How to access a server

Getting info about a server

Wikimedia users talk about servers like noc, ganglia, emery, etc. but the server name alone is insufficient to access it, even if you have been granted the rights to do so. You need to provide its domain, which is easier said than done.

• Some servers are in Category:Servers on this wiki, and the server's page may give its "node name"

• http://doc.wikimedia.org/puppet/ has autogenerated info for every server that has a puppet config (but not labs instances), usually including its fully-qualified path.

• https://labsconsole.wikimedia.org lists information about "instances"

• If you know whether a host is in production (wmnet) or is a labs instance (wmflabs), if you have rights you can ssh to its bastion and enter host name to get its fully-qualified name within the domain, e.g.

user@bast1001:~$ host srv42
srv42.pmtpa.wmnet has address N.NN.NN.NN

Public? private?

Many Wikimedia Foundation servers are not publicly accessible, and you have to be in the server's domain to access it. For example, the server "emery" has the fully-qualified name emery.wikimedia.org which is known to the world, but the server "vanadium" with the name vanadium.eqiad.wmnet is only accessible from within our private wmnet domain.

There are bastion hosts that provide access, such as bast1001.wikimedia.org for production servers and bastion.wmflabs.org for labs instances.

Avoiding the full server name

Help:Access on labsconsole explains how to configure ssh so you can leave off the domain.