Closed GoogleCodeExporter closed 8 years ago
The mapping between the MySQL status variables and the output of SHOW INNODB
STATUS
is not one-to-one. Some of the variables in one have a related value in the
other
that isn't quite the same thing. Looking at the source investigation I did for
innotop, I think pages_read has no exact equivalent in SHOW STATUS. I'll
remove this
from the list of overrides.
Original comment by baron.schwartz
on 25 Dec 2009 at 7:09
Actually I'm looking again just to be sure...
Innodb_buffer_pool_reads comes from export_vars.innodb_buffer_pool_reads, which
comes
from srv_buf_pool_reads. I don't think this is used for anything else, so
there is
nothing in SHOW INNODB STATUS to parse this from.
Innodb_pages_read comes from export_vars.innodb_pages_read, which comes from
buf_pool->n_pages_read, which is printed here:
2156 fprintf(stderr,
2157 "buf_pool size %lu\n"
2158 "database pages %lu\n"
2159 "free pages %lu\n"
2160 "modified database pages %lu\n"
2161 "n pending reads %lu\n"
2162 "n pending flush LRU %lu list %lu single page %lu\n"
2163 "pages read %lu, created %lu, written %lu\n",
2164 (ulong) size,
2165 (ulong) UT_LIST_GET_LEN(buf_pool->LRU),
2166 (ulong) UT_LIST_GET_LEN(buf_pool->free),
2167 (ulong) UT_LIST_GET_LEN(buf_pool->flush_list),
2168 (ulong) buf_pool->n_pend_reads,
2169 (ulong) buf_pool->n_flush[BUF_FLUSH_LRU],
2170 (ulong) buf_pool->n_flush[BUF_FLUSH_LIST],
2171 (ulong) buf_pool->n_flush[BUF_FLUSH_SINGLE_PAGE],
2172 (ulong) buf_pool->n_pages_read, buf_pool->n_pages_created,
2173 (ulong) buf_pool->n_pages_written);
So this is right:
'Innodb_pages_read' => 'pages_read',
I'll delete the other line. Someday I'll document the mapping between these
variables and SHOW INNODB STATUS.
Original comment by baron.schwartz
on 25 Dec 2009 at 7:21
Fortunately these overrides were processed in order so there was confusion but
no
actual error.
Original comment by baron.schwartz
on 25 Dec 2009 at 7:23
This issue was closed by revision r367.
Original comment by baron.schwartz
on 25 Dec 2009 at 7:26
Original issue reported on code.google.com by
joegrasse
on 17 Dec 2009 at 1:59