util.py: replace mountCgroups with checkCgroups, which
returns the cgroup version (cgroup, cgroup2)
node.py: handle cpu.max='quota period' for cgroup2
vs. cpu.{cfs_quota_us,cfs_period_us} for cgroup
net.py: make _iperfVals more robust in the case of a missing
UDP ACK
examples/cpu.py: user _iperfVals to handle new iperf behavior
of no output from telnet
This helps with virtualenv although it can open
up another security hole if you end up using an
unexpected python interpreter.
Overall it seems to make sense to err on the side
of usability but it's good to be aware of security.
However, for the remaining utility scripts that require
python 2, we explicitly note this with #!/usr/bin/python2.
Ubuntu 20.04 fixes:
- fixes sshd test
- speeds up examples/{treeping64,tree1024}.py
- debugging hacks/output for testLinkChange
- removes cfs from examples/popen.py
- improves nat in nodelib (netplan fixes?)
- makes some tests executable
- waits for switches to connect in tests to
avoid race conditions
-- adds mn -w option and wait CLI command
Changes:
- REMOVES default "-v" argument for Controller()
and adds verbose(=False) option; avoiding logging
makes it faster
- CHANGES waitConnected to wait for 5 seconds
as documented; we may wish to implement an argument
to -w to set this timeout
Issues?
- There may still be an issue with the ovs-netplan-clean
service causing the boot to hang ;-(
If we want to observe a monotonic affect, we should
make sure that we are in fact CPU limited where it
matters. In this case, we are CPU limiting the hosts,
and the iperf client uses a lot of CPU. We need to
reduce the CPU allocation so that iperf is in fact
CPU bound.
We also correct the CPU allocation so that the client
and server each receive 50% of the total. Previously
we were specifying the per-host CPU allocation, so
45% meant we were allocating 90% of the overall CPU,
which seems a bit confusing. On the other hand, now
at 40% each host gets 20% of the CPU, which could also
be considered slightly confusing!
Although the client transmit rate is going to be the
limiting factor, we still measure the received data
rate at the server, because that is more interesting
than the initial burst of buffering at the client.
Measuring at the server becomes more important as
we reduce the iperf time.
The output is also changed slightly, and the test has
been updated appropriately.
Although we could use print() from __future__, this messes
up scripts which use examples as modules.
The simple, if not nicest for 2.7, solution is to use
info(), output() and other mininet.log functions. The disadvantage
is that we may have to adjust things if we change info() to
add automatic newlines, but we can burn that bridge in Mininet
3.x.
lantz
Brad Walker
cody burkard
Brandon Heller