Portal:Cloud VPS/Admin/Runbooks/Kernel panic

Cloud VPS

• Cloud Services overview
• Cloud VPS user docs
• Horizon web interface (help)

Get started

• About Cloud VPS projects
• Cloud VPS user roles and rights
• List of Cloud VPS projects

Instances and access

• Set up a Cloud VPS instance
• Access Cloud VPS instances
• Add disk space to instances (detachable volumes)
• Server groups
• Sudo policies
• Unmanaged instances

Networking

• Security groups and firewall settings
• Floating IP addresses
• Web proxies for exposing Cloud VPS services to the internet

Puppet

• Puppet on Cloud VPS
• Project puppetserver

Managed storage

• Database instances (Trove)
• Object storage (S3 / Swift)

Get help

• Help and communication
• Recommendations for moving to production

Programmatic usage

• OpenStack APIs
• OpenTofu

Documentation for admins

• Cloud VPS infrastructure
• Administrator documentation
• Administrator runbooks

[]

The procedures in this runbook require admin permissions to complete.

Error / Incident

The server had a number of panic / error messages showing up in the dmesg kernel print buffer.

Depending on the error, you may want to take the server out of service immediately (depool).

Debugging

This is quite a generic alert, so there's many options. In general, a kernel panic may indicate any or all of:

• a linux kernel bug
• a hardware problem
• a firmware problem
• other severe misconfiguration

Try to ssh to the host and check the status, for example:

journactl -k --since today

Error-level logs:

journalctl -k -p err

Check all other logs:

journalctl --since today

You may also want to check idrac logs. TODO: add link.

Common issues

x86/cpu: VMX (outside TXT) disabled by BIOS

Unless you want to run VMs on that machine, this message is ok.

If you want to run VMs on that machine, you'll have to enable that flag in the BIOS

mpt3sas_cm0: Trace buffer memory 2048 KB allocated

This one is fixed in newer kernels (demoted to info): https://lkml.org/lkml/2023/1/18/1533

Acknowledging known issues

Currently the only way is to acknowledge the alert for 24h (adding a comment with !ACK mycomment or using the tick in the ui).

Note that the alert will reappear for any message that is shown on boot.

Related information

See https://wikitech.wikimedia.org/wiki/SRE/Dc-operations/Platform-specific_documentation

Communication and support

Support and administration of the WMCS resources is provided by the Wikimedia Foundation Cloud Services team and Wikimedia movement volunteers. Please reach out with questions and join the conversation:

Discuss and receive general support

• Chat in real time in the IRC channel #wikimedia-cloud ^connect or the bridged Telegram group
• Discuss via email after you have subscribed to the cloud@ mailing list

Stay aware of critical changes and plans

• Subscribe to the cloud-announce@ mailing list (all messages are also mirrored to the cloud@ list)
• Read the News wiki page

Track work tasks and report bugs

Use a subproject of the #Cloud-Services Phabricator project to track confirmed bug reports and feature requests about the Cloud Services infrastructure itself

Read stories and WMCS blog posts

Read the Cloud Services Blog (for the broader Wikimedia movement, see the Wikimedia Technical Blog)

Old incidents

Add your incident here: