GitLab

swapin this morning was well over the top.

Move Brian and Chad to alumni (ahh, the loss!).  Add affils for alumni.
No way to check my syntax; hope I didn't typo.

time to fix it.

allocated nodes.

when acting as a host for jails.

fill up my mail spool directory.

config files, named with the hosts' IPs. The idea is that it's
easier for some external program (ie. frisbeed) to change what nodes
boot in an environment that doesn't have the full testbed database.

the XML to stdout, user calls to puts confuse the converter, and get
lost so the user never sees them (before, they would end up in the
start/preload log). The converter now eats up lines till it sees a
valid XML header string, printing those lines to stdout so they end up
in the log like before.

Need to deal with this in a better manner at some point. I did not
want to write the XML to temporary file. Not sure why, but I thought
that was the right approach, to keep the user from getting his grubby
little hands on it. I'm thinking now thats silly, and I'll eventually
fix things up (since it requires more TCL hacking, and I don't have
the energy for that).

was not properly kicking the NS file back to the user.

has fixed assign problems with it.

block size. I was only allowing 4K block size FS's, and assumed that
the group descriptor map was in the second block (not true when the
blocksize if 1K), and that the number of blocks in a descriptor as 8
times the block size. Also moved some sanity checks around, and
cleaned up a couple of seeks. Now works on the baleep image, which is
where we first noticed the problem.

that are displayed to about 100 (filtering reduces the number
slightly).

experiments. Linked from the "Emulab Users" page.

2. misc: use more efficient way to get interface addr,
   ignore excess interfaces,
   only consider interfaces with IPV4 addresses assigned

Chad's tiptunnel stuff. Requires ssh-mime.pl in the current directory,
to be installed as a browser helper application on the users machine.
Copied Chad's instructions for the tiptunnel from the FAQ, and stuck
it into ssh-mime.html as a help file (not really FAQ material). The
intent of this of course is to make ssh into jailed nodes easier, but
not having to know port numbers, or directly log into ops first, when
the jails are using control network IPs in our private IP space (not
routable from outside).

this fails to download that from the original location (www.postel.org
for TG and www.isi.edu for NS)

of '-p' when mapping virtual nodes.

After building the set of pclasses normally, we make another pass
through the vnodes. The goal to create a pclass for each individual
node. We disable the node's 'own' pclass to begin with. Then, the
_first_ time it gets a vnode mapped to it, we remove it from the
'regular' pclass it belongs to, and enable it's own pclass. Then, if
it goes empty, we put it back in its regular pclass and disable it's
own.

The point of this is to replace -p for use with multiplexed nodes.
Instead of disabling pclasses altogheter, which has serious
performance implications, we can instead be smart about which pnodes
remain equivalent (because nothing's been mapped to them), and which
are now different. The result is that on my test topoloy, the time to
get a good mapping has gone from over 3 minutes to about 6 seconds.

This feature is enabled with the -d option, and the -P option is
pretty much mandatory when using it, since it greatly exacerbates the
problem of cruft in the ptop file.

This satisfies #14 from the todo file:
14.  do dynamic pclasses

Also bumped up the minimum neighborhood size from 500 to 1000. In some
tests I was doing, this resulted in better solutions.

set by the user. This was added for local jails, but is applicable to
widearea assignment too.

do the actual parse. The parser now spits out XML instead of DB
queries, and the wrapper on boss converts that to DB insertions after
verification. There are some makefile changes as well to install the
new parser on ops via NFS, since otherwise the parser could
intolerably out of date on ops!

experiments created in groups and experiments created in the main
project. Too easy for the queries to get out of sync, and the
advantage is minor anyway.

turns out that getting iterators to the STL hash_* data structures is
really slow, so for some that won't be very big, use the non-hash
version.

Buys something like a 30% speedup for large topologies.

The conversion from XML to an experiment representation updates the
DB, although the experiment has to exist, and only certain fields can
be updated. For the rest of the virtual toplogy, the old toplogy is
deleted, and the new one installed. Use with caution. This is intended
to be hooked to the parser, and later to the vis tool.

fixed the trivial_link problem for Lans.
Add an option to ptopgen that says to include the various virtnode
related things, so as not to slow down assign in the general case
that virtnodes are not being used.