In the Windows OS, when the disks mounted to the initiator client are being initialized or data on the disks are being replicated, the following phenomena occur:
The disks mounted to the initiator client has a slow I/O response (3 to 4 Mbit/s), but the network port on the server is in full-duplex mode and works at the auto-adaptive rate of 1 Gbit/s.
In the Windows OS, the disks mounted to the initiator client cannot be quickly formatted or normally formatted. At the same time, the system has slow or no response. The system log reports the iScsiPrt error and the event ID is 5.
The phenomena mentioned above causes that the customer cannot properly use the SAN storage network.
Step 1 Connect the redundant links.
Connect the port G0. Data is accessed on network segment 250.
Step 1 Change the target IP address.
Change the target IP address in the initiator client and log out of the initiator client. In this way, data can be accessed on network segment 251.
Step 1 Modify the VLAN settings of switches.
Delete the port of switch B that connects to network segment 251 from the VLAN of network segment 251.
Add the previous port to the VLAN of network segment 250.
Save the settings. Data on the storage array can be accessed on network segment 250.
Check the network link first. Check whether the storage array has alarms, and then check the system log and HBA drive version and working mode.
The troubleshooting procedure is as follows:
Step 1 Log in to the Windows OS and ping the service network port of the storage array.
> ping 184.108.40.206
Reply from 220.127.116.11: bytes=32 time<53ms TTL=64
Reply from 18.104.22.168: bytes=32 time<102ms TTL=64
After the ping command is executed, the displayed messages indicate that the service network port is normal but has long latency. The network link may be defective.
Step 2 Check whether the storage device has alarms.
Log in to the OSM and click Alarm. No exception or alarm exists.
Step 3 Check the HBA drive version.
The HBA drive is of the 2006 version. Download the latest version from the official web site and upgrade the HBA drive. The fault persists.
Step 4 Check whether the multipathing software is enabled.
The multipathing software is installed but disabled.
Step 5 Check whether the target IP address is correctly entered.
One mounted path is connected to another switch. The network cable of port G0 is disconnected.
The result proves that the fault is not caused by hardware drive and link. After the storage device is directly connected to the server without switches, I/O response becomes normal, which indicates that the fault is not caused by the storage array. The cause is the incorrect target IP address. Switch A is connected to the storage device over network segment 250, while switch B over network segment 251, and the switch management network port (100 Mbit/s) is connected to the core switch. When the network cable of port G0 is disconnected, the auto-adaptive rate among switches A and B, storage array, and server is 1 Gbit/s. Data is accessed on network segment 250 rather than network segment 251 by using the target IP address. The transfer path for data packages goes through the management network pot of switch A, the core switch, and then switch B. However, the auto-adaptive rate between the management network ports of switches A and B and the core switch is 100 Mbit/s. Therefore, the rate for copying small-sized files is still 3 to 4 Mbit/s and the error iScsiPrt ID 5 is reported.
To improve data security on a IP SAN, the multipathing software is often installed to implement redundant links. However, customers change redundant links to single link to improve the access density of the server. When adding a target IP address, customers may enter the IP address of another port by mistake, causing that the IP SAN cannot be accessed.