summaryrefslogtreecommitdiffstats
path: root/xlators/cluster/afr
Commit message (Collapse)AuthorAgeFilesLines
* Changed xattr format of afr changelog to support adding and removing of ↵Vikas Gorur2009-04-1611-287/+322
| | | | | | subvolumes while keeping existing data. Signed-off-by: Anand V. Avati <avati@amp.gluster.com>
* Send statbuf from the same subvolume, even across revalidates.Vikas Gorur2009-04-131-5/+14
| | | | Signed-off-by: Anand V. Avati <avati@amp.gluster.com>
* update cluster/afr with new readv writev prototypesAnand V. Avati2009-04-125-17/+20
| | | | Signed-off-by: Anand V. Avati <avati@amp.gluster.com>
* Compulsorily do self heal if file sizes differ.Vikas Gorur2009-04-096-23/+100
| | | | | | | | | If file sizes differ, then compulsorily do self-heal. If no 'wise' sources are found, then pick a 'fool' with the biggest file size. If even 'fools' aren't found, pick the 'innocent' source with the biggest file size. Signed-off-by: Anand V. Avati <avati@amp.gluster.com>
* Handle files which have no pending xattrs at all.Vikas Gorur2009-04-091-1/+27
| | | | | | | | If a pending xattr key is non-existent on a file (call such files 'ignorant'), make all other non-ignorant subvolumes point towards the ignorant one. Signed-off-by: Anand V. Avati <avati@amp.gluster.com>
* Erase xattr during self-heal based on original dict.Vikas Gorur2009-04-095-10/+27
| | | | | | | | Decrement xattr during self-heal based on the original dict instead of pending_matrix, as the pending_matrix might have been altered later. Signed-off-by: Anand V. Avati <avati@amp.gluster.com>
* Fix minor bug in afr_open_cbk.Vikas Gorur2009-04-081-3/+7
| | | | Signed-off-by: Anand V. Avati <avati@amp.gluster.com>
* Fix in changelog logic.Vikas Gorur2009-04-076-41/+254
| | | | | | If a writev fails, remember it by marking it in the fd context. Signed-off-by: Anand V. Avati <avati@amp.gluster.com>
* Consider a subvolume dead if an fop fails on itVikas Gorur2009-04-073-17/+19
| | | | | | | | | Transaction fops earlier called afr_transaction_child_died only if an fop failed due to ENOTCONN or EBADFD. Now they consider a child dead regardless of the reason for failure. This handles cases such as ENOSPC. Signed-off-by: Anand V. Avati <avati@amp.gluster.com>
* Fix in return value of afr_sh_mark_sourcesVikas Gorur2009-04-063-10/+26
| | | | | | | | | | | | afr_sh_mark_sources now returns: -1 if two wise subvols conflict (split-brain) 0 if all subvols are innocent (no self-heal needed) >0 if sources found Also, changes to callers of afr_sh_mark_sources to handle return value properly. Signed-off-by: Anand V. Avati <avati@amp.gluster.com>
* Made afr inode context a 64-bit packed value instead of a structure.Vikas Gorur2009-04-034-313/+205
| | | | Signed-off-by: Anand V. Avati <avati@amp.gluster.com>
* Load balance read operations among subvolumes in afrVikas Gorur2009-04-024-110/+513
| | | | Signed-off-by: Anand V. Avati <avati@amp.gluster.com>
* Defined afr_inode_ctx_t structure.Vikas Gorur2009-04-022-11/+76
| | | | | | | | Notification of a split-brain situation, which was earlier signalled by the mere presence of inode context is now signalled by the 'split_brain' member in the structure. Signed-off-by: Anand V. Avati <avati@amp.gluster.com>
* Print log message when AFR receives CHILD_UP/DOWN notificationVikas Gorur2009-04-021-0/+6
| | | | Signed-off-by: Anand V. Avati <avati@amp.gluster.com>
* cluster/replicate should propogate the ESTALE error to parent, when at least ↵Basavanagowda Kanur2009-04-021-4/+14
| | | | | | one subvolume returns ESTALE. Signed-off-by: Anand V. Avati <avati@amp.gluster.com>
* Fixed typo in warning string in afr.cVikas Gorur2009-03-261-1/+1
| | | | | | Fixed typo. Signed-off-by: Anand V. Avati <avati@amp.gluster.com>
* Made self heal logic more precise.Vikas Gorur2009-03-241-22/+233
| | | | | | | | | 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>
* Add extra 'volume' parameter to inodelk/entrylk callsVikas Gorur2009-03-126-20/+37
| | | | Signed-off-by: Anand V. Avati <avati@amp.gluster.com>
* check for fd ctx set in changelog_needed_post_op for flush() in afrVikas Gorur2009-02-272-95/+49
| | | | | | | Earlier the check was in afr_flush(), which caused race conditions with writev() Signed-off-by: Anand V. Avati <avati@amp.gluster.com>
* fixing warnings in previous patch (typecasting pointer to long)Anand V. Avati2009-02-271-1/+1
|
* reverting nested locks in posix-locks for inodelkVikas Gorur2009-02-271-0/+2
| | | | Signed-off-by: Anand V. Avati <avati@amp.gluster.com>
* updated copyright header to extend copyright upto 2009Basavanagowda Kanur2009-02-2618-18/+18
| | | | | | updated copyright header to include 2009. Signed-off-by: Anand V. Avati <avati@amp.gluster.com>
* protect fd_ctx get/set in afr_transaction and afr.c with locksVikas Gorur2009-02-262-16/+31
| | | | Signed-off-by: Anand V. Avati <avati@amp.gluster.com>
* Unset fd_ctx in afr_flush if it is setVikas Gorur2009-02-261-0/+2
| | | | | | | | | 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>
* Added all filesVikas Gorur2009-02-1820-0/+14019
generated by cgit (git 2.34.1) at 2024-11-16 15:29:53 +0000