[mongodb-user] Re: Cache activity vs Page faults on Wired tiger
Re: Cache activity vs Page faults on Wired tiger
Kevin Adistambha <
Tue, 10 May 2016 21:48:16 -0700 (PDT)
I am seeing quite a bit of “Read into” activity on the cache but my page
faults are 0. Shouldn’t the “Read into” activity of the cache be captured
as page faults?
Where are you seeing these statistics? Did you gather them from
is MMAPv1 <
statistics that tracks how often MongoDB must access the disk when the
associated data was not loaded into memory yet. Historically, this metric
was used to check whether there is physical memory pressure in the system.
This statistics is less relevant for the WiredTiger storage engine, since
WiredTiger tracks its statistics differently.
In WiredTiger, there is no single “page fault”-like metric that represents
memory usage. The relevant statistics in WiredTiger are the wiredTiger.cache
document from the output of db.serverStatus(). In particular, wiredTiger.cache.unmodified
shows the number of pages evicted from the cache. If it corresponds to an
increase in wiredTiger.cache.pages read into cache
it may indicate that pages are evicted from the cache to be replaced by
another page read from disk. Possible reason include large collection scans.
Please see the wiredTiger.cache page
for more information regarding this topic.
You received this message because you are subscribed to the Google Groups "mongodb-user"
For other MongoDB technical support options, see:
You received this message because you are subscribed to the Google Groups "mongodb-user" group.
To unsubscribe from this group and stop receiving emails from it, send an email to mongodb-user+unsubscribe@xxxxxxxxxxxxxxxx.
To post to this group, send email to mongodb-user@xxxxxxxxxxxxxxxx.
Visit this group at
To view this discussion on the web visit
For more options, visit
for more updates.