Configuring iSCSI Initiators and Multipathing on Enterprise Linux

When provisioning high-performance database servers across multi-site data centers (such as east-dc1 or west-dc1), establishing stable block storage access over TCP IP networks requires configuring redundant network pathways. If an ethernet switch fails or a storage processor reboots, single-path iSCSI connections will instantly drop application disk access.

In this operational manual, I outline how to assign dynamic route links, discover iSCSI Target portals (iqn), initialize redundant session bindings, and configure EMC Navisphere Host Agent services on CentOS Linux.

Step 1: Configuring Dedicated Storage Routing

Before discovering SAN target groups, establish static explicit gateway routes pointing storage traffic across dedicated backend network switches.

# Append static host gateway routing
route add -host 10.12.7.65 gw 10.12.13.5
route add -host 10.12.7.66 gw 10.12.13.5
route add -host 10.12.7.76 gw 10.12.13.5
route add -host 10.12.7.77 gw 10.12.13.5

# Validate routing paths and port responsiveness
netstat -rn
telnet 10.12.7.65 6389

---

Step 2: Deploying EMC Navisphere Host Agent Daemons

For intelligent array discovery and storage LUN reporting, install and bootstrap the EMC Navisphere Host Agent RPM binaries.

# Unpack agent archive tarballs
cd ~wazeem/
unzip NaviHostAgent-Linux-6.26.32.0.72.zip

# Install CLI helper daemon packages
rpm -ivh naviagentcli-6.26.32.0.72-1.noarch.rpm
vi /etc/Navisphere/agent.config

Verify service status and ensure daemon background processes are actively binding to required management ports (6389):

service naviagent restart
ps -ef | grep -i navi
netstat -an | grep 89

---

Step 3: Discovering and Logging into iSCSI Portals

Use the command-line administration tool iscsiadm to scan target storage portals and establish underlying session maps.

1. Initializing SendTargets Discovery

Search Target IP addresses on standard iSCSI port 3260 to enumerate available IQNs:

# Verify initiator daemons are running
service iscsi start
service iscsid status

# Execute sendtargets discovery
iscsiadm -m discovery -t sendtargets -p 10.10.8.12:3260

2. Executing Target Login and Automatic Startup

Attach logical block target names across redundant storage processors (such as cx.apm00121204249.a5 and b5) and set the nodes to reconnect automatically on server boot:

# Login to Primary Storage Processor A
iscsiadm --mode node --targetname iqn.1992-04.com.emc:cx.apm00121204249.a5 --portal 10.10.8.12:3260 --login
iscsiadm --mode node --targetname iqn.1992-04.com.emc:cx.apm00121204249.a5 --portal 10.10.8.12:3260 -o update -n node.startup -v automatic

# Login to Redundant Storage Processor B
iscsiadm --mode node --targetname iqn.1992-04.com.emc:cx.apm00121204249.b5 --portal 10.10.8.14:3260 --login
iscsiadm --mode node --targetname iqn.1992-04.com.emc:cx.apm00121204249.b5 --portal 10.10.8.14:3260 -o update -n node.startup -v automatic

---

Step 4: Validating Session Mappings and Disk Device Discovery

Verify active session parameters and force SCSI kernel bus rescans to discover incoming disk nodes.

# Inspect established session identities
iscsiadm -m session

# Force SCSI target topology rescans
echo "- - -" > /sys/class/scsi_host/host0/scan

Audit operating system kernel logs and verify incoming raw sd block devices:

tail -n 50 /var/log/messages
ls /dev/sd*
fdisk -l | more

References

• Managing Linux iSCSI Administration
• EMC Host Agent Installation Manual