java, threads, truss and other nightmares
- From: adirtymindisajoyforever <getridofthespam@xxxxxxxxx>
- Date: Thu, 7 Jun 2012 06:05:57 -0700 (PDT)
I have a java application that seems to show sometimes delays
of 3 seconds for no obvious reasons.
Trussing the app shows that one of the threads is waiting for 3
seconds
for a mutex lock to be release. However, in the trace I can't find the
thread
holding/releasing the lock.
Therefor I trussed, or better tried to, also the libraries. This
results in a hanging
application, sometimes crashing.
One of the nasty things is that the call I am interested in in now
failing with ETIME:
lwp_cond_wait(0x00281848, 0x00281830, 0xB3FFFD28, 0) Err#62 ETIME
condvar type: USYNC_THREAD
mutex type: USYNC_THREAD
timeout: 0.045380900 sec
From the man page:If the time of
day becomes greater than abstime, _lwp_cond_timedwait()
returns with the error code ETIME.
The man page is for underscored versions with different parameters...
Could someone explain this behaviour and if the time delay caused by
truss is the cause of this ETIME error.
Any other suggestions for tracing/debugging thsi behaviour.
In case needed: this is a camel application, invoked by maven.
.
- Follow-Ups:
- Re: java, threads, truss and other nightmares
- From: Stefan Krueger
- Re: java, threads, truss and other nightmares
- From: YTC#1
- Re: java, threads, truss and other nightmares
- Prev by Date: Re: Wall St Journal Thurs 31may12 pg B1: amazing article on Oracle * HP battle
- Next by Date: Re: java, threads, truss and other nightmares
- Previous by thread: Wall St Journal Thurs 31may12 pg B1: amazing article on Oracle * HP battle
- Next by thread: Re: java, threads, truss and other nightmares
- Index(es):
Relevant Pages
|
