Re: Postgres Replaying WAL slowly

Here's what 'perf top' said on streaming replica :

Samples: 26K of event 'cpu-clock', Event count (approx.): 19781
95.97% postgres [.] 0x00000000002210f3
0.41% perf [.] 0x000000000005f225
0.39% libc-2.12.so [.] __strstr_sse2
0.22% libc-2.12.so [.] memchr
0.22% [kernel] [k] kallsyms_expand_symbol
0.18% perf [.] symbols__insert
0.18% [kernel] [k] format_decode
0.15% libc-2.12.so [.] __GI___strcmp_ssse3
0.13% [kernel] [k] string
0.12% [kernel] [k] number
0.12% [kernel] [k] vsnprintf
0.12% libc-2.12.so [.] _IO_vfscanf
0.11% perf [.] dso__find_symbol
0.11% [kernel] [k] _spin_unlock_irqrestore
0.10% perf [.] hex2u64
0.10% postgres [.]
hash_search_with_hash_value
0.09% perf [.] rb_next
0.08% libc-2.12.so [.] memcpy
0.07% libc-2.12.so [.] __strchr_sse2
0.07% [kernel] [k] clear_page
0.06% [kernel] [k] strnlen
0.05% perf [.] perf_evsel__parse_sample
0.05% perf [.] rb_insert_color
0.05% [kernel] [k] pointer

On Mon, Jun 30, 2014 at 2:05 PM, Heikki Linnakangas <hlinnakangas(at)vmware(dot)com
> wrote:

> On 06/29/2014 03:43 PM, Soni M wrote:
>
>> top and sar says 100% cpu usage of one core, no sign of I/O wait.
>>
>
> Hmm, I wonder what it's doing then... If you have "perf" installed on the
> system, you can do "perf top" to get a quick overlook of where the CPU time
> is spent.
>
> - Heikki
>
>

--
Regards,

Soni Maula Harriz