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.
> 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.
It sounds like a job for dtrace (and no, I can't tell you how to use that :-) )
> On 06/ 7/12 02:05 PM, adirtymindisajoyforever wrote:
>> 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.
> It sounds like a job for dtrace (and no, I can't tell you how to use > that :-) )
Try plockstat(1M) rather than dtrace directly.
Not sure how much the output will help you for a java process though.
-- Andrew Gabriel
[email address is not usable -- followup in the newsgroup]
> 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.
