I'm seeing above average failure rates for 1079 and 1080
1079 looks like the archive create is requesting metrics that don't necessarily
exist and then the replay is not checking the archive correctly ... using an
instance identifier that is not defined in the archive should not happen, I
suspect.
*** 1079.out 2015-06-07 16:28:35.928574860 +1000
--- 1079.out.bad 2015-06-09 16:29:08.255585737 +1000
***************
*** 1505,1508 ****
--- 1505,1509 ----
Recording an archive using atop
Replaying an archive using atop
+ pcp-atop: lvm instances: Instance domain identifier not defined in the PCP
archive log
done
And I'm seeing lots of variants of valgrind failures for 1080. This is
representative
Invalid read of size 1
at 0x4C2DD64: __strcmp_sse42 (in
/usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so)
by 0x40ACAA: deviatsyst (deviate.c:883)
by 0x402F62: engine (atop.c:643)
by 0x402F62: main (atop.c:455)
Address 0x82c4320 is 0 bytes after a block of size 80 alloc'd
at 0x4C2B974: calloc (in /usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so)
by 0x40CEA6: sstat_alloc (photosyst.c:49)
by 0x402C7E: engine (atop.c:505)
by 0x402C7E: main (atop.c:455)
{
<insert_a_suppression_name_here>
Memcheck:Addr1
fun:__strcmp_sse42
fun:deviatsyst
fun:engine
fun:main
}
|