Symptom/Cause | Solution |
|---|---|
Non availability of telnet/ssh service | When the node is booting up, we disable all remote access. Upon the start of hostpd, the service xinetd starts. Make sure hostpd is running or started during init sequence of board initialization, and xinted service is running. Execute at the Linux prompt and verify listening socket: ip netns exec zebosfib1 netstat -tpln Active Internet connections (only servers) Proto Recv-Q Send-Q Local Address Foreign Address State PID/Program name tcp 0 0 127.0.0.1:705 0.0.0.0:* LISTEN 30044/snmpd tcp 0 0 0.0.0.0:199 0.0.0.0:* LISTEN 30044/snmpd tcp6 0 0 :::22 :::* LISTEN 29997/xinetd tcp6 0 0 :::23 :::* LISTEN 29997/xinetd tcp6 0 0 :::830 :::* LISTEN 29997/xinetd |
Failure to authenticate a user | If the basic files for Linux authentication of a user are missing/corrupted, the login to the node is denied. Using console root user, make sure the /etc/passwd file has an entry for the user trying to login. Look for authentication errors are in /var/log/messages, for more about such failures. |
Remote access to the node via telnet/ssh hangs | The shell imish/cmlsh is configured for all OcNOS users, except for user root, which is accessible via console only. If the module imi or cmld is not responding, then there will be no imish/cmlsh prompt after successful login. The system monitoring module (pservd) restarts such hung modules, recovering hang states of one of more modules. Look for the core directory (/var/log/crash/cores) and syslog messages in /var/log/messages to find the actions from system monitoring module. |
Continuous restart of any module | If any module is restarting continuously, disable monitoring such module via: no software-watchdog <module name> If the NSM/HSL module crashes or hangs, the system reboots. The system does not reboot automatically when the earlier two reboots were due to HSL or NSM crashes during the initial few minutes of board boot up. This is to stop continuous reboots of the system due to NSM/HSL crashes. There is no mechanism to disable this except for disabling pservd service. Stop the service pservd to disable it. If module pservd is hung, it will be restarted in 5 mins. |
Deleting ZebOS.conf looses management IP address | During ONIE installation, if you do not configure a static IP address, OcNOS boots and gets an IP address for eth0 (management port) through DHCP and updates the /etc/network/interfaces file. Once you configure a static IP address from the OcNOS command line and save the configuration, OcNOS updates /etc/network/interfaces and changes the method used to configure eth0 from dhcp to static. In this scenario, if you delete ZebOS.conf, then the management IP address is lost and you can only recover management access by assigning an IP address via the console. |
sys-update install <installer> failure | • No free space left on system. Minimum 1 GB space is needed:remove some files to make available space > 1GB on device. • Binaries not compatible with the board: use proper installer file for the respective board. • Installer not downloaded properly, try again: downloaded installer file is not complete. • Source Interface not found. • OcNOS version you are trying to upgrade is already Installed: no need to upgrade again, you have the same version already installed. • File not found on board: installer file is not present on board for given path, provide valid path for installer file. • File not found on server: installer file is not present on the server provided in the link, provide valid link for installer file. • Server connection timed out: waited 60 seconds for server to respond. • Unsupported protocol: the ftp, http, tftp, and file protocols are supported. • Invalid installer: installer file is not valid. • % Source interface is not up : Ensure source interface is UP. Note: When the sys-update operation stops without any error, check whether the IP reachability is there to download the installer. |
sys-update install <deb package> failure | • No free space left on system. Minimum 1 GB space is needed: remove some files to make available space > 1GB on device. • Unsupported protocol: the ftp, http, tftp, and file protocols are supported. • Unsupported OCNOS image format (need:*.deb): deb package name should be like <filename>.deb. • Kernel changes are present in this version, sysupdate not possible: upgrade using installer. • Binaries not compatible with the board: use proper installer file for the respective board. • OcNOS version you are trying to upgrade is already Installed: no need to upgrade again, you have the same version already installed. • Non-ZEBM to ZEBM upgrade using *.deb not allowed: use installer for non-ZEBM to ZEBM upgrade. • ZEBM to non-ZEBM upgrade using *.deb not allowed: use installer for ZEBM to non-ZEBM upgrade. |
When incompatible transceiver(s) are inserted it may result in Device Monitoring (DDM) to be disabled and may result in hang. | • When Incompatible transceiver(s) insertions results in device monitoring to be disabled and possible board hang, two possible recovery mechanisms are available: • Remove the incompatible transceiver • Power Cycle the board (Not reboot) • When Incompatible transceivers insertion issue is seen, system reboot or CMMd restart could lead to undefined behavior (includes inaccessible system). • When the system first boots up after upgrade with incompatible transceiver present which results in issue, the system might be inaccessible after the upgrade. |
Coherent optical line interface doesn't come up. | • There could be two possible reasons: • If the optical module is inserted/removed on a running system, need to power-cycle (not reboot) the system to recover as hot-swap is currently not supported. • Configuration mismatch - Mismatch in configuration of few of the attributes (Modulation, frequency etc.,) on both the ends will not bring up the line interfaces. Please make sure that configuration matches (show coherent-module <Slot Number>) on both the ends. |