After adding the NetRestrict file and waiting for a scheduled fileserver
restart, the two interface addresses are still listed in the VLDB. I
suppose this is normal, since I did not explicitly issue a 'vos
changeaddr -remove'.
However, when I do try to remove them, the VLDB refuses, saying there
are still volume listings. Clearly the process does not see a
difference between the address of the fileserver and the interface
cards. This behavior is the same when trying to remove by name,
afsfs12-iscsi0, or by IP address.
vos changea -remove 123.456.789.100
Could not remove server 123.456.789.100 from the VLDB
VLDB: volume Id exists in the vldb
The only way I can see to make it happen is to move all 5K volumes to
another server, remove the listing from the VLDB and return the
volumes. Does anyone else have a suggestion on how to safely remove
these addresses? Or am I overlooking something?
I will be adding another fileserver this week, and I'll see if I can
restrict the iSCSI addresses before they are registered.
Thanks for any help.
Jeffrey Altman wrote:
Andrew Bacchi wrote:
Hello all,
I've added some new fileservers to the cell recently that use iSCSI
disk for vicepX partitions. We've noticed the named disks listed in
listaddr, and I have one report of a Mac client trying to hit it as
a fileserver.
Here's the 'vos listaddr' output. Redundant paths on each server is
the reason for the iscsi0, and iscsi1 names.
afsfs12.server.rpi.edu
afsfs12-iscsi0.iscsi.rpi.edu
afsfs12-iscsi1.iscsi.rpi.edu
afsfs13.server.rpi.edu
afsfs13-iscsi0.iscsi.rpi.edu
My questions are:
Since they are listed in the VLDB, will clients try hitting them?
yes. If the VLDB reports the address to the client, the client will
use it.
If I remove them with 'vos changeaddr -remove' will they be
relisted somehow? Will they return?
As long as your file server continues to report those addresses as
being valid addresses for the file server, they will continue to be
inserted into the VLDB.
Use the NetRestrict file on each file server to limit the reported
addresses.
http://www.openafs.org/pages/manpages/5/NetRestrict.html
--
veritatas simplex oratio est
-Seneca
Andrew Bacchi
Systems Programmer
Rensselaer Polytechnic Institute
phone: 518.276.6415 fax: 518.276.2809
http://www.rpi.edu/~bacchi/
_______________________________________________
OpenAFS-info mailing list
[email protected]
https://lists.openafs.org/mailman/listinfo/openafs-info
