- 28 May, 2002 12 commits
-
-
Leigh B. Stoller authored
-
Leigh B. Stoller authored
-
Leigh B. Stoller authored
-
Leigh B. Stoller authored
state when tearing down a failed experiment.
-
Leigh B. Stoller authored
other than just PCs. Needs a more complete solution ...
-
Leigh B. Stoller authored
this last part! Senility is setting in. Anyway, change join project to not send the email to the projleader when the user is new; let that happen out of the verify page when the newuser does the verification. Basically, duplicate some code to generate the email message. XXX: I have not fixed up approveuser.php3 to check to make sure that the user has been verified. Basically, for this to happen the proj leader would have to generate the URL by hand, and thats not likely, and not really a dangerous thing anyway. However, it is confusing.
-
Leigh B. Stoller authored
node information web page.
-
Leigh B. Stoller authored
information page. Got tired of having to past the them all by hand between windows. Only admins can do this. Also, add subpage menu stuff to make the page look nicer.
-
Leigh B. Stoller authored
-
Leigh B. Stoller authored
generating a warning in the startexp logfile.
-
Leigh B. Stoller authored
happened (in the web page listing).
-
Shashi Guruprasad authored
and a simulated node, things failed
-
- 27 May, 2002 3 commits
-
-
Kirk Webb authored
bit set. It appears that cvs doesn't notice just permission changes (?)
-
Shashi Guruprasad authored
-
Leigh B. Stoller authored
Add sorting by number of nodes reserved (PCs column). Right now this is total number, so it includes widearea nodes, but I plan to split these apart. Sorting by PCs required changing in the initial query quite a bit. Also changed the final report to include all classes, and convert into a prettier table. Add a minor optimization; instead of querying the DB for the number of node classes used by experiment (which at this moment means 107 queries), do one query and split into a little hash table of per exper usage totals.
-
- 24 May, 2002 12 commits
-
-
Kirk Webb authored
-
Mac Newbold authored
Big changes. Add idlecheck, sdisrunning and sddeploy to configure and the make files. Install sd* above into sbin dir. Rename sddeploy.pl and isrunning.pl.
-
Leigh B. Stoller authored
-
Mac Newbold authored
Added default behavior so that when no pid/eid is given, it looks up all the nodes that haven't reported in the last 2 hours and runs it on those nodes. Also added magic around ssh calls. Fork and set an alarm, if it doesn't come back, declare the node unreachable and kill the child. (This checkin is for checkpointing, I'm now going to add this to configure so that it properly gets devel libdb.pm instead of hardcoded to the real one.)
-
Leigh B. Stoller authored
detects that at least one of the nodes in the lan includes a remote node. Of course, on links can include remote nodes, but thats okay.
-
Kirk Webb authored
1) Removed the ten iterations limit from main daemon loop I had set during testing a while back (duh). This was why slothd was mysteriously disappearing after half a day. 2) Updated the makefile to produce static slothd binaries 3) Added a locking mechanism to avoid accidentally starting two slothd's simulatneously.
-
Robert Ricci authored
Checks to make sure that all ports for an experiment are in the correct switch state. If they are enabled, they should have carrier. Note that enough code is shared between checkports and portstats that some of it could probably be moved to a library at this point.
-
Mac Newbold authored
-
Mac Newbold authored
-
Kirk Webb authored
than rc.testbed. Since rc.testbed calls cvsup, modifications to this file do not take effect until the next reboot occurs. This is why slothd was not starting on many nodes.
-
Leigh B. Stoller authored
an experiment fails to setup properly (we were leaving them running).
-
Leigh B. Stoller authored
-
- 23 May, 2002 11 commits
-
-
Robert Ricci authored
First, gives us a handy way to build all the tools, if there is ever more than one. Second, it's a workaround for a really annoying problem with configure. Since there was nothing in the tools/ directory itself, it wasn't getting created, so configure could not make tools/pcapper (since the parent directory didn't exist.)
-
Chad Barb authored
-
Kirk Webb authored
-
Kirk Webb authored
Section mentioning non-default daemons started by testbed scripts updated with slothd info, and corrections were made to the healthd info.
-
Robert Ricci authored
request.
-
Leigh B. Stoller authored
-
Leigh B. Stoller authored
-
Leigh B. Stoller authored
recommendation by Dave in email to testbed-ops).
-
Leigh B. Stoller authored
-
Leigh B. Stoller authored
-
Leigh B. Stoller authored
-
- 22 May, 2002 2 commits