insert query update delete getmore command flushes mapped vsize res faults locked db idx miss % qr|qw ar|aw netIn netOut conn set repl time
4 8 4 *0 12 22|0 0 44.7g 92.1g 28.4g 0 mailer:0.3% 0 0|0 0|0 42k 193k 237 rep1 PRI 09:17:46
6 7 4 *0 16 30|0 0 44.7g 92.1g 28.4g 2 mailer:0.4% 0 0|0 0|0 87k 283k 238 rep1 PRI 09:17:47
5 12 4 *0 14 23|0 0 44.7g 92.1g 28.4g 2 mailer:0.4% 0 0|0 0|0 68k 313k 238 rep1 PRI 09:17:48
3 9 2 *0 6 13|0 0 44.7g 92.1g 28.4g 1 mailer:0.3% 0 0|0 0|0 37k 208k 238 rep1 PRI 09:17:49
7 11 4 *0 18 29|0 0 44.7g 92.1g 28.4g 3 mailer:0.5% 0 0|0 0|0 100k 370k 238 rep1 PRI 09:17:50
6 7 4 *0 16 28|0 0 44.7g 92.1g 28.4g 2 mailer:0.3% 0 0|0 0|0 82k 334k 238 rep1 PRI 09:17:51
5 7 2 *0 10 17|0 0 44.7g 92.1g 28.4g 0 mailer:0.2% 0 0|0 0|0 68k 240k 238 rep1 PRI 09:17:52
2 7 4 *0 8 16|0 0 44.7g 92.1g 28.4g 1 mailer:0.3% 0 0|0 0|0 23k 150k 238 rep1 PRI 09:17:53
3 7 4 *0 10 17|0 0 44.7g 92.1g 28.4g 1 mailer:0.2% 0 0|0 0|0 43k 189k 238 rep1 PRI 09:17:54
2 8 4 *0 8 16|0 0 44.7g 92.1g 28.4g 0 mailer:0.3% 0 0|0 0|0 29k 199k 238 rep1 PRI 09:17:55
rep1:PRIMARY> db.serverStatus().extra_info
{
"note" : "fields vary by platform",
"heap_usage_bytes" : 569423584,
"page_faults" : 237554
}
free -m
total used free shared buffers cached
Mem: 32075 31874 201 0 327 29480
-/+ buffers/cache: 2066 30009
Swap: 31999 0 31999
Server have 32 gigabytes of memory and data it's only around 15 gigabytes and there is no other application run on this server except mongodb,
Might somebody can help me to solved this issue,
I'm having trouble with page fault in mongodb, this server is new and there is little transaction on database but i always have page fault
It is normal to have some page faults when mongod needs to read data that isn't in physical memory yet: http://docs.mongodb.org/
Your mongostat output has very low page fault activity and the other metrics don't suggest any performance issues. The accumulated page faults are hard to put in context without knowing how long your mongod has been running for .. these could be a very low average over time.
Unless you are noticing any specific performance issues it doesn't seem like there is anything to be concerned about.
Currently mongod service is up for 1068038, here is the detail,
"host" : "tkp85",
"version" : "2.6.5",
"process" : "mongod",
"pid" : NumberLong(22843),
"uptime" : 1068038,
"uptimeMillis" : NumberLong(1068038048),
"uptimeEstimate" : 1061445,
"localTime" : ISODate("2014-12-17T01:42:45.
I know mongodb will experience page fault when have exhausted memory, i'm just curious because we just start this machine about 2 weeks.
> I know mongodb will experience page fault when have exhausted memory, i'm
> just curious because we just start this machine about 2 weeks.
> just curious because we just start this machine about 2 weeks.
As per the docs page I mentioned earlier, the page fault counter in MongoDB is part of normal operation and does not only increase when physical memory is exhausted:
Page faults can occur as MongoDB reads from or writes data to parts of its data files that are not currently located in physical memory. In contrast, operating system page faults happen when physical memory is exhausted and pages of physical memory are swapped to disk.
Unless you see a high (and sustained) number of page faults per second, this is generally not an indicator of a performance problem. For example, if you restart mongod you will see more frequent page faults as newly accessed data is loaded into memory.
Typically page faults (and most monitoring metrics) don't tell a full story in isolation. If you have some monitoring in place you will normally see some correlation with other activity such as op counters (inserts, update, queries, deletes, ..), I/O stats, queues, etc.
If your mongostat output was a representative sample, it doesn't look like you have any performance concern at the moment.
Yeah i think this is not a problem because after I check all goes well in the system.
Thank You for your help Stephen
댓글 없음:
댓글 쓰기