Lenovo Flex System x220, Installation And Service Manual

Page: 635 / 714

4. Make sure that the management node has an IP address, is on the same subnet as the CMM, and is
able to ping the CMM.
5. Make sure that the system-management processor in each compute node has a valid IP address by
checking the chassis map of the remote chassis.
Note: If the CMM recently lost connection to the DCHP server, you must reset the IMM by using the
CMM interface so that a new IP address can be acquired.
If a system-management processor does not have a valid IP address, complete the following steps:
a. Restart the CMM.
b. Check http://www.lenovo.com/support for any firmware or software updates that might apply to
this issue. You can view the release notes for a firmware update to determine the issues that are
addressed by the update.
c. Remove and reinstall CMM.
d. Replace the CMM.
6. Check http://www.lenovo.com/support for any service bulletins that are related to network connectivity.
7. Remove the management node and make sure that the connectors on the chassis midplane and the
management node are not damaged.
8. (Trained service technician only) Complete the following steps:
a. Force the link/duplex speed.
b. Check the connectors on the nodes and the chassis midplane to make sure that no pins are bent.
c. Replace the I/O expansion card in the management node.
d. Replace the management node.
Multiple nodes cannot ping the management node in a different chassis
Use the information in this section to troubleshoot multiple nodes that cannot ping the Flex System Manager
management node in a different chassis on the management network.
Complete the following steps until the problem is solved:
1. Make sure that the management node is powered on and is not hung, by logging in to the management
node. If you are unable to log in from the management network, attempt to log in from the data network.
If the management node is powered on and hung, perform a virtual reseat of the management node
from the CMM.
2. Make sure that the CMM is powered on and is not hung by pinging the CMM from the management node:
a. If the CMM is powered on and hung, reset the CMM.
b. Make sure that the IMM, the management node, and the CMM are all on the same subnet.
3. Make sure that all cables between the CMMs and the top-of-rack switch are correctly connected and
secure and that the activity LEDs are lit on the applicable ports
4. Make sure that the CMMs (the CMM in the same chassis as the management node and the CMM in the
remote chassis) have the same subnet address and can ping one another. If they cannot, make sure
that the ports on the top-of-rack switch are enabled. Attempt to connect the CMMs directly if possible.
5. Make sure that the management node has fully discovered the nodes within its chassis by viewing the
chassis map. If it has not fully discovered the nodes, follow the steps in “Multiple nodes cannot ping the
management node in the same chassis” on page 626.
6. Make sure that the management processor for each compute node has a valid IP address by checking
the chassis map of the remote chassis.
Note: If the CMM recently lost connection to the DCHP server, you must reset the IMM by using the
CMM interface so that a new IP address can be acquired.
Chapter 6
.
Troubleshooting
627