tokio.connectors.lfshealth module

Connectors for the Lustre lfs df and lctl dl -t commands to determine the health of Lustre file systems from the clients’ perspective.

class tokio.connectors.lfshealth.LfsOstFullness(*args, **kwargs)

Bases: tokio.connectors.common.SubprocessOutputDict

Representation for the lfs df command. Generates a dict of form

{ file_system: { ost_name : { keys: values } } }

__repr__()

Serialize object into an ASCII string

Returns a string that resembles the input used to initialize this object:

snx11025-OST0001_UUID 90767651352 63381521692 26424604184 71% /scratch1[OST:1]

load_str(input_str)

Parse the output of lfs df to initialize self

class tokio.connectors.lfshealth.LfsOstMap(*args, **kwargs)

Bases: tokio.connectors.common.SubprocessOutputDict

Representation for the lctl dl -t command. Generates a dict of form

{ file_system: { ost_name : { keys: values } } }

This is a generally logical structure, although this map is always almost fed into a routine that tries to find multiple OSTs on the same OSS (i.e., a failover situation)

__repr__()

Serialize object into an ASCII string

Returns a string that resembles the input used to initialize this object

get_failovers()

Identify OSSes with an abnormal number of OSTs

Identify OSTs that are probably failed over and return a list of abnormal OSSes and the expected number of OSTs per OSS.

load_str(input_str)

Parse the output of lctl dl -t to initialize self