On Mon, Jun 22, 2009 at 10:09 PM, Jason King<[email protected]> wrote:
> On Mon, Jun 22, 2009 at 9:40 PM, Charles Hedrick<[email protected]> wrote:
>> Thanks for the suggestions. I'm using u14. (Actually I think I'm using 1.7 
>> at this moment, but normally 1.6.0u14.) We're moving to u14 in production, 
>> in hopes of beginning to test G1. (We're going to be *very* conservative 
>> about that though.) On our production systems we've had a bit too many long 
>> GC pauses. I'm hoping G1 will help us with that. Of course it may be that 
>> just going from 1.5 to 1.6 will be enough. (We're just in the process of 
>> that now.)
>>
>> It doesn't look like a lock issue:
>>
>> plockstat: pid 2952 has exited
>>
>> Mutex block
>>
>> Count     nsec Lock                         Caller
>> -------------------------------------------------------------------------------
>>   10   209803 libc.so.1`libc_malloc_lock   
>> libjava.so`JNU_GetStringPlatformChars+0x4d0
>>   20    72016 libc.so.1`libc_malloc_lock   
>> libjava.so`JNU_ReleaseStringPlatformChars+0x1e
>>   50    27446 libc.so.1`libc_malloc_lock   libjvm.so`__1cCosEfree6Fpv_v_+0x1e
>>   24    27960 libc.so.1`libc_malloc_lock   libzip.so`zcfree+0x1e
>>   12    25816 libc.so.1`libc_malloc_lock   
>> libjvm.so`__1cCosGmalloc6FI_pv_+0x2e
>>    5    26522 libc.so.1`libc_malloc_lock   libverify.so`verify_method+0x99a
>>    2    55999 libc.so.1`libc_malloc_lock   
>> libzip.so`Java_java_util_zip_Inflater_inflateBytes+0xb4
>>    6    17213 libc.so.1`libc_malloc_lock   libzip.so`freeZip+0x93
>>    5    18190 libc.so.1`libc_malloc_lock   libzip.so`freeMetaNames+0x3b
>>    1    85590 libc.so.1`libc_malloc_lock   
>> libverify.so`VerifyClassForMajorVersion+0x46c
>>    1    50780 libc.so.1`libc_malloc_lock   
>> libzip.so`Java_java_util_zip_Deflater_deflateBytes+0x1b7
>>    2    24013 libc.so.1`libc_malloc_lock   
>> libverify.so`signature_to_fieldtype+0x93b
>>    2    20947 libc.so.1`libc_malloc_lock   
>> libverify.so`VerifyClassForMajorVersion+0x306
>>    2    20475 libc.so.1`libc_malloc_lock   
>> libzip.so`Java_java_util_zip_Deflater_deflateBytes+0x101
>>    1    39596 libc.so.1`libc_malloc_lock   libverify.so`CCinit+0x22
>>    1    39205 libc.so.1`libc_malloc_lock   libverify.so`class_to_ID+0x74d
>>    1    37187 libc.so.1`libc_malloc_lock   
>> libverify.so`class_name_to_ID+0x323
>>    1    37001 libc.so.1`libc_malloc_lock   libc.so.1`calloc+0x4c
>>    4     8815 libc.so.1`libc_malloc_lock   libzip.so`ZIP_FreeEntry+0x48
>>    1    32371 libc.so.1`libc_malloc_lock   libzip.so`ZIP_FreeEntry+0x7c
>>    4     4338 libc.so.1`libc_malloc_lock   
>> libzip.so`Java_java_util_zip_Inflater_end+0x2e
>>    1    17218 libc.so.1`libc_malloc_lock   libCrun.so.1`__1c2n6FI_pv_+0x38
>>    1    10798 libc.so.1`libc_malloc_lock   
>> libverify.so`VerifyClassForMajorVersion+0x54d
>>    1    10341 libc.so.1`libc_malloc_lock   libjava.so`readBytes+0x17f
>>    1    10170 libc.so.1`libc_malloc_lock   libzip.so`freeCEN+0x21
>>    1     9553 libc.so.1`libc_malloc_lock   libzip.so`newEntry+0x1f
>>    1     9060 libc.so.1`libc_malloc_lock   libzip.so`freeMetaNames+0x5b
>>    1     8116 libc.so.1`libc_malloc_lock   
>> libverify.so`cp_index_to_class_fullinfo+0x6c2
>>    1     3288 libc.so.1`libc_malloc_lock   
>> libverify.so`VerifyClassForMajorVersion+0xb48
>
> How long did it run, and how many threads?  The original post suggests
> a 5 second runtime.   Unless I'm forgetting the meaning of the output
> values from plockstat, that's almost a whole second (900msec) of
> blocking on the libc malloc lock (all by various callers, but the same
> lock).  If the code is only running for 5 seconds, that seems
> significant (though depends on how many threads).
>
> try doing "LD_PRELOAD=libmtmalloc.so java ....." and
> "LD_PRELOAD=libumem.so java ...." to run your commands and see how
> each fares (each has different pathological cases so if you can,
> doesn't hurt to see if one does better than the other).
>

I just reread the original post (wasn't visible with all the other
threads) and just noticed I confused units on the original times, my
apologies :)

Still might be worth trying it just to see.
_______________________________________________
opensolaris-discuss mailing list
[email protected]

Reply via email to