It was possible for the name resolution to steal an address from another section.
Fix by having explicit IP addressing consume and then
purge any violaters after concurrent evaluation completes.
Specify connect and login timeouts
to avoid sessions being held open.
Also, in blocking_scan, wrap everything so that finally can ensure the scan is recognized as complete.
For auto discovery, banner extraction, and interactive ssh, refactor to commen sshclient class.
This hooks the pubkeys.ssh in a manner compatible with confluent 3.x
Going from generic-redfish as most specific, then generic-https, and generic-ssh being for ssh-only targets.
For generic-https and generic-ssh, the available ports are specified so code can know if https *also* has port 22 available.
Refactor functions to let netutil depend on neighutil.
Add a suite of functions to take a mac and try to figure out some viable ip for the mac.
Provide a ping6 mainly to support a ping to ff02::1, and follow up with unicast UDP discard packets to trigger neighbor table population.
Try to use the neighbor table to figure out an ip and scope for a mac, preferring LLA.
If this fails, go for a try of converting the mac to lla arithmetically, then try all the nics to see which one seems to work.
Newer systemd prints '(unset)' where it used to print 'n/a', so an
unset console keymap was handed to the installer verbatim. An unset
System Locale has no '=' either, and the whole line was being taken
as the locale.
Linux-PAM reads vendor defaults from /usr/lib/pam.d and distributions are
migrating there package by package: systemd and polkit already ship into
it on both EL and Debian, and on SUSE 16 openssh has followed. There the
old code left a dangling /etc/pam.d/confluent and every pam authentication
against it failed.
The deb postinst carries the same logic, so fix it in step. ln -sf rather
than ln -s because -e is false for a dangling link, so the old code retried
the symlink and failed with 'File exists' instead of repairing it.
init_sdr assigned self._sdr before initialize() ran, so a failure left the
half built object in the cache. The next call saw a non-None _sdr and handed
back that partial repository rather than trying again.
The visible symptom is a first call raising and the second appearing to
succeed. The real cost is on a bmc where the read fails once: the client
keeps the incomplete sdr for the life of the session and every later sensor
lookup answers from it without complaint.
get_webclient falls off its end when /api/login answers anything but 200, so
it returned None. wc() passes that back, and thirty of the thirty-four call
sites use it unchecked, so a refused login arrived as "'NoneType' object has
no attribute 'grab_json_response'" from wherever it landed.
Raised where the failure is known, and with the status: 404 is firmware with
no web api, or a Redfish-only capture of one, while 401 is credentials it
will not take. Neither was distinguishable before.
The other four call sites are the inventory reads, and they always did check.
They answer partially when the web interface is out of reach, which is why
nodeinventory still says something useful. They ask through wc_if_available
now, so that tolerance is stated rather than resting on a None.
Jarrod Johnson
Jarrod Johnson
Markus Hilger