No Network on GCE instance

[kobi segev]

Hey,

I have a Centos7 instance on GCE that just stopped accept any connection. (no ping, ssh, http, nothing..) from serial port I can see these errors:

1. ntpd_intres[336]: host name not found: metadata.google.internal

2. NetworkManager: /usr/sbin/NetworkManager: symbol lookup error: /lib64/libgudev-1.0.so.0: undefined symbol: g_type_class_adjust_private_offset

3. systemd: NetworkManager.service: main process exited, code=exited, status=127/n/a

4. systemd: Failed to start Network Manager.

5. Bringing up interface eth0: ERROR : [/etc/sysconfig/network-scripts/ifup-eth] Device does not seem to be present, delaying initialization.

6. /etc/sysconfig/network-scripts/ifup-eth: Device does not seem to be present, delaying initialization.

Any idea what the issue might be or I can I get in to try and fix this? Any help or guidance will be appreciated

[Carlos (Cloud Platform Support)]

Hi Kobi,

Thank you for your message.

I would initially try to take a snapshot of the original disk [1]. From it I would create a new disk and spin a new VM [2][3]. Google startup scripts might or not fix the problem for you.

If the previous steps do not work, you need to attach the original disk (or a copy of it) to a VM you are able to access.  After mounting the file system you will have access to the configuration files and logs.  You can then modify those files or retrieve your data to spin a new VM. The details are explained in this article [4].

Let me know if that helps.

[1] https://cloud.google.com/compute/docs/disks/create-snapshots

[2] https://cloud.google.com/sdk/gcloud/reference/compute/disks/create

[3] https://cloud.google.com/sdk/gcloud/reference/compute/instances/create

[4] https://cloud.google.com/compute/docs/troubleshooting#pdboot