GitLab

wrap. Also spellchecked while I was in there.

Added site variables ('sitevars').
These are stored in the sitevariables database table.
Each one has a name, a description (NOT OPTIONAL!), a
default value, and a current value.
If the current value is NULL, the default value is used.

Soon, a mechanism will be added to the install process to
ensure all needed site variables exist before installing;
more on that when it is committed.

 - Added 'editsitevars.php3' page, accessable to admins
   via the 'Edit Site Variables' menu option.

 - Added 'setsitevar' script,
   an interface for listing, viewing in detail, and setting
   site variables.

 - Web interface now uses 'web/nologins' and 'web/message'
   instead of one-off database tables.

NOTE that setting a variable to the default value and
setting a variable to a value which is string-identical
are NOT the same thing.
(This doesn't matter yet, but when we push default values out to
 remote sites as part of our install, it will.)

soon be used by admins only to grant swappability. For now, it is open to
anyone who has expt modify access as well.

The change is visible everywhere SHOWEXP is called, like showexp.php3 for
instance.

Also introduced toggle.php, a generic toggle script, similar to
adminmode.php, but useful for just about anything. To add something, just
add it to the list, specify the valid values it can have, and add its
permission check and action. Now you don't need a different php page for
each toggle switch. This page is already set up to take over from
adminmode.php, but I didn't change things over just yet.

(Should this go in the changelog?)

to call ISADMIN(uid) with.

hang and get restarted (i.e., join again with a different ID).
Note that this count is only used for stats reporting so it isn't vital
that its value be correct.

Sort by vname is now the default sort, and you can sort by node_id (the
old sort), and by vname, by clicking on the column title, as usual.

and /share, all rooted in /netbed. This is convenient! New images have
symlinks pointing /{proj,groups,share} into /netbed.

* Allow users to set default group and default shell via web page and
  pass out to nodes when creating accounts.

not ever need to create a OSID directly, so do not give them the link.
Will need a mechanism to allow this later for Eric and others with a
clue. Also add pointer to customos section of the tutorial at the top
of the page.

Changed sitevariables table to use 'varchar(255)' for
name column instead of 'text'.
Made name column primary key.

This is pretty much a comment field for now;
NULL is a safe value, but if a node is a RON node
"ron:" followed by the name of the node in the ron
hosts table (in the external-to-emulab ron db) is
the preferred value. (This may facilitate automatic
db synchronization later.)

none is set.

Set TBAUTOMAILEMAIL in defs-default to testbed-automail@flux.utah.edu

Change quotamail to use bcc automail instead of tb-ops. Still says from
tb-ops, though, so that they reply to the right place.

Changed widearea_nodeinfo address columns to tinytext,
to make them consistent with users table.

Added 'sitevariables' table.

This is the table prompted by Mac's email yesterday, which
can contain site-specific configuration options.

'default' _must_ be set for each row, and will be used if
'value' is NULL.

'description' _should_ be set for each row.

The table doesn't have a primary key, but is indexed to
32 characters on the "name" column, so it should be plenty efficient.

libdb functions and web interface forthcoming...

* Add a web page to recreate an image from a node and an existing
  image descriptor. Need to add a "kill runnin frisbee" function to
  create_image though (well, need this in general cause I always get
  bit by it). Might need some locking too (experiment should be locked
  down).

* Append Eric suggestion to bug tracking TODO item (already existed).

cron daily, on the fs node (ops) as root. Uses lots of configure variables
so that it does the right thing in any installation, without any
customization.

One possible caveat regarding quotas: If FS_NODE != USERNODE, they don't
have a login on the fs machine. So checking their quota won't work unless
the rpc.rquotad(8) daemon is running on FS_NODE, which we currently don't
do.

In order to do this right, I had to add a new configure var,
FS_WITH_QUOTAS, that has a space separated list of file systems that have
quotas enabled. (Ie the default is 'FS_WITH_QUOTAS="/q /users"'.) It
doesn't have any default, since I couldn't come up with a reasonable one.
All the defs files have been updated appropriately to define this new
variable.

When a user tries to create an imageid without
specifying a node to suck the image from, they are given
a warning screen explaining their situation,
with "confirm" and "back" buttons.

(also made the confirmation mechanism general so if we
 need to add more warning conditions, it'll be no problem.)

Make use of all available memory via the -M option to frisbee

Changed form respit error message:
"no such group" --> "group 'foo' is not in project 'bar'".

hopefully never to return again!

A couple of the changes Jay wanted:

 - move 'Node to Create Image From' above OS Features
 - make all Node Types default to checked

Also cleaned up footnotes a touch.

1. Implement PREQUEST message which passes a bit map of desired blocks.
   We still use the REQUEST message (start block + number of blocks) for
   full chunk requests as that is more efficient.  This message also
   includes a flag indicating whether it is a retry of a request we
   originally made or not.  This gives the server more accurate loss info.

2. More stats and tracing goo.


Frisbee client:

1. Add 'C' and 'W' command line options to specify amount of memory
   for chunk buffers (network buffering) and for write buffers (disk
   buffering).  The Emulab frisbee startup script uses these to partition
   up all the available memory on a machine.  Previously we were just
   using a fixed ~128MB even though our machines have 256 or 512MB of
   memory.  Also add the 'M' option which specifies the overall memory,
   internally dividing it up between chunk buffers and write buffers.

2. Add 'S' command line option to explicitly specify the server.  This
   allows us to make a feeb...um, "lightweight" authentication check
   on incoming messages.

3. Use the common BlockMap data struct to track which pieces of a chunk
   we have received.  This is easily inverted to make PREQUESTS and it is
   also smaller than the older byte-per-block technique.

4. Allow partial request-ahead.  Previously, we only issued request-ahead
   if there were enough empty chunk buffers for a maximum (2) request-ahead.

Frisbee server:

1. Use BlockMap for workQ elements.  An easy way to allow a complete merge
   of incoming requests with existing ones.

2. Check for overlap of incoming requests with the request currently
   being serviced.  This happens surprisingly often.

3. Dubious: burst gap becomes burst interval.  The latter takes into
   account the time required to read data, etc., in other words, we now
   have variable-sized gaps and put out bursts at specific times rather
   than having fixed gaps and putting out bursts at variable times.
   This gives us more accurate pacing over shorter time periods.  I
   thought this might be important for dynamic pacing.

4. Add 'W' command line option to specify a target bandwidth.  Frisbeed
   will use this to calculate a burst size/interval.

5. Rewrote the dynamic pacing code.  It is now easily as bad as before
   if not worse.  But it does have fewer magic constants!  Needs to be
   redone by someone who understands the TCP-friendly rate equation.

Imagezip:

1. add 'R' option to specify one or more partitions for which to force
   raw (naive) compression even if the FS format is understood.  Useful
   for benchmarking.

2. add 'D' option to allow "dangerous" writes.  In this mode, we don't
   do the fsync's or retries of failed writes.  Overrides the hack we put
   in for NFS.  Use this if writing to a local filesystem (or /dev/null).

3. Eliminate an extra copy of every chunk header.

Imageunzip:

1. Eliminate extra copy of decompressed data that we were doing between
   the single decompression buffer and the disk buffers.  Helps on slow
   machines (like gatech's 300Mhz machines with 66MHz memory bus).

2. Allow dynamic number of variable-sized write buffers.  Total memory
   not to exceed the writebufmem limit.  Previously we had a small number
   of fixed-size (256K) buffers.

3. Add debugging 'C' option to just compute a single CRC of the decompressed
   image.  Back-ported to older imageunzip and used to make sure my write
   buffer changes were correct.  Maybe handy for similar massive changes
   in the future.

request counter.

It means nothing and just takes up space and may cause confusion, and is
easy to reenable as soon as we want to.

Added processor type, location, inet connection type to widearea listing.

blog.

start, but was a dope!