[performancecopilot/pcp] reported pmDestroyContext: pthread_mutex_destroy(c_lock) failed: Device or resource busy, failed second try as well (#118)

["Frank Ch. Eigler" <notifications@xxxxxxxxxx>]

Filing here with incomplete information, so issue is not forgotten:

hc000 on irc reported an instance where, using pcp 3.11.5, possibly two copies of vector talking to the same pmwebd & same remote pmcd, occasionally suffering 1.8s latencies for queries.

pmwebd logs indicate this, running with pmwebd -Dpdu, and maybe also pmcd -t1 to induce timeout errors for those 1.8s requests.

pmwebd(14086): context web103761=pm2) expired. pmDestroyContext: 
pthread_mutex_destroy(c_lock) failed: Device or resource busy, failed second try as well

Given that this part of pmwebd is strictly single-threaded, this seems to implicate libpcp remote/pdu error handling, and/or the sharing of sockets within libpcp toward the same remote pmcd. It may be worth finally dropping the latter functionality, as a prophylactic.

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub, or mute the thread.