glusterfs.git/xlators/storage/posix/src, branch v3.1.0qa3

fixes for spurious entry self-heal in afr

2010-07-29T06:58:09+00:00

Signed-off-by: Pranith Kumar K 
Signed-off-by: Anand V. Avati 

BUG: 974 (Spurious self-heal?)
URL: http://bugs.gluster.com/cgi-bin/bugzilla3/show_bug.cgi?id=974

some check added to the variables after GF_CALLOC

2010-07-28T10:34:48+00:00

handles some NULL dereference problems
(reported by clang when ran with code where '#define GF_CALLOC NULL').

Signed-off-by: Amar Tumballi 
Signed-off-by: Anand V. Avati 

BUG: 966 (NULL check for avoiding NULL dereferencing of pointers..)
URL: http://bugs.gluster.com/cgi-bin/bugzilla3/show_bug.cgi?id=966

proper way to do defrag of a mountpoint

2010-07-14T21:44:26+00:00

Usage: "glusterfs-defrag "

there are new features added to distribute with this patch:

* bash# getfattr -n trusted.distribute.linkinfo 
  Gives output in the format ":", if there is a
  linkfile present, where hostname is server, directory is backend
  directory where the actual linkfile is present.

* bash# getfattr -n trusted.glusterfs.pathinfo 
  Gives layout information if directory, and info about actual
  location of file in backend servers, in the form of
  'hostname:directory'. (TODO: should extend it to all xlators)

* bash# getfattr -n trusted.distribute.fix.layout 
  scales out the directory layout in distribute, so when new
  servers are added, layouts are fixed to include them.

* removed 'scale-n-defrag.sh' as its no more required.

* changed 'defrag.sh' to have a feature to specify target bricks,
  so defrag happens to only those nodes. moved the file to
  'glusterfs-defrag', which now gets installed to '${prefix}/bin'

* storage/posix:
  takes new option 'hostname' so it can resolve to same hostname
  given during 'gluster volume create'.
  with 'trusted.glusterfs.pathinfo' posix returns the value in
  'hostname:real_path' format.

Signed-off-by: Amar Tumballi 
Signed-off-by: Anand V. Avati 

BUG: 1073 ('gluster defrag ' fails in mainline)
URL: http://bugs.gluster.com/cgi-bin/bugzilla3/show_bug.cgi?id=1073

removed 'fop->checksum' from codebase as its not required anymore

2010-07-06T12:58:56+00:00

Signed-off-by: Amar Tumballi 
Signed-off-by: Anand V. Avati 

BUG: 734 (keep only the working/usable code in build tree to focus more on development)
URL: http://bugs.gluster.com/cgi-bin/bugzilla3/show_bug.cgi?id=734

NULL dereference fixes in code base after running with 'clang'

2010-07-02T12:17:03+00:00

* 212 logical (NULL deref/divide by zero) errors reduced to 28
  (27 of them in contrib/ and lex part of codebase, 1 is invalid)
* 11 API errors reduced to 0

Signed-off-by: Amar Tumballi 
Signed-off-by: Anand V. Avati 

BUG: 966 (NULL check for avoiding NULL dereferencing of pointers..)
URL: http://bugs.gluster.com/cgi-bin/bugzilla3/show_bug.cgi?id=966

mem-types: include *-mem-types.h in noinst_HEADERS variables in all Makefile.am

2010-06-08T12:25:07+00:00

Signed-off-by: Anand V. Avati 
Signed-off-by: Anand V. Avati 

BUG: 329 (Replacing memory allocation functions with mem-type functions)
URL: http://bugs.gluster.com/cgi-bin/bugzilla3/show_bug.cgi?id=329

dynamic volume changes for graph replacement

2010-06-07T18:58:50+00:00

Signed-off-by: Anand V. Avati 
Signed-off-by: Anand V. Avati 

BUG: 971 (dynamic volume management)
URL: http://bugs.gluster.com/cgi-bin/bugzilla3/show_bug.cgi?id=971

OS X: adjustments, minor fixes to eliminate warnings

2010-05-21T07:31:51+00:00

Signed-off-by: Csaba Henk 
Signed-off-by: Anand V. Avati 

BUG: 361 (GlusterFS 3.0 should work on Mac OS/X)
URL: http://bugs.gluster.com/cgi-bin/bugzilla3/show_bug.cgi?id=361

added features/mac-compat

2010-05-13T19:12:08+00:00

Signed-off-by: Csaba Henk 
Signed-off-by: Anand V. Avati 

BUG: 800 (Problem on OSX with NFS and CIFS exports)
URL: http://bugs.gluster.com/cgi-bin/bugzilla3/show_bug.cgi?id=800

posix: Support thread-safe vectored writes/reads

2010-05-05T11:36:05+00:00

..by maintaining internal offsets and using pwrite/pread
instead of writev/read.

The recent io-threads change is causing concurrent writes to
a file from different io-threads. This results in a race between
two threads both of which are trying to call lseek-writev. The
lseek-writev sequence should be a critical section. This is where
pwrite syscall helps. pwrite also requires an offset to be given so
for supporting pwrites for multiple iovecs we maintain an offset
within posix for the duration of the write fop and pass this internal
offset to pwrite.

The same logic also applies to read, only difference is that we do
not need to read multiple iovecs but only provide an atomic
lseek-read sequence.

Signed-off-by: Shehjar Tikoo 
Signed-off-by: Anand V. Avati 

BUG: 883 (Data corruption due to thread unsafe reads and writes)
URL: http://bugs.gluster.com/cgi-bin/bugzilla3/show_bug.cgi?id=883