watchman frequently times out

NEW
Unassigned

Status

Developer Services
General
10 days ago
10 days ago

People

(Reporter: sfink, Unassigned)

Tracking

Details

(Reporter)

Description

10 days ago
watchman has been problematic for me for quite a while. Example: I just did

$ hg amend
warning: Watchman unavailable: timed out waiting for response, while executing ('query', '/home/sfink/src/mozilla', {'fields': ['mode', 'mtime', 'size', 'exists', 'name'], 'since': 'c:1508545429:8873:2:7407942', 'expression': ['not', ['anyof', ['dirname', '.hg'], ['name', '.hg', 'wholename']]], 'empty_on_fresh_instance': False, 'sync_timeout': 2000})

I think this watchman log message probably came from the same command:

2017-11-14T11:28:55,771: [client=0x9f4cc0:stm=0x9ee800:pid=10724] PERF: {"ru_nvcsw": 63, "ru_oublock": 0, "ru_inblock": 400, "ru_msgsnd": 0, "ru_nswap": 0, "ru_majflt": 50, "ru_nivcsw": 0, "ru_minflt": 0, "ru_idrss": 0, "ru_msgrcv": 0, "version": "4.9.0", "start_time": 1510687732.9232359, "meta": {"args": ["query", "/home/sfink/src/mozilla", {"sync_timeout": 2000, "fields": ["mode", "mtime", "size", "exists", "name"], "since": "c:1508545429:8873:2:7407942", "expression": ["not", ["anyof", ["dirname", ".hg"], ["name", ".hg", "wholename"]]], "empty_on_fresh_instance": false}], "client": {"pid": 10724}, "root": {"watcher": "inotify", "ticks": 7480897, "number": 2, "path": "/home/sfink/src/mozilla", "case_sensitive": true, "recrawl_count": 0}}, "description": "dispatch_command:query", "elapsed_time": 2.848096, "user_time": 0.001384, "ru_maxrss": 0, "system_time": 0.0068580000000000004, "ru_nsignals": 0, "pid": 8873, "ru_ixrss": 0}

I get mysterious pauses quite often, which I suspect may be related to watchman, and a few timeouts every day.
You need to log in before you can comment on or make changes to this bug.