HP StorageWorks Scalable File Share Release Notes - Version 2.3

New and changed features in HP SFS Version 2.31–10
1.4.9 Server management commands fail when targeting more than 24 servers
at a time
sfsmgr server management commands, such as boot/shutdown server, scan mac, and so on, may fail
on some servers when targeting more than 24 servers at a time. On very large systems, HP strongly
recommends to split such commands in groups targeting at most 24 servers.
Example: For a 48-node server:
sfs> shutdown server 2-24 wait=0
...
Command has finished:
atlas[2-24] -- <Success>
...
sfs> shutdown server 25-48 wait=0
...
Command has finished:
atlas[25-48] -- <Success>
1.4.10 Need to avoid DHCP conflicts
SFS OSS nodes network boot from the SFS Admin node using DHCP and PXE.
Ensure that no other DHCP server in your organization responds to boot requests from the OSS nodes or it
will interfere with their operation
1.4.11 The SFS Web Server updates sometimes fail
The SFS Web Server publishes server configuration information, for use by SFS clients. This information can
be displayed on clients using the sfsview command. It is used by sfsconfig and sfsmount for
generating the correct Lustre commands.
In rare cases, after changing the SFS server configuration, the updates may fail to be recorded on the Web
server. The symptoms are sfsconfig or sfsmount errors on the client side, using or reporting the wrong
configuration. (The one before the change on the server.) This can be confirmed by searching on the server
in the event log for the message "Failed to update Web server configuration."
To fix it, on the server run the sfsmgr command:
sfs> update webserver
1.4.12 OFED 1.2 does not support RH4U6
When building the client-side components for RH4U6 with OFED, manually build OFED-1.3, since OFED-
1.2 does not support RH4U6. The steps to build a RH4U6 client with OFED can be found later in these
release notes.
1.4.13 What to do in case of large numbers of 'database not responding'
messages
A few cases have occurred where, when the system is slow, sfsmgr commands often display database
not responding messages, and sfsmgr commands fail randomly.
This was the result of a corrupted system event log which in turn caused the admin service to restart
repeatedly in an attempt to fix the problem. During the intervals when the admin service is down, the system
fails to operate normally.
This diagnostic can be verified by looking at the tail of file /var/log/hpls_admin_2.log on the admin server,
and seeing repetitive restarts of hpls_admin, and error messages each time:
Failure of action: status for service: /usr/opt/hpls/bin/hpls_logd