shared vg 4 /var/opt/ignite

[Holger Knoppik]

Hi IUX experts!
Short description: I have 4 L-Class boxes attached to 2 IBM ESS via fiber channel. 2 Clusters (2x2 nodes).
Two servers should become ignite servers (ignite01 = a node of cluster1, ignite02 = a node of cluster2). I want to create a VG that is accessible by the 2 ignite servers. (The VG will contain 2 PVs, prime & mirror, so /var/opt/ignite will reside in the ESS). Is that possible, and if, does it make sense ?
TIA, RGDS
Holger

[Sanjay_6]

Hi Holger,

You cannot have a VG that is part of two clusters. Since your two ignite server belong to different cluster, sharing the same VG among them is not possible.

A Possible configuration would be to NFS mount the file system you want to share from one System to another after exporting it from the first system.

hope this helps.

Regds

[Holger Knoppik]

Sanjay, thanx 4 ur quick response.
I took a look into "vgchange (1M)". Seems like i have to put /var/opt/ignite into the first cluster. What i want to do is to backup all 4 boxes without having different versions (i'm lazy!). The only way to backup all 4 servers is to setup 2 ignite servers. Could u gimme more deatils?
Thanx, Holger

[harry d brown jr]

I have no idea how you handle clusters, but check this out:



I have no idea as to the impact of this (how the kernel on node2 will behave), but so far it seems to behave ok, so I urge you to use this at YOUR OWN RISK!

Scenario:

node1 has vgxx with two lv???s = lvol1 and lvol2
node2 can access the same physical disks as node1 via san, fc???s, whatever


(01) From node1, use the vgexport command to export the volume group to get some definitions, where xx is the volume group you want to share:
# vgexport ???p -m /tmp/vgxx_lvmapfile ???f /tmp/vgxx_dskfile /dev/vgxx

(02) From node1, copy the mapfile and lvfile to node2:
# rcp or ftp to node2 on the /tmp directory

(03) On node2, issue the following command:
# mkdir /dev/vgxx

(04) (node2) Create a control file named group in the directory /dev/vgxx, as in the following:
# mknod /dev/vgxx/group c 64 0xhh0000
The major number is always 64, and the hexadecimal minor number has the form 0xhh0000 where hh must be unique to the volume group you are creating. Use the next hexadecimal number that is available on your system, after the volume groups that are already configured.

(05) (node2) THIS IS VERY IMPORTANT if the device numbers are different (ie, on my system node1 has disk /dev/dsk/ c7t9d0 where node2 sees it as /dev/dsk/c6t9d0) vi the /tmp/vgxx_dskfile and correct the device.

(06) (node2) Use the vgimport command, specifying the map file you copied from the configuration node and the lv file:
# vgimport -v -m /tmp/vgxx_lvmapfile ???f /tmp/vgxx_dskfile /dev/vgxx

(07) (node2) Activate the volume:
# vgchange -a y /dev/vgxx

(08) (node2) Create mount points for lv???s:
# mkdir /rolvs (ro for readonly ??? just for readability sake, name it anything you want)
# mkdir /rolvs/lv01
# mkdir /rolvs/lv02

(09) (node2) MOUNT READ ONLY
mount -F vxfs -o ro /dev/vgxx/lvol1 /rolvs/lv01
mount -F vxfs -o ro /dev/vgxx/lvol2 /rolvs/lv02

Now, if node1 makes any changes to lv01 or lv02, you must umount on node2 and remount to see the changes!!!


live free or die
harry

[harry d brown jr]

No one commented on my SAN readonly vg disks. Damn! No thought provoking ideas?

??

live free or die
harry

[Holger Knoppik]

Hmmm, maybe there's a flaw in my thinkin', i just opened a call for my idea. I'm curious what the cracks will offer...
Thank you for your help. I'll let you know...
RGDS, Holger