ONTAP Discussions
ONTAP Discussions
I rebuilt my 4-node lab cluster the other day using 8.3RC1, and I've run into an odd issue that I hope the community can help me with. I setup a SAN SVM with both FCP and iSCSI. I created the requisite volumes, LUNs, portsets, LIFs, igroups, etc. to export an iSCSI and FCP LUN to my CentOS 6 client. This client had previously mounted FCP and iSCSI LUNs from this cluster when it was 8.2. I was forced to rebuild this cluster from scratch, so this is essentially a clean build and not an upgrade.
My problem is that using either iSCSI or FCP I can see the client login to the filer on all 4 nodes:
labcluster::system license> event log show
Time Node Severity Event
------------------- ---------------- ------------- ---------------------------
2/16/2015 21:24:02 labcluster-01 NOTICE scsitarget.fct.portLogin: Login at target FC port:'0a' by initiator port '21:00:00:24:ff:31:46:fe'-addres 0x7d0200. The target virtual port is:'NetApp FC Target Port (2532) san_vserver:labcluster-01_fc_lif_1'.
2/16/2015 21:24:02 labcluster-02 NOTICE scsitarget.fct.portLogin: Login at target FC port:'0a' by initiator port '21:00:00:24:ff:31:46:fe'-addres 0x7d0200. The target virtual port is:'NetApp FC Target Port (2532) san_vserver:labcluster-02_fc_lif_1'.
2/16/2015 21:24:02 labcluster-03 NOTICE scsitarget.fct.portLogin: Login at target FC port:'0a' by initiator port '21:00:00:24:ff:31:46:fe'-addres 0x7d0200. The target virtual port is:'NetApp FC Target Port (2532) san_vserver:labcluster-03_fc_lif_1'.
2/16/2015 21:24:02 labcluster-04 NOTICE scsitarget.fct.portLogin: Login at target FC port:'0a' by initiator port '21:00:00:24:ff:31:46:fe'-addres 0x7d0200. The target virtual port is:'NetApp FC Target Port (2532) san_vserver:labcluster-04_fc_lif_1'.
However, on the client side I only see two nodes:
[root@san_client ~]# sanlun lun show -p
ONTAP Path: san_vserver:/vol/fcp_volume/lun0
LUN: 0
LUN Size: 100g
Product: cDOT
Host Device: 3600a098051763175562b46746d634345
Multipath Policy: round-robin 0
Multipath Provider: Native
--------- ---------- ------- ------------ ----------------------------------------------
host vserver
path path /dev/ host vserver
state type node adapter LIF
--------- ---------- ------- ------------ ----------------------------------------------
up primary sdf host1 labcluster-01_fc_lif_1
up secondary sde host1 labcluster-02_fc_lif_1
My SAN is non-redundant, but I should still see 4 paths -- one for each controller. My client /etc/multipath.conf is setup per NetApp best-practices.
I've tried just about everything I can think of to troubleshoot this. I'm thinking it might be a license issue (licenses are demos from NetApp) that would limit the protocols to 2 nodes. Has anyone else experienced this? Is there a way to check the license node limits in 8.3?
Thanks!
Solved! See The Solution
In 8.3 selective lun mapping is enabled by default.
In an upgrade scenario, the old behaviour is retained.
In 8.3 selective lun mapping is enabled by default.
In an upgrade scenario, the old behaviour is retained.
I didn't even realize this was a new behavior. I did find this documentation link that explains the behavior. Thank you!
https://library.netapp.com/ecmdocs/ECMP1636035/html/GUID-62ABF745-6017-40B0-9D65-CE9F7FF66AB3.html