Just to trigger #8
Closed
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
Can one of the admins verify this patch? |
mdiep25
force-pushed
the
mdiep-test
branch
7 times, most recently
from
56cc1c0 to
a0a5a6d
Compare
Signed-off-by: Minh Diep <[email protected]>
|
testing |
|
Can one of the admins verify this patch? |
mdiep25
force-pushed
the
mdiep-test
branch
8 times, most recently
from
5f3f73c to
d4ca69b
Compare
|
just one |
mdiep25
force-pushed
the
mdiep-test
branch
4 times, most recently
from
0a58fdf to
951b00a
Compare
mdiep25
force-pushed
the
mdiep-test
branch
2 times, most recently
from
41a6dad to
d4cc914
Compare
Signed-off-by: Minh Diep <[email protected]>
jolivier23
added a commit
that referenced
this pull request
Mar 23, 2020
#8) clang-format doesn't work on older clang compilers. Rather than skipping the step, instead use fallback of Mozilla if the version isn't suitable. Signed-off-by: Jeff Olivier <[email protected]>
jolivier23
pushed a commit
that referenced
this pull request
Apr 21, 2022
#8780) It is possible that the error is not reset to 0 when a process is retrieving the private event in a multi-threaded environment since another thread might have processed the event completion. add a lock on daos_event_t for events that are not part of an event queue, including the thread private event. This is required because in a multithreaded environment like dfuse, the event can be completed and the error set non atomically, while the main thread sees the completion but before the error has actually set, causing the owner of the event to reset the event error, with the other thread setting the error after. this affects future use of the event that launches with ev_error already set to the previous error. Signed-off-by: Mohamad Chaarawi <[email protected]>
jolivier23
pushed a commit
that referenced
this pull request
Apr 21, 2022
#8779) It is possible that the error is not reset to 0 when a process is retrieving the private event in a multi-threaded environment since another thread might have processed the event completion. add a lock on daos_event_t for events that are not part of an event queue, including the thread private event. This is required because in a multithreaded environment like dfuse, the event can be completed and the error set non atomically, while the main thread sees the completion but before the error has actually set, causing the owner of the event to reset the event error, with the other thread setting the error after. this affects future use of the event that launches with ev_error already set to the previous error. Signed-off-by: Mohamad Chaarawi <[email protected]>
liw
added a commit
that referenced
this pull request
Jun 7, 2022
Makito and Samir observed the following assertion failure after restarting engines. #0 raise () from /lib64/libc.so.6 #1 abort () from /lib64/libc.so.6 #2 __assert_fail_base () from /lib64/libc.so.6 #3 __assert_fail () from /lib64/libc.so.6 #4 pool_map_get_version (map=0x0) at src/common/pool_map.c:2852 #5 ds_pool_get_version (pool=0x7f0ca063c690, pool=0x7f0ca063c690) at src/include/daos_srv/pool.h:296 #6 pc=rpc@entry=0x7f0ca0998d30, p_rpt=p_rpt@entry=0x7f0ca83a77b0) at src/rebuild/srv.c:2101 #7 rebuild_tgt_scan_handler (rpc=0x7f0ca0998d30) at src/rebuild/scan.c:954 #8 crt_handle_rpc (arg=0x7f0ca0998d30) at src/cart/crt_rpc.c:1654 #9 ABTD_ythread_func_wrapper (p_arg=0x7f0ca83a78a0) at arch/abtd_ythread.c:21 #10 make_fcontext () from /usr/lib64/libabt.so.1 #11 ?? () The ds_pool_get_version call passed a NULL map argument to pool_map_get_version. The ds_pool.sp_map field may be NULL after the pool is started but before the pool receives the initial pool map from the pool service. This patch fixes ds_pool_get_version to return 0, which is less than all valid pool map versions, when sp_map is NULL, resulting in rebuild retries like this: Rebuild [queued] (pool=3bf68c9c ver=2) tgts=2 Rebuild [started] (pool 3bf68c9c ver=2) Rebuild [failed] (pool 3bf68c9c ver=2 status=DER_BUSY(-1012): 'Device or resource busy') Rebuild [queued] (pool=3bf68c9c ver=2) tgts=2 Rebuild [started] (pool 3bf68c9c ver=2) Rebuild [failed] (pool 3bf68c9c ver=2 status=DER_BUSY(-1012): 'Device or resource busy') Rebuild [queued] (pool=3bf68c9c ver=2) tgts=2 Rebuild [started] (pool 3bf68c9c ver=2) Rebuild [scanning] (pool 3bf68c9c ver=2, toberb_obj=0, rb_obj=0, [...] Rebuild [scanning] (pool 3bf68c9c ver=2, toberb_obj=0, rb_obj=0, [...] Rebuild [completed] (pool 3bf68c9c ver=2, toberb_obj=0, rb_obj=0,[...] Target[2] (rank 2 idx 0 status 16 ver 1) is excluded. Also, this patch removes some rebuild code that handles NULL ds_pool.sp_group fields. Those can not happen as we always initialize sp_group (as well as sp_iv_ns) before putting a ds_pool object into the LRU. Signed-off-by: Li Wei <[email protected]>
shimizukko
pushed a commit
that referenced
this pull request
Jun 28, 2022
Makito and Samir observed the following assertion failure after restarting engines. #0 raise () from /lib64/libc.so.6 #1 abort () from /lib64/libc.so.6 #2 __assert_fail_base () from /lib64/libc.so.6 #3 __assert_fail () from /lib64/libc.so.6 #4 pool_map_get_version (map=0x0) at src/common/pool_map.c:2852 #5 ds_pool_get_version (pool=0x7f0ca063c690, pool=0x7f0ca063c690) at src/include/daos_srv/pool.h:296 #6 pc=rpc@entry=0x7f0ca0998d30, p_rpt=p_rpt@entry=0x7f0ca83a77b0) at src/rebuild/srv.c:2101 #7 rebuild_tgt_scan_handler (rpc=0x7f0ca0998d30) at src/rebuild/scan.c:954 #8 crt_handle_rpc (arg=0x7f0ca0998d30) at src/cart/crt_rpc.c:1654 #9 ABTD_ythread_func_wrapper (p_arg=0x7f0ca83a78a0) at arch/abtd_ythread.c:21 #10 make_fcontext () from /usr/lib64/libabt.so.1 #11 ?? () The ds_pool_get_version call passed a NULL map argument to pool_map_get_version. The ds_pool.sp_map field may be NULL after the pool is started but before the pool receives the initial pool map from the pool service. This patch fixes ds_pool_get_version to return 0, which is less than all valid pool map versions, when sp_map is NULL, resulting in rebuild retries like this: Rebuild [queued] (pool=3bf68c9c ver=2) tgts=2 Rebuild [started] (pool 3bf68c9c ver=2) Rebuild [failed] (pool 3bf68c9c ver=2 status=DER_BUSY(-1012): 'Device or resource busy') Rebuild [queued] (pool=3bf68c9c ver=2) tgts=2 Rebuild [started] (pool 3bf68c9c ver=2) Rebuild [failed] (pool 3bf68c9c ver=2 status=DER_BUSY(-1012): 'Device or resource busy') Rebuild [queued] (pool=3bf68c9c ver=2) tgts=2 Rebuild [started] (pool 3bf68c9c ver=2) Rebuild [scanning] (pool 3bf68c9c ver=2, toberb_obj=0, rb_obj=0, [...] Rebuild [scanning] (pool 3bf68c9c ver=2, toberb_obj=0, rb_obj=0, [...] Rebuild [completed] (pool 3bf68c9c ver=2, toberb_obj=0, rb_obj=0,[...] Target[2] (rank 2 idx 0 status 16 ver 1) is excluded. Also, this patch removes some rebuild code that handles NULL ds_pool.sp_group fields. Those can not happen as we always initialize sp_group (as well as sp_iv_ns) before putting a ds_pool object into the LRU. Skip-test: true Signed-off-by: Li Wei <[email protected]>
liw
added a commit
that referenced
this pull request
Jul 6, 2022
Makito and Samir observed the following assertion failure after restarting engines. #0 raise () from /lib64/libc.so.6 #1 abort () from /lib64/libc.so.6 #2 __assert_fail_base () from /lib64/libc.so.6 #3 __assert_fail () from /lib64/libc.so.6 #4 pool_map_get_version (map=0x0) at src/common/pool_map.c:2852 #5 ds_pool_get_version (pool=0x7f0ca063c690, pool=0x7f0ca063c690) at src/include/daos_srv/pool.h:296 #6 pc=rpc@entry=0x7f0ca0998d30, p_rpt=p_rpt@entry=0x7f0ca83a77b0) at src/rebuild/srv.c:2101 #7 rebuild_tgt_scan_handler (rpc=0x7f0ca0998d30) at src/rebuild/scan.c:954 #8 crt_handle_rpc (arg=0x7f0ca0998d30) at src/cart/crt_rpc.c:1654 #9 ABTD_ythread_func_wrapper (p_arg=0x7f0ca83a78a0) at arch/abtd_ythread.c:21 #10 make_fcontext () from /usr/lib64/libabt.so.1 #11 ?? () The ds_pool_get_version call passed a NULL map argument to pool_map_get_version. The ds_pool.sp_map field may be NULL after the pool is started but before the pool receives the initial pool map from the pool service. This patch fixes ds_pool_get_version to return 0, which is less than all valid pool map versions, when sp_map is NULL, resulting in rebuild retries like this: Rebuild [queued] (pool=3bf68c9c ver=2) tgts=2 Rebuild [started] (pool 3bf68c9c ver=2) Rebuild [failed] (pool 3bf68c9c ver=2 status=DER_BUSY(-1012): 'Device or resource busy') Rebuild [queued] (pool=3bf68c9c ver=2) tgts=2 Rebuild [started] (pool 3bf68c9c ver=2) Rebuild [failed] (pool 3bf68c9c ver=2 status=DER_BUSY(-1012): 'Device or resource busy') Rebuild [queued] (pool=3bf68c9c ver=2) tgts=2 Rebuild [started] (pool 3bf68c9c ver=2) Rebuild [scanning] (pool 3bf68c9c ver=2, toberb_obj=0, rb_obj=0, [...] Rebuild [scanning] (pool 3bf68c9c ver=2, toberb_obj=0, rb_obj=0, [...] Rebuild [completed] (pool 3bf68c9c ver=2, toberb_obj=0, rb_obj=0,[...] Target[2] (rank 2 idx 0 status 16 ver 1) is excluded. Also, this patch removes some rebuild code that handles NULL ds_pool.sp_group fields. Those can not happen as we always initialize sp_group (as well as sp_iv_ns) before putting a ds_pool object into the LRU. Signed-off-by: Li Wei <[email protected]>
jolivier23
pushed a commit
that referenced
this pull request
Jul 19, 2022
Makito and Samir observed the following assertion failure after restarting engines. #0 raise () from /lib64/libc.so.6 #1 abort () from /lib64/libc.so.6 #2 __assert_fail_base () from /lib64/libc.so.6 #3 __assert_fail () from /lib64/libc.so.6 #4 pool_map_get_version (map=0x0) at src/common/pool_map.c:2852 #5 ds_pool_get_version (pool=0x7f0ca063c690, pool=0x7f0ca063c690) at src/include/daos_srv/pool.h:296 #6 pc=rpc@entry=0x7f0ca0998d30, p_rpt=p_rpt@entry=0x7f0ca83a77b0) at src/rebuild/srv.c:2101 #7 rebuild_tgt_scan_handler (rpc=0x7f0ca0998d30) at src/rebuild/scan.c:954 #8 crt_handle_rpc (arg=0x7f0ca0998d30) at src/cart/crt_rpc.c:1654 #9 ABTD_ythread_func_wrapper (p_arg=0x7f0ca83a78a0) at arch/abtd_ythread.c:21 #10 make_fcontext () from /usr/lib64/libabt.so.1 #11 ?? () The ds_pool_get_version call passed a NULL map argument to pool_map_get_version. The ds_pool.sp_map field may be NULL after the pool is started but before the pool receives the initial pool map from the pool service. This patch fixes ds_pool_get_version to return 0, which is less than all valid pool map versions, when sp_map is NULL, resulting in rebuild retries like this: Rebuild [queued] (pool=3bf68c9c ver=2) tgts=2 Rebuild [started] (pool 3bf68c9c ver=2) Rebuild [failed] (pool 3bf68c9c ver=2 status=DER_BUSY(-1012): 'Device or resource busy') Rebuild [queued] (pool=3bf68c9c ver=2) tgts=2 Rebuild [started] (pool 3bf68c9c ver=2) Rebuild [failed] (pool 3bf68c9c ver=2 status=DER_BUSY(-1012): 'Device or resource busy') Rebuild [queued] (pool=3bf68c9c ver=2) tgts=2 Rebuild [started] (pool 3bf68c9c ver=2) Rebuild [scanning] (pool 3bf68c9c ver=2, toberb_obj=0, rb_obj=0, [...] Rebuild [scanning] (pool 3bf68c9c ver=2, toberb_obj=0, rb_obj=0, [...] Rebuild [completed] (pool 3bf68c9c ver=2, toberb_obj=0, rb_obj=0,[...] Target[2] (rank 2 idx 0 status 16 ver 1) is excluded. Also, this patch removes some rebuild code that handles NULL ds_pool.sp_group fields. Those can not happen as we always initialize sp_group (as well as sp_iv_ns) before putting a ds_pool object into the LRU. Signed-off-by: Li Wei <[email protected]>
liw
added a commit
that referenced
this pull request
Jul 20, 2022
Makito and Samir observed the following assertion failure after restarting engines. #0 raise () from /lib64/libc.so.6 #1 abort () from /lib64/libc.so.6 #2 __assert_fail_base () from /lib64/libc.so.6 #3 __assert_fail () from /lib64/libc.so.6 #4 pool_map_get_version (map=0x0) at src/common/pool_map.c:2852 #5 ds_pool_get_version (pool=0x7f0ca063c690, pool=0x7f0ca063c690) at src/include/daos_srv/pool.h:296 #6 pc=rpc@entry=0x7f0ca0998d30, p_rpt=p_rpt@entry=0x7f0ca83a77b0) at src/rebuild/srv.c:2101 #7 rebuild_tgt_scan_handler (rpc=0x7f0ca0998d30) at src/rebuild/scan.c:954 #8 crt_handle_rpc (arg=0x7f0ca0998d30) at src/cart/crt_rpc.c:1654 #9 ABTD_ythread_func_wrapper (p_arg=0x7f0ca83a78a0) at arch/abtd_ythread.c:21 #10 make_fcontext () from /usr/lib64/libabt.so.1 #11 ?? () The ds_pool_get_version call passed a NULL map argument to pool_map_get_version. The ds_pool.sp_map field may be NULL after the pool is started but before the pool receives the initial pool map from the pool service. This patch fixes ds_pool_get_version to return 0, which is less than all valid pool map versions, when sp_map is NULL, resulting in rebuild retries like this: Rebuild [queued] (pool=3bf68c9c ver=2) tgts=2 Rebuild [started] (pool 3bf68c9c ver=2) Rebuild [failed] (pool 3bf68c9c ver=2 status=DER_BUSY(-1012): 'Device or resource busy') Rebuild [queued] (pool=3bf68c9c ver=2) tgts=2 Rebuild [started] (pool 3bf68c9c ver=2) Rebuild [failed] (pool 3bf68c9c ver=2 status=DER_BUSY(-1012): 'Device or resource busy') Rebuild [queued] (pool=3bf68c9c ver=2) tgts=2 Rebuild [started] (pool 3bf68c9c ver=2) Rebuild [scanning] (pool 3bf68c9c ver=2, toberb_obj=0, rb_obj=0, [...] Rebuild [scanning] (pool 3bf68c9c ver=2, toberb_obj=0, rb_obj=0, [...] Rebuild [completed] (pool 3bf68c9c ver=2, toberb_obj=0, rb_obj=0,[...] Target[2] (rank 2 idx 0 status 16 ver 1) is excluded. Also, this patch removes some rebuild code that handles NULL ds_pool.sp_group fields. Those can not happen as we always initialize sp_group (as well as sp_iv_ns) before putting a ds_pool object into the LRU. Signed-off-by: Li Wei <[email protected]>
jolivier23
pushed a commit
that referenced
this pull request
Jul 21, 2022
Makito and Samir observed the following assertion failure after restarting engines. #0 raise () from /lib64/libc.so.6 #1 abort () from /lib64/libc.so.6 #2 __assert_fail_base () from /lib64/libc.so.6 #3 __assert_fail () from /lib64/libc.so.6 #4 pool_map_get_version (map=0x0) at src/common/pool_map.c:2852 #5 ds_pool_get_version (pool=0x7f0ca063c690, pool=0x7f0ca063c690) at src/include/daos_srv/pool.h:296 #6 pc=rpc@entry=0x7f0ca0998d30, p_rpt=p_rpt@entry=0x7f0ca83a77b0) at src/rebuild/srv.c:2101 #7 rebuild_tgt_scan_handler (rpc=0x7f0ca0998d30) at src/rebuild/scan.c:954 #8 crt_handle_rpc (arg=0x7f0ca0998d30) at src/cart/crt_rpc.c:1654 #9 ABTD_ythread_func_wrapper (p_arg=0x7f0ca83a78a0) at arch/abtd_ythread.c:21 #10 make_fcontext () from /usr/lib64/libabt.so.1 #11 ?? () The ds_pool_get_version call passed a NULL map argument to pool_map_get_version. The ds_pool.sp_map field may be NULL after the pool is started but before the pool receives the initial pool map from the pool service. This patch fixes ds_pool_get_version to return 0, which is less than all valid pool map versions, when sp_map is NULL, resulting in rebuild retries like this: Rebuild [queued] (pool=3bf68c9c ver=2) tgts=2 Rebuild [started] (pool 3bf68c9c ver=2) Rebuild [failed] (pool 3bf68c9c ver=2 status=DER_BUSY(-1012): 'Device or resource busy') Rebuild [queued] (pool=3bf68c9c ver=2) tgts=2 Rebuild [started] (pool 3bf68c9c ver=2) Rebuild [failed] (pool 3bf68c9c ver=2 status=DER_BUSY(-1012): 'Device or resource busy') Rebuild [queued] (pool=3bf68c9c ver=2) tgts=2 Rebuild [started] (pool 3bf68c9c ver=2) Rebuild [scanning] (pool 3bf68c9c ver=2, toberb_obj=0, rb_obj=0, [...] Rebuild [scanning] (pool 3bf68c9c ver=2, toberb_obj=0, rb_obj=0, [...] Rebuild [completed] (pool 3bf68c9c ver=2, toberb_obj=0, rb_obj=0,[...] Target[2] (rank 2 idx 0 status 16 ver 1) is excluded. Also, this patch removes some rebuild code that handles NULL ds_pool.sp_group fields. Those can not happen as we always initialize sp_group (as well as sp_iv_ns) before putting a ds_pool object into the LRU. Signed-off-by: Li Wei <[email protected]>
18 tasks
mlawsonca
pushed a commit
that referenced
this pull request
May 16, 2023
Signed-off-by: Łukasz Sitkiewicz <[email protected]>
18 tasks
osalyk
pushed a commit
to osalyk/daos
that referenced
this pull request
Oct 18, 2023
…caffolding dumps: dtx_update_abort scaffolding
18 tasks
18 tasks
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Signed-off-by: Minh Diep [email protected]