Project Home
Project Home
Wiki
Wiki
Discussion Forums
Discussions
Project Information
Project Info
Forum Topic - mount QNET vs starting with io-pkt: (4 Items)
   
mount QNET vs starting with io-pkt  
I have a system with 4 units running QNET. What I would like to do is use a resolver file, so that specific units can 
only be seen by other specific units. I.e 4 units are really 2 partner sets, and I'd like only partners to see each 
other and not the rest of the network.

I've created a resolver file qnet_hosts (each unit has 2 nics)
------------------------------------------------------------------------------
#
# QNET Host Database
#
UNITNAME1	00:80:74:00:13:2c,00:80:74:00:13:2d
------------------------------------------------------------------------------

If I start io-pkt this way:
io-pkt-v4 -dppc885_mii channel=1,phy=1 -dppc885_mii channel=2,phy=3 -ptcpip -pqnet auto_add=0,resolve=file:/bin/
qnet_hosts

Everything works as expected.

However if i start it this way:
mount -Tio-pkt lsm-qnet.so auto_add=0,resolve=file:/bin/qnet_hosts

QNET starts, but it sees every other node on the system. I'd like to mount QNET by itself, because the file is in 
another mount point not available at the time io-pkt starts.

Does mount not support the command line options or am I not entering the options properly? 

Thank you for your input in advance.
Re: mount QNET vs starting with io-pkt  
> 
> However if i start it this way:
> mount -Tio-pkt lsm-qnet.so auto_add=0,resolve=file:/bin/qnet_hosts

Try:

mount -Tio-pkt -o auto_add=0,resolve=file:/bin/qnet_hosts lsm-qnet.so

-seanb
Re: mount QNET vs starting with io-pkt  
That did the trick. Thanks Sean!
RE: mount QNET vs starting with io-pkt  
I hate mount.  Everyone loses the command line 
options, and you run with less stack.

--
aboyd