GitLab

First, a liblocsetup.pm for Linux (no target to install it yet.)
Mostly ripped off from the local-node liblocsetup .

Second, some changes to the mkemuman script, so that it will work
on Linux as well as FreeBSD.

fail (usually cause of network dropout), continue to try every minute
until it suceeds.

/var/run/slothd to /var/run/slothd.pid

Lets hope I've finally put it in the right place..

First, we'll no longer have seperate sup releases for each revision,
so the supfiles use the release 'current'

Second, since we no longer have seperate sup releases, the prepare
scripts no longer need to remove checkouts files.

* Add ability to control timeout when invoking tmcc, and whether to
  fatally die if tmcc fails. Related change was to allow more options
  to be passed into tmcc by way of an additional options argument.

* Change to pubkey handling. tmcd now returns a list of pubkeys for
  each login. The entire list is written to the authkeys files, and
  the old is backed up (with mod time preserved).

* tmcd now returns a serial number for each use to indicate when
  user info has changed (via the web pages). Only users with updated
  serial numbers get modified (and new authkeys file created). This
  should vastly reduce the passwd/authkeys churning on the ron nodes,
  which periodically check for account updates. The little db file now
  records the serial number too.

* Change to os_usermod; Change the actual password string! Was not
  doing that before.

* Conditionalize all the tmcc.foo stuff that Shashi added; don't want
  any of this on remote nodes yet.

* Remove a bunch of printfs to reduce output clutter.

actually doing that on local nodes before. I guess I'm very ssh
centric (I never type passwords) but that appears not to be the case
for most people!

well. We use a UDP packet to keep it lightweight. If it does not get
through, thats okay, obviously. The return value is just a yes/no flag
that says an update needs to run. Right now, thats just accounts.
This allows us to churn a little less on accounts.
Other cleanups.

when remote is set timeouts when trying to update accounts, and fail
more gracefully.

Return the update_accounts flag so that the client knows to
update accounts. This flag is set in node_update for remote nodes.
Once the client picks up accounts, decrement the update_accounts flag.
This functions a simple barrier (up/down counter so that clients
do not miss).

mode.

the usr_modified field of the user table, which the nodes can use
as a serial number to determine if something has changed.
Minor cleanup of last revision. Just moved things around a bit.

waiting for a reply. I'm using this as part of the (30 minute hack)
widearea node keepalive support.

node connects with UDP, update the nodes table status with 'up' and the
current time. This is the only thing that can happen when a remote
node connects with UDP (since there is no ssl). The idea is that a
daemon on the remote nodes will wake up periodically and send in a UDP
packet that says its alive. Since the idea is to be low overhead, I'm
using a UDP packet for now, which means I can run it fairly often on
all the clients, without it being too much of a drain. By its nature,
if the remote node can start up tmcc and get a udp packet out, its
probably in good shape. Maybe we will find out this does not work, but
if so I will have lost only "30 minutes". See related changes in
db/node_status.in.

Also, add the code that kicks out remote nodes that connect with tcp
but no ssl (it was commented out while I originally updated the ron
nodes with the new tmcc stuff).

as a script in init.d that gets called when the node goes into
runlevel 6 (reboot.)

(Note, it has to get 'started' earlier, when going into runlevel 3,
because if it has not created a lockfile in /var/lock/subsys,
it will not get called when shutting down.)

This is done by creating a new directory, /etc/testbed/rc.d .
This works just like /usr/local/etc/rc.d . Right now, it
only has one script, which does a 'tmcc state REBOOTING' when
called with the 'stop' argment.

While I'm at it, checked in rc.conf - it contains some testbed
specific things now, so it should be in CVS.

down on the amount of password file churning by adding new accounts
from the watchdog, but not updating mundane stuff like the gecos and
group list, except via the explicit push from emulab. Eventually this
needs a more complex solution to avoid more work than necessary,
probably by some batch queuing system that handles updates to nodes
that are down.

have checked in some other file with that long comment.

list of pubkeys per user from doaccounts. pub keys are stored in
the user_pubkeys table in the database. Currently, doaccounts is the
version support winner, with different return values for all 5
versions!

watchdog at the moment, but it will be. Right now it does boot time
stuff; issues tmcc state event so the testbed knows (REBOOTED), does
an account update to get any accounts missed while dead, then sets up
and vnodes (tunnels and such) that where supposed to be running on
the node, then issues a tmcc ISUP event.

After that, goes into a loop doing periodic account update. At some
point it would be good to look for stale vnodes (that could not be
torn down because of network connectivity problems), but there are
some race conditions that I need to work out first.

node reboots. Currently, starts the watchdog daemon.

instead of account update. Doing an account update at boot is nice,
but hardly sufficient.

instead of waiting. This option used by new watchdog daemon that
runs on the RON nodes. No point in waiting if another is already
happening.

more thought, so just leave the directories.

files are generated.

currently assigned virtual nodes, say when a node reboots.

- killing off any currently running slothd (potentially started by rc.testbed)
- removing the lockfile
- staring the daemon

Slothd should now startup properly after a machine crash.

bit set.  It appears that cvs doesn't notice just permission changes (?)