1081 Commits

Author SHA1 Message Date
Vikas Gorur
26af4e1229 Print log message when AFR receives CHILD_UP/DOWN notification
Signed-off-by: Anand V. Avati <avati@amp.gluster.com>
2009-04-02 18:12:00 +05:30
Basavanagowda Kanur
8e621bdfff cluster/stripe to propogate ESTALE to parent.
Signed-off-by: Anand V. Avati <avati@amp.gluster.com>
2009-04-02 18:09:46 +05:30
Basavanagowda Kanur
2b63980c5d cluster/unify propogates ESTALE to parent translator, if any subvolume returns ESTALE during revalidate.
Signed-off-by: Anand V. Avati <avati@amp.gluster.com>
2009-04-02 18:09:42 +05:30
Basavanagowda Kanur
60615a0a78 cluster/replicate should propogate the ESTALE error to parent, when at least one subvolume returns ESTALE.
Signed-off-by: Anand V. Avati <avati@amp.gluster.com>
2009-04-02 18:09:32 +05:30
Basavanagowda Kanur
705c7c9bac cluster/{dht,nufa} propogate ESTALE (when at least one subvolume returns ESTALE for revalidate) to parent translator.
Signed-off-by: Anand V. Avati <avati@amp.gluster.com>
2009-04-02 18:09:28 +05:30
Basavanagowda Kanur
e8cce0bed6 cluster/{nufa,dht} fail revalidates with ESTALE, in case a stale layout is found in inode->ctx.
Signed-off-by: Anand V. Avati <avati@amp.gluster.com>
2009-04-02 18:09:23 +05:30
Anand V. Avati
36e39bf9cb Enhancements to distribute selfheal
- create missing directories instead of creating linkfiles when
  entry missing  on hashed subvol

- detect cases where there are dirs and linkfiles for a name and
  make them all dirs

Signed-off-by: Anand V. Avati <avati@amp.gluster.com>
2009-03-26 17:09:08 +05:30
Vikas Gorur
ec5a5fa208 Fixed typo in warning string in afr.c
Fixed typo.

Signed-off-by: Anand V. Avati <avati@amp.gluster.com>
2009-03-26 16:28:38 +05:30
Vikas Gorur
6e8017479f Made self heal logic more precise.
Discard earlier patch sent for the same error. This patch fixes
it more comprehensively.

This solves the spurious split-brain seen by many users.

Signed-off-by: Anand V. Avati <avati@amp.gluster.com>
2009-03-24 21:53:19 +05:30
Raghavendra G
8f590ad883 dht memory leak fix
Signed-off-by: Anand V. Avati <avati@amp.gluster.com>
2009-03-24 21:53:12 +05:30
Amar Tumballi
cf15c8ab0f fixes issue of hashing to wrong subvolumes in case when a subvolume is down
When a hashed subvolume is down, variable 'subvol' was NULL, but was sent to itransform.
This patch solves this, and readdir is sent to proper subvolume now.

Signed-off-by: Anand V. Avati <avati@amp.gluster.com>
2009-03-21 04:05:52 +05:30
Anand V. Avati
103d827004 dht_layout_dir_cmp - if xattr is missing, it is an error only if cached layout has entry for that subvolume. 2009-03-19 02:22:20 +05:30
Amar Tumballi
2d7f44e397 Moving few translators which are not well tested, are in beta stage to 'testing/' directory.
This way, users will be aware which are in 'beta' stage, and we can keep on
adding new translators (if any) seemlessly to stable codebase and once tested
can move them to proper places.

To use these translators, everyone will have to prefix 'testing/' to existing
type of translator (in volumefile)

Signed-off-by: Anand V. Avati <avati@amp.gluster.com>
2009-03-18 08:26:29 +05:30
Basavanagowda Kanur
473d02d169 implement forget for cluster/unify
inode_ctx_put() would set a list allocated on heap and would not be free()ed anywhere.

Signed-off-by: Anand V. Avati <avati@amp.gluster.com>
2009-03-13 11:01:42 +05:30
Basavanagowda Kanur
215a91c472 unify revalidate should propogate the error from any of the subvolume upto the parent, so that parent sends a fresh lookup.
unify revalidate fails if one of the subvolume returns error.

Signed-off-by: Anand V. Avati <avati@amp.gluster.com>
2009-03-13 11:01:30 +05:30
Vikas Gorur
7af8d8edbd Add extra 'volume' parameter to inodelk/entrylk calls
Signed-off-by: Anand V. Avati <avati@amp.gluster.com>
2009-03-12 16:29:45 +05:30
Anand V. Avati
fdfdd7cf05 dht_readdir_cbk - retry on same subvol if no entries match the hash
Signed-off-by: Anand V. Avati <avati@amp.gluster.com>
2009-03-09 17:40:35 +05:30
Vikas Gorur
3aff44fab8 check for fd ctx set in changelog_needed_post_op for flush() in afr
Earlier the check was in afr_flush(), which caused race conditions
with writev()

Signed-off-by: Anand V. Avati <avati@amp.gluster.com>
2009-02-27 22:49:17 +05:30
Anand V. Avati
da9664587d Skip self-heal when subvolumes are down
Signed-off-by: Anand V. Avati <avati@amp.gluster.com>
2009-02-27 16:12:07 +05:30
Anand V. Avati
596b48e14e fixing warnings in previous patch (typecasting pointer to long) 2009-02-27 02:50:48 +05:30
Vikas Gorur
2dbcb3c793 reverting nested locks in posix-locks for inodelk
Signed-off-by: Anand V. Avati <avati@amp.gluster.com>
2009-02-27 02:36:04 +05:30
Basavanagowda Kanur
f097e77ffb moved dht_hashfn_tea() to libglusterfs/hashfn.c as gf_dm_hashfn() (dm - davies-meyer).
moved dht_hashfn_tea() to libglusterfs/src/hashfn.c as gf_dm_hashfn().

Signed-off-by: Anand V. Avati <avati@amp.gluster.com>
2009-02-27 01:37:08 +05:30
Basavanagowda Kanur
3d8bc3cbaf updated copyright header to extend copyright upto 2009
updated copyright header to include 2009.

Signed-off-by: Anand V. Avati <avati@amp.gluster.com>
2009-02-26 22:18:46 +05:30
krishna
431617ef19 Layout is sorted on volume names before forming the layout ranges. But when anomalies are to be detected layout is sorted based on "start" of the range (done in the selfheal during lookup).
Signed-off-by: Anand V. Avati <avati@amp.gluster.com>
2009-02-26 19:46:23 +05:30
Vikas Gorur
c445012fa9 protect fd_ctx get/set in afr_transaction and afr.c with locks
Signed-off-by: Anand V. Avati <avati@amp.gluster.com>
2009-02-26 18:03:18 +05:30
Vikas Gorur
523d23b94c Unset fd_ctx in afr_flush if it is set
If fd_ctx is set, it means pending array needs to be decremented.
However, flush might be called many times and it used to lead
to multiple decrements. Fix is to unset fd_ctx on the first flush
received

Signed-off-by: Anand V. Avati <avati@amp.gluster.com>
2009-02-26 18:03:12 +05:30
Krishna Srinivas
417dfa48df DHT clean up. Lot of selfheal code is not applicable to newly created directory. Newly created directory will use only relavant part of DHT selfheal code - only for xattr application.
DHT clean up. Lot of selfheal code is not applicable to newly created directory. Newly created directory will use only relavant part of DHT selfheal
code - only for xattr application.

Signed-off-by: Anand V. Avati <avati@amp.gluster.com>
2009-02-25 17:47:12 +05:30
Krishna Srinivas
6609b1d37b DHT selfheal fix during lookup - create missing directories.
Signed-off-by: Anand V. Avati <avati@amp.gluster.com>
2009-02-25 16:48:45 +05:30
Anand V. Avati
d47eb5d681 DHT TEA hashfn bugfix
Signed-off-by: Anand V. Avati <avati@amp.gluster.com>
2009-02-23 21:02:27 +05:30
Basavanagowda Kanur
91df2dbcb1 cluster/ha to handle EBADFD for fd based operations as transport disconnected.
Signed-off-by: Anand V. Avati <avati@dev.gluster.com>
2009-02-19 07:08:12 -08:00
Vikas Gorur
77adf4cd64 Added all files 2009-02-18 17:36:07 +05:30