reserved port-checking

It is actually a param mount.checkport and the default should be ok

I think there is also a white paper about DFNS on Powerlink with more info

Thank you Rainer!

I went through the doc and the doc says just reserved port checking should be disabled.

and we dont have option to do this.

Should we do via # server_param   (or) through server_export for the NFS file system.

Thanks & have a nice day.

I believe all you need is the param "transChecksum". This option is what enabled support for DNFS on Celerra by ensuring the data mover uses a unique ID for each transaction so you avoid conflicting IDs that result from reuse of relinquished ports..

Do enable this log into the cli and give this a whirl to check it's current setting:

server_param server_X -facility nfs -list

We need it to be enabled (1)...  to do that you simple:

server_param server_X -facility nfs -modify transChecksum -value 1

Thank you Mark.

Right now it is:

transChecksum                        nfs          0          0

So, I will  run the following:

This change will simply make the datamover support DNFS.  It shouldn't have any impact on NFS shares which aren't for DNFS,  but it is datamover wide. I can't recall if it requires a reboot,  it will tell you after you change the param.  You won't need to change your server exports at all.

I found this blurb:

Thank you Mark.

After changing the parameter transChecksum too, the NFS mount for insecure option(Reserved port checking) for file system didn't work.

did we need anyother finetuning, as we have parameeters like: secureExportMode

$ server_param server_2 -facility nfs -list
server_2 :
param_name                       facility  default     current   configured
NTcred.trace                         nfs          3          3
NTcred.size                          nfs       1009       1009
secureExportMode                     nfs          0          0
NTcred.LDAP                          nfs          0          0
NTcred.winDomain                     nfs
spExportLookup                       nfs          0          0
NTcred.TTL                           nfs         20         20
rpcgss.discardReplay                 nfs          1          1
nthreads                             nfs        512        512
ofCachesize                          nfs     360000     360000
checkForExistingFileFirst            nfs          0          0
transChecksum                        nfs          0          1          1
v3xfersize                           nfs      32768      32768

What do you mean by „didn’t work” ?

The paramater secureExportMode just tells the datamover how to deal with conflicts between read-only and read-write exports with more than one netgroup.  Basically if you create a read only export for netgroup X and a read write export for the same netgroup X this setting depermins which one wins.

If it's set to 0 it will allow read-write access.  If it's set to 1 it will only access read-only.

what he meant is mount requests from non-privileged ports

that’s actually a mount parameter – that’s why it isn’t shown in the output from the nfs parameters that he was showing

still doesn’t explain what he meant with “doesn’t work”

Hi Rainer

My bad on the clarification

Actually I refer that, when client linux oracle server after mounts the NFS share and attempting  ... it comes with the error message

[1363846299666946] skgnfs_bind:1295: sockfd 32 port 58149 address 10.106.107.43

[1363846299666983] kgnfs_connect:10131:  skgnfs_connect() failed

so, the above is the block now.

Thanks

sorry - but that message isnt saying much

I think you need to troubleshoot further

maybe like this http://www.pythian.com/blog/dnfs-wtmax-procfsnfsdmax_block_size/