CVS log for pvfs2/src/io/trove/trove-mgmt.c
	Help

Request diff between arbitrary revisions / Display revisions graphically

merge with Orange-Branch

checkpoint on the uuid road

move PVFS_ds_position to different types when appropriate, pass a flag around instead of using special canary values in the handle. Things will be broken for a while.

added IO byte counters for small_io
added IO byte counters for ONLY small_io and ONLY flows
minor format edit
debug fixes to usrint code

initial transactional bdb work

initial merge with Orange-Branch. much will be broken

Merge with work from this summer.  Metadata and data storage spaces can now be located on different disks.  Fixed a small bug in string encoding.  Updated test code to reflect new Trove dual-storage space API.  Genconfig updated as well as affected admin apps. In the correct branch this time.

Undo accidental commit to MAIN.  Reverted each file to pre-commit version.

Merge with work from this summer.  Metadata and data storage spaces can now be located on different disks.  Fixed a small bug in string encoding.  Updated test code to reflect new Trove dual-storage space API.  Genconfig updated as well as affected admin apps.

merged in changes from summer at LANL

Properly shutdown trove collections that have different methods than the global method.  For example, filesystems specified with directio need to call the directio cleanup to shutdown the threads involved.

Reorganized so that mgmt infrastructure is not started unless the directio
method is activated.  The resulting threads consume cpu time even when unused.

reverse merge of direct IO changes to he-branch.

merge of direct IO code to PVFS HEAD.

dspace, keyval, coll_attr dbs are now created within the environment.
A whole lot of other things try to make replication work

Reverse merged and ported to HEAD.

changes to trove layer to add direct I/O funcionality (O_DIRECT).

reverse merge latest trunk changes to small-file-branch and update msgpair
usage where needed

reverse merge of latest changes from HEAD to he branch.

Updated branch with code from HEAD

Adding a new trove method for debugging purposes called "null-aio".
Exactly like alt-aio, except that it doesn't actually do any file I/O
(ie, writes don't hit the disks and reads return garbage).  It does
perform all metadata operations and modifies file sizes (if needed)
as if data were written, though.

Possibly useful for benchmarking what the network throughput capability
of a file system would be if there were no disk overhead.

Update migration branch to current CVS version

Upgrade to current CVS version

Trove TAS modul added.

reverse merge of HEAD to WALT3 branch.

commit of trove method capabilities.  We should now be able to support multiple trove methods.  The TroveMethod option to the config file can set the method to something other than the default (dbpf).

revert style changes back to previous versions.

run maint/pvfs2indent-80col.sh on all .c files to correct style :)

- make sure that our error routines strip any known class error bits
  instead of deciding that they're unrecognized if set
- improved trove error handling
- improved pvfs2-server error handling
- fixed crash bug in our config parser that tried to strdup strings
  that might be NULL
- pvfs2-types header file simplification and cleanups

- fixed trove error codes (yet another pass; this time things look
  sane)
- give trove flush and resize operations the ability to create the
  bstream they're working with if it doesn't exist instead of
  returning an error (easily triggerred on multiple servers if a small
  amount of data is written, followed by a sync -- bstreams that
  weren't yet created fail, causing the call to fail as a whole);
  pointed out by RobL
- fixed bug in (threaded) trove that did not properly reinitialize
  condition variables after the first finalize
- improvements to the storage space and collection removal code
  (rmspace)
- extensive trove cleanups and cruft removal
- check for error codes when doing server side getattr and reading
  metafile information; stop processing if an error is reached.
  pointed out by RobL
- invalidate attr cache on write_at calls (never used)
- changed pvfs2-showcoll coll_id prints from hex to decimal

- applied Nathan's expandtab patch (expandtab-2.patch.gz)

  first referenced at:
  http://www.beowulf-underground.org/pipermail/pvfs2-developers/2004-July/000745.html

add explicit header path

added a named reserved value for fs ids (collection ids), so that we can
cleanly test for uninitialized fs id values

- added versions to collections at create time, and verification at lookup time
- cosmetic cleanups
- more verbose error reporting in some trove error paths

remove some gossip; reset keyword_array_size on finalize

move mtrace call to server if enabled.

added a coll_id arg to trove_[open|close]_context calls.  threaded trove.
things may be badly broken now as it's on by default.  even if things appear
to work, don't be fooled.  much work is left to be done.

still in progress on io client state machine.

gossip tweaks; now server is more verbose than ever.  (removed some gossip
calls added to trove that cloberred the gossip output in the server)

more gossip tweaks

okay, properly gossipified trove -- seems to work now

gossipified parts of trove; added trove handle null constant

put in a message to tell us when trove_[open|close]_context is called before
trove_initialize.  this happens in the flow init on the client side, where
trove isn't needed.

...my bad, working around the niceties of trove...

Add storage specific open/close_context methods.  Move flow initialization
in pvfs2-server to be after trove initialization since a trove context is
needed.

Added stubs for using contexts in trove.  None of this is working right now,
but the code *should* still work as before.  The idea is to get contexts into
the trove interface to play better with the job interface like the existing
bmi and flow interfaces.  This is the first step toward that end.

move trove_handle_mgmt_initialize a bit earlier and deal w any of its
(unlikely) errors.  we want to be able to use trove_init as a test for
existance of a trove space.

making progress on utilizing the existing handle management (i.e. ledger) code.

Implemented and tested (for simple cases only) collection iterate.  Had to change TROVE_ITERATE_START and TROVE_ITERATE_END values.

Monster checkin:
- removed the trove "fs" files, which are no longer in use.
- added dbpf-collection.c, which keeps up with multiple collections in use
  at one time.
  - hooks in place in dbpf implementation to theoretically handle multiple
    collections; untested (works for one collection).
- added trove-create-stress.c, a stress test for object creation.
- added TROVE_SYNC flag as an option, which is handled in most calls.
- found and removed a number of problems in error handling in dbpf calls.
- created pvfs2codecheck to identify issues in the code to be cleaned up.
- eliminated a number of memory leaks in trove dbpf.
- moved flags into top of the op structure in trove.
- implemented (but did not test) dspace_verify().

Enabled use of "malloc trace" via --with-mtrace to configure.  Currently enabled in the trove initialize; we should eventually get the initialize call moved into some higher level initialize call in the library and in the server.

Cleaned up a lot of debugging messages, etc.

Added initialization lock to trove, so it only gets initialized once.

Working on setattr, checking in changes to get fs ops to work (for now; will maybe remove later).

Added vim (and emacs) formatting option comments where missing in trove subdir.

Imported sources

Initial revision