Details
-
Task
-
Resolution: Unresolved
-
Major
-
None
-
None
-
1
Description
It was observed in CBSE-11735 that when we issue a "shutdown" of the indexer process it takes around ~15 secs for the entire service shutdown to complete.
7110:[ns_server:debug,2022-04-04T16:41:37.386Z,babysitter_of_ns_1@cb.local:<0.18165.101>:ns_port_server:terminate:196]Shutting down port indexer
|
|
7111-[ns_server:debug,2022-04-04T16:41:58.020Z,babysitter_of_ns_1@cb.local:indexer-goport<0.18171.101>:goport:handle_eof:582]Stream 'stdout' closed
|
|
7112-[ns_server:debug,2022-04-04T16:41:58.020Z,babysitter_of_ns_1@cb.local:indexer-goport<0.18171.101>:goport:handle_eof:582]Stream 'stderr' closed
|
|
7113-[ns_server:info,2022-04-04T16:41:58.020Z,babysitter_of_ns_1@cb.local:indexer-goport<0.18171.101>:goport:handle_process_exit:563]Port exited with status 0.
|
|
7114-[ns_server:debug,2022-04-04T16:41:58.020Z,babysitter_of_ns_1@cb.local:<0.18165.101>:ns_port_server:terminate:199]indexer has exited
|
Interestingly this is not a one off occurrence and I see similar abnormally large shutdown times for indexer in the past (for the same node) and other nodes too.
rg "Shutting down port indexer" ns_server.babysitter.log -A 4
|
|
5554:[ns_server:debug,2022-03-21T21:59:48.933Z,babysitter_of_ns_1@cb.local:<0.1615.0>:ns_port_server:terminate:196]Shutting down port indexer
|
|
5555-[ns_server:debug,2022-03-21T22:00:08.311Z,babysitter_of_ns_1@cb.local:indexer-goport<0.1621.0>:goport:handle_eof:582]Stream 'stdout' closed
|
|
5556-[ns_server:debug,2022-03-21T22:00:08.311Z,babysitter_of_ns_1@cb.local:indexer-goport<0.1621.0>:goport:handle_eof:582]Stream 'stderr' closed
|
|
5557-[ns_server:info,2022-03-21T22:00:08.311Z,babysitter_of_ns_1@cb.local:indexer-goport<0.1621.0>:goport:handle_process_exit:563]Port exited with status 0.
|
|
5558-[ns_server:debug,2022-03-21T22:00:08.312Z,babysitter_of_ns_1@cb.local:<0.1615.0>:ns_port_server:terminate:199]indexer has exited
|
|
--
|
|
6701:[ns_server:debug,2022-04-03T06:02:40.160Z,babysitter_of_ns_1@cb.local:<0.4129.70>:ns_port_server:terminate:196]Shutting down port indexer
|
|
6702-[ns_server:debug,2022-04-03T06:02:54.942Z,babysitter_of_ns_1@cb.local:indexer-goport<0.4135.70>:goport:handle_eof:582]Stream 'stderr' closed
|
|
6703-[ns_server:debug,2022-04-03T06:02:54.942Z,babysitter_of_ns_1@cb.local:indexer-goport<0.4135.70>:goport:handle_eof:582]Stream 'stdout' closed
|
|
6704-[ns_server:info,2022-04-03T06:02:54.942Z,babysitter_of_ns_1@cb.local:indexer-goport<0.4135.70>:goport:handle_process_exit:563]Port exited with status 0.
|
|
6705-[ns_server:debug,2022-04-03T06:02:54.942Z,babysitter_of_ns_1@cb.local:<0.4129.70>:ns_port_server:terminate:199]indexer has exited
|
|
--
|
It would be good if we can add logging to goport code to precisely know where the issue is happening (or know that the issue isn't happening in goport and is elsewhere).
Area's to investigate the original issue in CBSE-11735:
1) I see exec package in golang mentioned that the Wait command (internally calls the wait syscall to notify when a child process spawned via exec.Start has terminated) waits till all the stdout and stderr buffers are completely drained and we have logic in goport which waits as well for eof's on stdout and stderr.
Attachments
Gerrit Reviews
For Gerrit Dashboard: MB-51813 | ||||||
---|---|---|---|---|---|---|
# | Subject | Branch | Project | Status | CR | V |
180323,4 | MB-51813: Adding log to goport | master | ns_server | Status: NEW | 0 | -1 |