db/pg_probackup
1
0
Fork 0
mirror of https://github.com/postgrespro/pg_probackup.git synced 2026-06-21 01:34:15 +02:00
Code Issues Releases Activity
3,810 Commits 217 Branches 72 Tags
17badb0c26683d7bcd50fa72b75e73ae187d2bae
Commit Graph
3810 Commits
This Branch
This Branch
All Branches
Author SHA1 Message Date
Yura Sokolov 17badb0c26 Revert "PBCKP-415 use pio for catalog_get_backup_list" 
This reverts commit f0215bbcc4.
 2022-12-18 21:32:44 +03:00
Yura Sokolov e79e5f0728 [PBCKP-416] pioOpenDir + pioDirNext 
Looks like pioListDir were no so great idea. Step to return to iterator
style.
 2022-12-18 21:32:44 +03:00
Yura Sokolov abdfdd2777 ft_strbuf_cat_path 2022-12-18 21:32:21 +03:00
Yura Sokolov d2843d0a11 fix for fio_receive_pio_err 2022-12-18 21:32:15 +03:00
Yura Sokolov 5341b9d0a7 remove pioAsyncWrite/pioAsyncError 2022-12-18 21:32:04 +03:00
Yura Sokolov 133c2275d5 fix cleanup_tablespace: it is for DB, not for BACKUP 2022-12-18 16:39:32 +03:00
Yura Sokolov 04c3159edf ft_utils: more useful staff. 2022-12-18 16:39:32 +03:00
Yura Sokolov 6bfdc0b024 adapt fobj missing argument compile-time error generation 
And it actually found error in data.c.
 2022-12-18 01:33:58 +03:00
Yura Sokolov bce53bbd31 fu_util/fobj: make missing argument compilation error 2022-12-18 01:33:38 +03:00
Yura Sokolov 33f7c10d2d fix clang warning about format(gnu_printf) 2022-12-18 01:33:38 +03:00
Yura Sokolov d0d06f3e91 fix ft_strbuf_ensure for zero-initialized ft_strbuf_t 2022-12-18 01:32:08 +03:00
Yura Sokolov 47de7fea75 ft_util: easier to parse and compose from C structs 
FT_BYTES_FOR - bytes buffer for variable
ft_bytes_dup - alloc copy of bytes
ft_bytes_shift_to/shift_must - cut to bytes from the beginning.
 2022-12-18 01:31:56 +03:00
Yura Sokolov 2b274c7a6a move pioRename from pioDrive to pioDBDrive 
it has single invocation in archive.c to move "<WAL>.ready" to "<WAL>.done"
on database host.
 2022-12-17 19:47:45 +03:00
Yura Sokolov 0852bf9abe [PBCKP-417] move sync paramter from pioClose to pioOpenWrite/Rewrite 
pioClose was not so comfortable place for. We always know at the file
opening will it be sync or not.
 2022-12-17 02:21:37 +03:00
Yura Sokolov eaead56049 use ft_strbuf_catc_zt + ft_bytes_shift_zt 2022-12-17 01:34:17 +03:00
Yura Sokolov 61c4d90316 ft_strbuf_cat_zt and ft_bytes_shift_str to easier remote operations 2022-12-17 01:28:19 +03:00
Yura Sokolov 81e2f7fe75 [PBCKP-416] do_catchup calls db_list_dir with backup_logs = false. 2022-12-16 23:22:14 +03:00
Sergey FukanchikandYuriy Sokolov f0215bbcc4 PBCKP-415 use pio for catalog_get_backup_list 2022-12-16 16:26:08 +00:00
Yura Sokolov 68113e38d8 fu_utils: add test for correct thread termination 2022-12-16 02:20:49 +03:00
Yura Sokolov f8fbbbbed3 ... 2022-12-16 01:57:41 +03:00
Yura Sokolov 47e21044c8 Merge branch 'REL_2_5' into REL_2_6 2022-12-15 22:52:12 +03:00
Yura Sokolov 764680e23f Merge branch 'REL_2_5' 2.5.11 2022-12-15 22:47:46 +03:00
Sergey Fukanchik 4b46382014 fix error message in merge_test 2022-12-15 19:31:12 +03:00
Yura Sokolov b84bed3ecb fix Makefile 2022-12-15 19:01:03 +03:00
Yura Sokolov dcf90d689c merge.c: unused buffer 2022-12-15 18:59:16 +03:00
Yura Sokolov 401bccf0b7 fu_util/fobj: fix autorelease pool cleanup in glibc 
glibc's pthread_exit runs stack unwinding for C++ destructor's handling,
and calls destructors for pthread specifics only after unwinding.

This way fobj_destroy_thread_AR is called too lately.

We have to compile with '-fexception' to register all our
attribute((cleanup)) for autorelease pools as "C++ desctructors".

Found in debug session with Sergey Fucanchik <s.fukanchik@postgrespro.ru>
 2022-12-15 18:57:40 +03:00
e.garbuz bbe41a403d Fix tests test_restore_from_fullbackup_to_new_location and test_restore_from_fullbackup_to_new_location_5_jobs 2022-12-15 13:37:15 +03:00
Yura Sokolov 8efe4d4bc7 remove fio fd-style functions and gzFile-style functions as unused 2022-12-15 11:13:15 +03:00
Yura Sokolov 692f0c0e7b SimpleXlogPageRead: replace fileExists with pioExists 2022-12-15 11:02:01 +03:00
Yura Sokolov d231fc6be5 fix pioReadFilter_pioRead 
1. without this change it attempt to another one pioRead on wrapped even if
  we already filled wbuf.
2. this lead to strange decompression error in SimpleXlogPageRead. I didn't
  investigate it properly, but this change fixes things.
 2022-12-15 11:02:01 +03:00
Yura Sokolov d175255933 use pioOpenRead and pioWrapForReSeek in SimpleXlogPageRead 2022-12-15 11:02:01 +03:00
Yura Sokolov fdab5a9149 prepare to replace fio in parsexlog.c: seekable decompressor 
Seekable decompressor is a cheat: read from underlying file starts from
beginning if we seek backward.

But it is best we can do as gzseek does the same.
 2022-12-15 10:33:11 +03:00
Yura Sokolov d3ad1f4318 remove a lot of unused fio functions 2022-12-15 07:33:34 +03:00
Yura Sokolov fdda848511 [PBCKP-404] use pioCopy instead of restore_non_data_file_internal 2022-12-15 07:04:12 +03:00
Yura Sokolov ad178a80d0 [PBCKP-404] use pioReader in restore_data_file and validate_file_pages 
reorganize common code in those functions into backup_page_iterator.
 2022-12-15 07:04:12 +03:00
Yura Sokolov 36335b6047 remove pioReadFull 
looks like all pioRead methods already tries to read all possible data
 2022-12-15 05:45:34 +03:00
Yura Sokolov f8286da568 [PBCKP-404] use pioDBWriter for restore_data_file/restore_non_data_file 2022-12-15 03:46:41 +03:00
Yura Sokolov 482c5d9933 fix pioTruncate/pioSeek signature: it is better to accept uint64_t 2022-12-15 01:01:40 +03:00
s.fukanchikandYuriy Sokolov 4f153eb70f PBCKP-403 use pio in write_page_headers 2022-12-14 23:39:18 +03:00
Alexander BurtsevandGitHub 640e7a5dcd Update README.md 2022-12-14 17:42:15 +03:00
Yura Sokolov f3a5de561a merge_test: duplicate fio_remove breakpoints to pioRemove__do 2022-12-14 16:13:25 +03:00
Alexander BurtsevandGitHub 30e3e37c7b Update README.md 2022-12-14 16:09:53 +03:00
Yura Sokolov 3f18ffab8b fix pioWriteCompressed 2022-12-14 15:59:22 +03:00
Yura Sokolov fceeb0abf2 Merge branch 'REL_2_5' into REL_2_6 2022-12-14 11:12:37 +03:00
Yura Sokolov 1f773d6d54 change pioWrite signature since it always return error on short write 
and we don't use short write in any meaningful way.
 2022-12-14 11:09:33 +03:00
Yura Sokolov 42ef001b9d add pioWriteCompressed to pioDBWriter 2022-12-14 10:55:42 +03:00
Yura Sokolov 7fa6986306 [PBCKP-395] split pioOpen to pioOpenRead and pioOpenStream 
it simplifies pioCopy as well
 2022-12-14 09:59:46 +03:00
Yura Sokolov ddb0d3329d [PBCKP-395] force async erorr fetching in pioRemoteWriteFile_pioClose 2022-12-14 09:26:01 +03:00
Yura Sokolov 36e52ff9ab fix for pioLocalFileWriter 2022-12-14 09:19:09 +03:00
Yura Sokolov e25f8ab1cc [PBCKP-395] intermediate step: forbid writting on pgFile 
All write activity is moved to pioOpenWrite and pioOpenRewrite.
pgFile will be renamed to pgReader and pgReadStream.
 2022-12-14 08:59:28 +03:00