29.01.2013 Views

GPFS: Administration and Programming Reference - IRA Home

GPFS: Administration and Programming Reference - IRA Home

GPFS: Administration and Programming Reference - IRA Home

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

The output shows the device name of the file system <strong>and</strong> the file system manager’s node number <strong>and</strong><br />

name.<br />

file system manager node [from 19.134.68.69 (k164n05)]<br />

---------------- ------------------<br />

fs1 19.134.68.70 (k164n06)<br />

See the “mmlsmgr Comm<strong>and</strong>” on page 212 for complete usage information.<br />

You can change the file system manager node for an individual file system by issuing the mmchmgr<br />

comm<strong>and</strong>. For example, to change the file system manager node for the file system fs1 to k145n32,<br />

enter:<br />

mmchmgr fs1 k145n32<br />

The output shows the file system manager’s node number <strong>and</strong> name, in parenthesis, as recorded in the<br />

<strong>GPFS</strong> cluster data:<br />

<strong>GPFS</strong>: 6027-628 Sending migrate request to current manager node 19.134.68.69 (k145n30).<br />

<strong>GPFS</strong>: 6027-629 Node 19.134.68.69 (k145n30) resigned as manager for fs1.<br />

<strong>GPFS</strong>: 6027-630 Node 19.134.68.70 (k145n32) appointed as manager for fs1.<br />

See the “mmchmgr Comm<strong>and</strong>” on page 107 for complete usage information.<br />

Starting <strong>and</strong> stopping <strong>GPFS</strong><br />

For new <strong>GPFS</strong> clusters, see Steps to establishing <strong>and</strong> starting your <strong>GPFS</strong> cluster in: General Parallel File<br />

System: Concepts, Planning, <strong>and</strong> Installation Guide.<br />

For existing <strong>GPFS</strong> clusters, before starting <strong>GPFS</strong>, ensure that you have:<br />

1. Verified the installation of all prerequisite software.<br />

2. Compiled the GPL layer, if Linux is being used.<br />

3. Properly configured <strong>and</strong> tuned your system for use by <strong>GPFS</strong>. This should be done prior to starting<br />

<strong>GPFS</strong>.<br />

For details see the General Parallel File System: Concepts, Planning, <strong>and</strong> Installation Guide.<br />

Start the daemons on all of the nodes in the cluster by issuing the mmstartup -a comm<strong>and</strong>:<br />

mmstartup -a<br />

The output is similar to this:<br />

Tue Aug 24 15:54:56 edt 2004: 6027-1642 mmstartup: Starting <strong>GPFS</strong> ...<br />

Check the messages recorded in /var/adm/ras/mmfs.log.latest on one node for verification. Look for<br />

messages similar to this:<br />

mmfsd initializing ...<br />

<strong>GPFS</strong>: 6027-300 mmfsd ready<br />

This indicates that quorum has been formed <strong>and</strong> this node has successfully joined the cluster, <strong>and</strong> is now<br />

ready to mount file systems.<br />

If <strong>GPFS</strong> does not start, see the General Parallel File System: Concepts, Planning, <strong>and</strong> Installation Guide<br />

<strong>and</strong> search for the <strong>GPFS</strong> daemon will not come up.<br />

See the “mmstartup Comm<strong>and</strong>” on page 270 for complete usage information.<br />

Chapter 2. Managing your <strong>GPFS</strong> cluster 11

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!