you so far.

Now the concrete situation:

output volprint:


Disk group: rootdg

TY NAME ASSOC KSTATE LENGTH PLOFFS STATE TUTIL0 PUTIL0
dg rootdg rootdg - - - - - -

dm dsk0f dsk0f - 0 - - - -
dm dsk0g-AdvFS dsk0g - 8192000 - - - -
dm dsk0h dsk0h - 23767576 - - - -
dm dsk1f dsk1f - 0 - - - -
dm dsk1g-AdvFS dsk1g - 8192000 - - - -
dm dsk1h dsk1h - 23767576 - - - -
dm dsk5 dsk5 - 71127888 - - - -
dm dsk6 dsk6 - 71127888 - - - -
dm root01 dsk0a - 524288 - - - -
dm root02 dsk1a - 524288 - - - -
dm swap01 dsk0b - 3073024 - - - -
dm swap02 dsk1b - 3073024 - - - -

v rootvol root ENABLED 524288 - ACTIVE - -
pl rootvol-01 rootvol ENABLED 524288 - ACTIVE - -
sd root01-01p rootvol-01 ENABLED 16 0 - - -
sd root01-01 rootvol-01 ENABLED 524272 16 - - -
pl rootvol-02 rootvol ENABLED 524288 - ACTIVE - -
sd root02-02p rootvol-02 ENABLED 16 0 - - -
sd root02-02 rootvol-02 ENABLED 524272 16 - - -

v swapvol swap ENABLED 3073024 - ACTIVE - -
pl swapvol-01 swapvol ENABLED 3073024 - ACTIVE - -
sd swap01-01 swapvol-01 ENABLED 3073024 0 - - -
pl swapvol-02 swapvol ENABLED 3073024 - ACTIVE - -
sd swap02-02 swapvol-02 ENABLED 3073024 0 - - -

v vol-dsk0g fsgen ENABLED 8192000 - ACTIVE - -
pl vol-dsk0g-01 vol-dsk0g ENABLED 8192000 - ACTIVE - -
sd dsk0g-01 vol-dsk0g-01 ENABLED 8192000 0 - - -
pl vol-dsk0g-02 vol-dsk0g ENABLED 8192000 - ACTIVE - -
sd dsk1g-01 vol-dsk0g-02 ENABLED 8192000 0 - - -

v vol-dsk0h fsgen ENABLED 23767576 - ACTIVE - -
pl vol-dsk0h-01 vol-dsk0h ENABLED 23767576 - ACTIVE - -
sd dsk0h-01 vol-dsk0h-01 ENABLED 23767576 0 - - -
pl vol-dsk0h-02 vol-dsk0h ENABLED 23767576 - ACTIVE - -
sd dsk1h-01 vol-dsk0h-02 ENABLED 23767576 0 - - -

v vol-dsk5 fsgen ENABLED 71124992 - ACTIVE - -
pl vol-dsk5-01 vol-dsk5 ENABLED 71124992 - ACTIVE - -
sd dsk5-01 vol-dsk5-01 ENABLED 71124992 0 - - -
pl vol-dsk5-02 vol-dsk5 ENABLED 71124992 - ACTIVE - -
sd dsk6-01 vol-dsk5-02 ENABLED 71124992 0 - - -
pl vol-dsk5-03 vol-dsk5 ENABLED LOGONLY - ACTIVE - -
sd dsk5-02 vol-dsk5-03 ENABLED 1170 LOG - - -

Disk group: dg1

TY NAME ASSOC KSTATE LENGTH PLOFFS STATE TUTIL0 PUTIL0
dg dg1 dg1 - - - - - -

dm dg101 dsk7 - 35560968 - - - -
dm dg102 dsk3 - 35560968 - - - -

v vol-baan fsgen ENABLED 35559424 - ACTIVE - -
pl vol-baan-01 vol-baan ENABLED 35559424 - ACTIVE - -
sd dg101-01 vol-baan-01 ENABLED 35559424 0 - - -
pl vol-baan-02 vol-baan ENABLED 35559424 - ACTIVE - -
sd dg102-01 vol-baan-02 ENABLED 35559424 0 - - -


Failures have been detected by the Logical Storage Manager:
failed disks:
dsk1h
failed plexes:
rootvol-02
swapvol-02
vol-dsk0g-02
vol-dsk0h-02


This is, what I would do:

1. # volplex -g rootdg -o rm dis rootvol-02 swapvol-02 vol-dsk0g-02 vol-dsk0h-02
2. # voldg -g rootdg rmdisk dk1a dsk1b dsk1f dsk1g-AdvFS dsk1h
3. # voldisk rm dsk1a dsk1b dsk1f dsk1g dsk1h
4. === physical replacement ==
5. # hwmgr scan scsi
6. # hwmgr show scsi => New unused disk dskX
7. # dsfmgr -e dskX dsk1
8. # disklabel -rw dskX
9. # volrootmir -a dskX

Now some questions:

a) Seems that dsk1f is a dummy (length=0).
However, i think it must be removed in 2, ok?
b) In 2. i have to use the disk_media_names.
Is dsk1g-AdvFS correct?
c) Is there a nice command to find out
the physical disk for dsk1?
d) I feel that there will be problems,
if the replacement disk isn't new
and was used in another LSM system.
Is a previous format neccessary?
e) What about 7. ?
The new disk will not be dsk1?

Any comments are greately appreciated!

[Arnulf Gercke]

i performed the replacing of the mirrored boot disk, but i had one problem.
i feel it is appropriate to open a new thread for this.