revconn: don't touch thedb when a worker exits with the db - #6211
markhannum wants to merge 1 commit into
Conversation
roborivers
left a comment
roborivers
left a comment
There was a problem hiding this comment.
Cbuild submission: Success ✓.
Regression testing: Success ✓.
The first 10 failing tests are:
logfill [db unavailable at finish] **quarantined**
tsa
consumer_non_atomic_default_consumer_generated **quarantined**
bind_query_plan
sc_downgrade [timeout] **quarantined**
roborivers
left a comment
roborivers
left a comment
There was a problem hiding this comment.
Cbuild submission: Error ⚠.
Regression testing: Success ✓.
The first 10 failing tests are:
ssl_san
consumer_non_atomic_default_consumer_generated **quarantined**
ssl_set_cmd
ssl_prefer
ssl_dbname
sc_downgrade [timeout] **quarantined**
sc_timepart [timeout] **quarantined**
reco-ddlk-sql [timeout] **quarantined**
tsa [timeout]
roborivers
left a comment
roborivers
left a comment
There was a problem hiding this comment.
Cbuild submission: Error ⚠.
Regression testing: Success ✓.
The first 10 failing tests are:
consumer_non_atomic_default_consumer_generated **quarantined**
sc_downgrade [timeout] **quarantined**
reco-ddlk-sql [timeout] **quarantined**
Workers aren't thrman-registered or joined, so exit frees thedb while one can still be blocked in send_reversesql_request. Signed-off-by: Mark Hannum <mhannum@bloomberg.net>
29c26d1 to
cc10d3c
Compare
roborivers
left a comment
roborivers
left a comment
There was a problem hiding this comment.
Cbuild submission: Success ✓.
Regression testing: Success ✓.
The first 10 failing tests are:
reco-ddlk-sql **quarantined**
incoherent_startup
consumer_non_atomic_default_consumer_generated **quarantined**
sc_downgrade [timeout] **quarantined**
sql_logfill_autodisable [timeout]
|
would it be better to register the worker threads instead so the clean exit waits for them? |
roborivers
left a comment
roborivers
left a comment
There was a problem hiding this comment.
Cbuild submission: Success ✓.
Regression testing: Success ✓.
The first 10 failing tests are:
consumer_non_atomic_default_consumer_generated **quarantined**
sc_downgrade [timeout] **quarantined**
truncatesc_offline_generated [timeout] **quarantined**
reco-ddlk-sql [timeout] **quarantined**
3 participants
Revconn workers aren't thrman-registered or joined, so clean exit frees thedb while one can still be blocked in send_reversesql_request; its DONE event then derefs NULL (backend_thread_event(dbenv=0x0), glue.c:4437). Skip the event when the db is exiting.