Still doesn't look like the content you would like to see:
r...@leopard-6:~# mdb -p $(pgrep idmapd)
Loading modules: [ ld.so.1 libumem.so.1 libavl.so.1 libuutil.so.1
libnvpair.so.1 ]
>
> ::findleaks
BYTES LEAKED VMEM_SEG CALLER
4096 2 fe5a7000 MMAP
16384 1 feb2b000 MMAP
------------------------------------------------------------------------
Total 2 oversized leaks, 20480 bytes
CACHE LEAKED BUFFER CALLER
08199010 38536 08248000 ?
08199590 55153 081db540 ?
0819a010 70049 081af008 ?
0819a590 49489 081eb4a0 ?
0819b010 42626 081cc3a0 ?
0819f590 33258 081e96c0 ?
081a0010 21859 081b3328 ?
081a1010 16629 081cf0e8 ?
081aa590 16629 08200000 ?
------------------------------------------------------------------------
Total 344228 buffers, 85752344 bytes
>
>
-----Original Message-----
From: Nicolas Williams [mailto:[email protected]]
Sent: Thursday, March 05, 2009 4:46 PM
To: John Rogers
Cc: John Keiffer; '[email protected]'; '[email protected]'
Subject: Re: [cifs-discuss] idmapd leaking?
On Thu, Mar 05, 2009 at 04:40:55PM -0800, John Rogers wrote:
> Here is some kind of result:
> r...@leopard-6:/tmp# pldd /tmp/idmapd-core.27374 |grep mem
> /lib/libumem.so.1
> r...@leopard-6:/tmp# echo ::findleaks|mdb /tmp/idmapd-core.27374
> BYTES LEAKED VMEM_SEG CALLER
> 16384 1 feb2b000 MMAP
> ------------------------------------------------------------------------
> Total 1 oversized leak, 16384 bytes
>
> CACHE LEAKED BUFFER CALLER
> 08199590 3341 081d6010 ?
Sigh. I guess the default core content was not sufficient.
You can also attach to the running idmapd:
# mdb -p $(pgrep idmapd)
> ::findleaks
...
Sorry, I thought gcore included the necessary core content.
_______________________________________________
cifs-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/cifs-discuss
