Call failures when lockfs -f is ran of /opt/SMAW

Question

During our update scenario, we initially disable local mirrors for all the local file systems like, /,/opt/SMAW,/var,/export/home.
During this disabling was call failures observed when "lockfs" was executed on the /opt/SMAW file system.
There was no call failure observed during disabling of other local file systems.
As a result of this some of the processes are blocked.
A Perf tool was run on the node to collect information like CPU.Load etc. At the time lockfs -f /opt/SMAW was executed , they was temporarily no response.
One point important here is that the disabling of all the local mirrors finished successfully.
Next a fresh test with disabling the lockfs option resulted in no call failures. So the question here is why lockfs behave like this only on /opt/SMAW .
My understanding of lockfs is -- . Lockfs with "-w" option, is used to put a "write lock" { all lock operations implicitly perform a flush } on the mirror peice on which fallback will happen if need be.So it is necessary that the data in the unmirrored part is intact.
The steps we follow while disabling are as follows:---
1> We first check if all the mirrors for a volume is in state OKAY.
2> Then it is checked in vfstab and "mount" output if the file system is mounted.
3> The lockfs will be done for the file system to ensure file systems are synced.
4> Next the first mirror is detached. fsck is ran on the mirror piece to ensure data is in good state.
Edited by: rbruby on Mar 25, 2009 11:19 PM

Answer 1

Thanks Lasse,
When running that tool, all the links are showing zero utilization - not sure if the tool is able to show any more detailed information? 
I should state that this Lync implementation is not live, there no Lync clients out on the corporate network, and I'm not expecting any traffic on the WAN site links at this point.
It just seems to be an issue when accessing as a remote user.   At this point I think it must be because I have misconfigured CAC.  And because it is happening for a conference and not peer to peer for the remote user, then it is probably the relationship
between the front end server pool and Edge server / remote user.  
I have a local / central site defined in CAC with no bandwidth policy applied.  Initially the subnets assigned to this central site were just the LAN subnets for the central site that I knew would have Lync clients running on them.  Then as part
of troubleshooting decided to add the Edge Internal IP, and then Edge external IPs as subnets associated with this central site.  Still getting the same problem though.
Its around about times like now where I wonder if it is something simple like a service needing restarting so I'll double check things like this.   If you have any other ideas it would be greatly appreciated - a bit stuck.
Regards,
James