![]() | Oracle System Handbook - ISO 7.0 May 2018 Internal/Partner Edition | ||
|
|
![]() |
||||||||||||||||
Solution Type Problem Resolution Sure Solution 1514153.1 : Pillar Axiom: Solaris APM Host Status Not Communicating due to Duplicate iSCSI Qualified Name
iSCSI Qualified Names are supposed to be world-wide unique identifiers. When Solaris systems are prepared from common source or cloned drives, the iSCSI Qualified Name is also duplicated if copying of the /etc/iscsi/iscsi_v1.dbc is not suppressed. In this Document
Created from <SR 3-6552931151> Applies to:Pillar Axiom 300 Storage System - Version All Versions to All Versions [Release All Releases]Pillar Axiom 600 Storage System - Version All Versions to All Versions [Release All Releases] Pillar Axiom 500 Storage System - Version All Versions to All Versions [Release All Releases] Information in this document applies to any platform. SymptomsA Solaris host running the Axiom Path Manager may show as Not Communicating in the Axiom GUI. If any of the following happen, the hosts that are affected can change:
The "Not Communicating" status can have other causes. This article documents how to identify and resolve the issue when this failure is due to duplicate iSCSI Qualified Names in multiple Solaris hosts. This Not Communicating will happen any time there are multiple Solaris hosts using the same iSCSI Qualified Name, whether or not the host is using iSCSI. Only a single host with the same iqn will be allowed to log in. The rest will fail the APM login due to duplicate iqn and be in "Not Communicating" status. If the host that successfully logs in loses the TCP connection for any reason, some other Solaris host with the same iqn may log in successfully, which can result in the Communicating/Not Communicating status changing. CauseThe host status is displayed as Not Communicating if that host does not currently have an APM daemon login session to the Axiom Pilot, but has logged in at some point in the past. The three statuses are:
These error messages are due to two or more Solaris hosts having the same iSCSI Qualified Name for their initiator. Even if the hosts do not use iSCSI this can still be an issue and must be fixed. The APM daemon running on Solaris gets the iSCSI initiator-node name from the the iSCSI Management API library (/usr/lib/libima.so). SolutionThe solution is to cofigure each Solaris host with a unique iSCSI initiator node name. There are two methods to verify this issue.
Checking the iSCSI Qualified Name, or iqn To check the configured initiator node name on a Solaris host, run the following command and look at the Initiator node name field: For more information on the iscsiadm command, see iscsiadm(1M) # iscsiadm list initiator-node
Initiator node name: iqn.1986-03.com.sun:01:002128fa2e16.50904a7b Initiator node alias: ca-ftsun-10 Login Parameters (Default/Configured): Header Digest: NONE/- Data Digest: NONE/- Authentication Type: NONE RADIUS Server: NONE RADIUS access: unknown Tunable Parameters (Default/Configured): Session Login Response Time: 60/- Maximum Connection Retry Time: 180/- Login Retry Time Interval: 60/- Configured Sessions: 1 Collecting APM Host Debug Logs and verifying log entries Since the APM host is not communicating with the Axiom pilot, debug logs must be generated and collected at the host. To create debug logs, type /opt/pillar/sbin/axiompmd -d. This will produce a logset directory in /var/log/axiompmd, named for the time of log creation. For example: Tue_Dec_18_14_20_24_2012/. The Daemon_Trace is in that directory, with the hostname prefixed. For example, if the hostname is shost3, the file name will be shost3_DaemonTrace.txt. This file can be viewed directly on the Solaris host. The typical sequence in the axiompmd debug daemon trace log is: some lines removed date/time INF 00000004 PrtIscsiInfo: Initiator Name iqn.1986-03.com.sun:01:0010e0214064.50bf35d7 If you see the above Connected to and Logged in messages, the communications path from the APM host and the Axiom are available. Then the APM host and Axiom exchange APM version negotiation. The current supported version is "2" MccSendRequest: Queue Request MGMTILIB_NEGOTIATE_VERSION to 10.10.10.10 The APM daemon sends the Configure San Host request. This request includes the hostname, a list of all host HBAs and HBA WWNs discovered, all iSCSI initiator identities (whether or not iSCSI is in use), and the LUNs and paths that queries to the Host multipath driver API can detect. MccSendRequest: Queue Request MGMTILIB_PERFORM_CONFIGURE_SAN_HOST_V2 to 172.32.221.19
Modifying the iSCSI Qualified Name, or iqn Following is an example to disable APM, modify the initiator node name via the 'iscsiadm modify initiator-node --node-name <initiator-node name (should be globally unique)>' command, and re-enable APM: # svcadm disable axiompmd
# svcs axiompmd (Wait until the axiompmd service is in disable STATE) # iscsiadm modify initiator-node --node-name iqn.1986-03.com.sun:01:0123456789ab:cdef0123 Make sure this value is unique on all hosts running APM # svcadm enable axiompmd Verify changes to the initiator name took place: # iscsiadm list initiator-node
At this point APM should be able to log into the Axiom pilot and the host should report a Communicating state in the GUI. If this does not solve the issue, you may be asked to provide the output of iscsiadm list initiator-node and another set of APM debug logs to the Support Center for assistance. Attachments This solution has no attachment |
||||||||||||||||
|