librbd/cache/pwl/ssd/WriteLog: decrement m_bytes_allocated when retiring #41068
Conversation
|
cc @MahatiC |
|
cc @CongMinYin |
| @@ -708,6 +708,8 @@ bool WriteLog<I>::retire_entries(const unsigned long int frees_per_tx) { | |||
| m_first_valid_entry = first_valid_entry; | |||
| ceph_assert(m_first_valid_entry % MIN_WRITE_ALLOC_SSD_SIZE == 0); | |||
| this->m_free_log_entries += retiring_entries.size(); | |||
| ceph_assert(this->m_bytes_allocated >= allocated_bytes); | |||
| this->m_bytes_allocated -= allocated_bytes; | |||
There was a problem hiding this comment.
This logic does exist in RWL, but it is omitted here.
There was a problem hiding this comment.
I'm not sure I understand the comment. Do you mean that it is omitted intentionally?
There was a problem hiding this comment.
No, it's not omitted intentionally. This PR looks correct. I actually noticed this issue and fixed it locally along with a few more fixes and haven't upstreamed them yet. And I am currently running some tests as well. I will review this change once again by tomorrow. Thank you for this PR!
There was a problem hiding this comment.
FWIW I have another PR for the ssd mode almost ready -- fixing power cycle issues. Should be able to submit by tomorrow. I'll CC you, thanks for taking a look!
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Currently if ssd cache is filled to capacity, all future I/O hangs indefinitely because even though the cache eventually becomes clean and retires enough entries to get back under RETIRE_HIGH_WATER, this isn't communicated to AbstractWriteLog::check_allocation(). Fixes: https://tracker.ceph.com/issues/50560 Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Follow rwl mode and use AbstractWriteLog::m_bytes_allocated_cap instead of m_log_pool_ring_buffer_size specific to ssd. This fixes "bytes available" calculation in STATS output. Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
…evel Bump "Waiting for allocation" to 5. "Retiring" is at 20 for rwl and 1 for ssd. Bump the latter to 20 as well. Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
While at it, reduce the number of calls to operator<< and drop the trailing comma. Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
idryomov
force-pushed
the
wip-rbd-pwl-ssd-capacity
branch
from
b5b8150
to
2a974fd
Compare
|
Added a fix for an existing free()/delete mismatch. |
|
Thanks! Looks good to me. |
Currently if ssd cache is filled to capacity, all future I/O hangs
indefinitely because even though the cache eventually becomes clean
and retires enough entries to get back under RETIRE_HIGH_WATER, this
isn't communicated to AbstractWriteLog::check_allocation().
Fixes: https://tracker.ceph.com/issues/50560
Signed-off-by: Ilya Dryomov idryomov@gmail.com