GitLab

much better.

Switch around 'value' and 'defaultvalue' fields for "watchdog/rusage" and
set value field to NULL.

schema.

CR_SERVER_GONE_ERROR). Instead of bailing, sit and loop trying to
reconnect, given that this is known to be a transient error, and we do
not really want our daemons to go belly up during that brief time when
the watchdog is getting it restarted. The query is then resent.

For the perl version of this change, I was a lot more pedantic since
we use this library from a zillion places. Also, there is some special
handling cause of the mysqld watchdog which would become useless if
the test query hung trying to reconnect to the server forever.

As a side effect of this change, we should see way less email when
mysqld goes catanoic since the new code will just loop instead of
generating tons of errors.

Might actually increase overall rebustness. On the other hand, could
end up being a total disaster!

CR_SERVER_GONE_ERROR). Instead of bailing, sit and loop trying to
reconnect, given that this is known to be a transient error, and we do
not really want our daemons to go belly up during that brief time when
the watchdog is getting it restarted.

This turns out to be nice in my elabinelab experiments, where very
idle daemons have a habit of timing out, although there were other
ways to deal with that, but I'm lazy.

Most programs already use that library; getting the stragglers before
I whack the library for mysql 5.X

Most programs already use that library; getting the stragglers before
I whack the library for mysql 5.X

and report the numbers without passing judgement on the results.

This way, the linktest instances can know who will be running linktest
and who will not and thus we can run on a subset of nodes.

there is a query (by Mike I think) that has this form:

          SELECT * FROM t1, t2 JOIN t3 ON  ...

Prior to 5.X, the comma operator had the same precedence as join, and
so it was evaluated left to right. Well, now comma is lower then join
and this query breaks cause the right side is evaluated before the
left side. The solution is simply:

          SELECT * FROM (t1, t2) JOIN t3 ON  ...

but as the manual points out, why not just use:

          SELECT * FROM t1 JOIN t2 JOIN t3 ON ...

and avoid precedence issues.

are now reserved words mysql 5.X. Skip to the next entry ...

5.0 client side is very picky about this.

5.0 client side is very picky about this.

Move Univ of Newcastle nodes to Europe class.  Oops!  These nodes actually
route over commodity Inet, which is what threw me off.

now and the -FUTURE kernel delay-agent has too many problems.

Set the magic variable in auto-pelab.ns and swapin your experiment.
Logs will wind up in /proj/blahblah/exp/blahblah/logs/<node>/<name>.tcpdump
after you run stop-experiment.

I did this with the program agent and duct tape.  If it turns out that
you can already do this with the link tracing framework, then someone in
Oregon is going to whomp me upside the head with a clue-stick and he'll
never let me forget...at least til he forgets in a day or two.

Fixed bug where a single packet in a quantum caused an infinite loop. Fixed a bug where only socket buffer size updated in a quanum caused an infinite loop. Outstanding issue: Odd emulab side latency. Way too low.

Added ping error handling: timeout, ttl exceeded, and unknown.
 Values returned to ops are negative numbers. The meanings of these
 is listed in the top section of bgmon.pl
Note, that these changes have not been tested but were not complex.
testing is a definite TODO.

IP and not the real plab node IP.

Minor bugfix - quote the "NULL" strings.

Commit the option I added to runbgmon/bgmon.pl that allows you to specify
which port iperf is to use (-i <port>).  Also fix a minor issue with latency
and bandwidth period option parsing.

of the planet-* links.

know what is actually being reported.