2
0
mirror of https://github.com/xcat2/confluent.git synced 2024-11-25 02:52:07 +00:00
confluent/TODO
Jarrod Johnson 5bcde84e14 Fix cpu churn when web sessions open and idle
It turns out that eventlet.green.threading.Event() doesn't behave very efficiently in this context for whatever reason.
Use eventlet.event.Event() instead.  It was not used before due to lack of timeout and clear, but that is overcome by
disposing of it rather than reusing and using with eventlet.Timeout() to add timeout to wait that doesn't have built in timeout.
2014-05-15 16:22:28 -04:00

65 lines
3.5 KiB
Plaintext

-other auth
-pam if user exists but has no passphrase
-keystone?
-ad? (specialized to the AD case)
-expressionkeys never gets smaller - perf impact
-When a user account is changed, have httpapi and sockapi notified of changes
to kill off related sessions. password changes are given a pass, but user
deletion will result in immediate session termination
-When the time comes to dole out configuration/discovery, take page from xCAT
'flexdiscover' command, if possible bring an ipmi device under management
by way of ipv6 to eliminate requirement for ip to be specified.
Requires the polling event support (which is required for security anyway)
-this stack trace (happened with method was set to ""):
Traceback (most recent call last):
File "/usr/lib/python2.6/site-packages/eventlet/wsgi.py", line 402, in handle_one_response
for data in result:
File "/home/jbjohnso/Development/confluent/confluent/httpapi.py", line 301, in resourcehandler
cfgmgr, querydict)
File "/home/jbjohnso/Development/confluent/confluent/pluginapi.py", line 273, in handle_path
passvalue = pluginmap[plugpath].__dict__[operation](
KeyError: ''
-have pyghmi and friends do multiprocessing pools
-investigate console behavior when target off (causing disconnects/reconnects
on input)
-test to make sure that 'Session Disconnected' is working
-confetty to do right thing with respect to status updates when run right on a
console. It currently clutters up the screen with data that would land in
a titlebar
-audit log did not show confetty activity for starting console
-read exclusive and full exclusive console access modes
-when an http session exists, confluent cpu usage on my vm idles at 2 %.
profile exactly what is spinning and try to make it nicer
- an ipmi console was once stuck in 'connecting' state, no idea how, this traceback might have occurred:
Traceback (most recent call last):
File "/usr/lib/python2.6/site-packages/eventlet/hubs/hub.py", line 346, in fire_timers
timer()
File "/usr/lib/python2.6/site-packages/eventlet/hubs/timer.py", line 56, in __call__
cb(*args, **kw)
File "/usr/lib/python2.6/site-packages/eventlet/semaphore.py", line 121, in _do_acquire
waiter.switch()
File "/usr/lib/python2.6/site-packages/eventlet/greenthread.py", line 194, in main
result = function(*args, **kwargs)
File "/usr/lib/python2.6/site-packages/confluent/consoleserver.py", line 177, in _connect_backend
self._console.connect(self.get_console_output)
File "/usr/lib/python2.6/site-packages/confluent/plugins/hardwaremanagement/ipmi.py", line 143, in connect
iohandler=self.handle_data)
File "/usr/lib/python2.6/site-packages/pyghmi/ipmi/console.py", line 65, in __init__
session.Session.wait_for_rsp(0)
File "/usr/lib/python2.6/site-packages/pyghmi/ipmi/private/session.py", line 914, in wait_for_rsp
cls._route_ipmiresponse(sockaddr, data)
File "/usr/lib/python2.6/site-packages/pyghmi/ipmi/private/session.py", line 991, in _route_ipmiresponse
sockaddr=sockaddr)
File "/usr/lib/python2.6/site-packages/pyghmi/ipmi/private/session.py", line 1035, in _handle_ipmi_packet
self._handle_ipmi2_packet(data)
File "/usr/lib/python2.6/site-packages/pyghmi/ipmi/private/session.py", line 1063, in _handle_ipmi2_packet
self.k1, rawdata[4:-12], hashlib.sha1).digest()[:12]
File "/usr/lib64/python2.6/hmac.py", line 133, in new
return HMAC(key, msg, digestmod)
File "/usr/lib64/python2.6/hmac.py", line 68, in __init__
if len(key) > blocksize:
TypeError: object of type 'NoneType' has no len()