mon/PGMonitor: batch filter pg states; add sanity check #9394
Conversation
Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>
xiexingguo
force-pushed
the
xxg-wip-batchpgstates
branch
from
3d17c23
to
17be6f0
Compare
|
@tchaikov Some mon-related minor fixes, could you kindly help to review? Thanks:-) |
| return; | ||
| } | ||
|
|
||
| MStatfsReply *reply; |
There was a problem hiding this comment.
could move this line down to where reply is used for the first time.
xiexingguo
force-pushed
the
xxg-wip-batchpgstates
branch
from
9b2e143
to
443a9e2
Compare
| propose = true; | ||
| assert(pg_map.last_osdmap_epoch < epoch); | ||
| pending_inc.osdmap_epoch = epoch; | ||
| (void)map_pg_creates(); |
There was a problem hiding this comment.
check_osd_map() is the only caller of these functions. maybe we can change them so that they return void instead?
xiexingguo
force-pushed
the
xxg-wip-batchpgstates
branch
from
443a9e2
to
2b67303
Compare
|
@tchaikov Done and thanks. |
| @@ -1754,24 +1754,17 @@ void PGMap::generate_test_instances(list<PGMap*>& o) | |||
| } | |||
| } | |||
|
|
|||
| void PGMap::get_filtered_pg_stats(const string& state, int64_t poolid, int64_t osdid, | |||
| void PGMap::get_filtered_pg_stats(__u32 state, int64_t poolid, int64_t osdid, | |||
There was a problem hiding this comment.
__u32 is a linux specific type, maybe we can try to use uint32_t instead? i see the underlying type of pg_state_t::state is __u32. but we can be more portable by using these ANSI types.
There was a problem hiding this comment.
Done. And @tchaikov , I see you are busy with cmake, thank you for your time:-)
Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>
Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>
We'll assure pg_map.last_osdmap_epoch is smaller than osdmap's epoch at the function entry, so we are doomed to update pending_inc.osdmap_epoch, which means the "propose" variables will be always set to true. Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>
xiexingguo
force-pushed
the
xxg-wip-batchpgstates
branch
from
2b67303
to
2c8a2f6
Compare
We use "backfill_wait" everywhere except this one. Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>
| @@ -824,7 +824,7 @@ std::string pg_state_string(int state) | |||
| oss << "repair+"; | |||
| if ((state & PG_STATE_BACKFILL_WAIT) && | |||
| !(state &PG_STATE_BACKFILL)) | |||
| oss << "wait_backfill+"; | |||
| oss << "backfill_wait+"; | |||
There was a problem hiding this comment.
could update the doc/dev/osd_internals/recovery_reservation.rst accordingly.
There was a problem hiding this comment.
Done. Thank you for your reminder.
xiexingguo
force-pushed
the
xxg-wip-batchpgstates
branch
from
2c8a2f6
to
fbfd38c
Compare
|
|
||
| /** | ||
| * recalculate creating pg mappings | ||
| * | ||
| * @return true if we updated pending_inc |
There was a problem hiding this comment.
could you update the comment accordingly?
There was a problem hiding this comment.
Well, I don't understand what do you mean. It does not return anything now.
There was a problem hiding this comment.
ah, sorry, i thought you missed this.
|
lgtm |
|
@tchaikov Thanks for the merge. |
No description provided.