GitLab

failure, and that the state of the experiment is set back to NEW. Also
some minor print and email formatting changes to be more informational
when things go wrong.

Remove hot list for osids from big node listing to cut down on the
size of the html that gets spit out.

use.

but simply entered into the DB record for the experiment until we know
what to do with them. Add to batchexp script arguments, since all that
stuff is done outside the web interface. Add a swapexp perl script to
swap an an experiment in/out form the command line. Add web links on
the Experiment Information page to do this from the web interface. A
bunch of locking changes. Previously expt_terminating in the
experiment record prevented multiple calls to terminate an experiment,
but now we have a more general locking problem with
start,swapin,swapout, and terminate, so change expt_terminating to
expt_locked (still a datetime field) and add locking to all of
startexp, swapexp, and endexp. Note that batch experiments cannot be
swapped yet because of locking issues still to be resolved. Minor
cleanup in tbreport to make email message look better.

projects that the user has project_root permissions in, and give a
list of those to choose from.

default OSID from the node_types table, to a specific OSID from the
partition table on the actual node. This is to avoid setting the boot
OSID to RHL_STD when the node is released, which causes a boot
failure. Okay, so I added a library routine to do this (yanked out of
os_setup where I did the code originally). This would solve most of
the problems, except where there was no OS loaded that would satisfy
the mapping, in which case the user must have done an os_load, and now
that auto schedules a reload. Anyway, seemed like this should work.
Ha! Mysql locking is downright dumb; all tables used within a lock
region must be locked. nfree was already locking 9 tables, and in
order to call out to library routines (which might use anything) I
would have to lock the world, which is not actually possible anyway.
Why all this locking in nfree in the first place? The idea is that
there is a race between releasing the node from reserved, and cleaning
up all those tables (interfaces, delays, nodes, etc). We don't want to
free a node, and have it get allocated to another experiment before
the cleanup is done, since that would mess up the state of the node.
The solution (albiet a crufty one) was to lock just the reserved table
(which guards against multiple people trying to nfree the same node at
the same time) and switch the reservation out of the pid,eid and into
a holding reservation. This effectively removes the node from the
users control, but keeps it reserved. Then I unlock the reserved
table. With that done, I can clean up all those tables without any
locking, since the node is still reserved. After cleanup, I can either
delete the reservation, or move it to the next reserve or reload
reservation if those were pending. No locking is needed at this point
since single table changes are atomic (and nalloc locks reserved
anyway). Okay, so now we sit back and see if this was a good idea.

have entries in scheduled_reloads. Also changed hard-coded reload
types to use the constants in libdb for flexibility.

-v is for verbosity,
-t specifies a timeout number of seconds before exiting if there are no clients.

the image file. Previously, if the user did not, the frisbeed process would
not start properly, but nothing would complain.

arguments; do not need to specify the filename of the image since that
is in the imageid descriptor in the DB. Also reverse the arguments to
be consistent with every other script in the system!

is a complete filename.

reloads for nodes in an experiment.
Change os_load to schedule a default image reload whenever a mereuser
loads an image that is not the default image for that node type.
Add some support stuff in libdb (TBSetSchedReload) and some constant
definitions for sched_reload and for nodelog.

viewing the log.

fails. In batch mode, the daemon takes care of this, but in immediate mode
the parent script has exited, so it has to be done here.

Add 'update users set admin=1 where uid='flest';' to the dumped db so that flest is set as admin for its runs. Gives it magical powers to clear out all the expts. We may want to someday change this to be UID instead of 'flest', so that whoever is running it becomes admin. But for now, I think everyone who runs the test suite already is an admin.

done a couple of weeks ago).

experiment code. No longer uses another table. Rather, the experiment
record contains a couple of extra fields for the batch system. Also
combined some of the backend code (no longer a killbatch script).
Also added scriptable experiments; the batchexp program in the bin
directory can start an experiment from the command line, and in fact
is used from the web page for both batch experiments and immediate
experiments (-i option). All of the DB code that was in the web
interfaces was moved to batchexp.

Now back to its old behavior, where frisbee reloads are handled no differently
than other reloads, because this difference is now taken care of in the
reload_daemon.

Note that the '-w' flag has been removed. This is due to an error that perl
reports on line 123, which, as far as I can tell, is not real. The code works
fine, anyway.

reloads to finish.

commit.

reserved to get its load info.

Did this with a little DB table that lists any aux machines that
groups need to be built on.

group, you need a -G "" argument? Had to read the source code to learn
that.