eG Help

To reduce expensive disk reads and to improve cache usage, the MongoDB cache should be populated with data that is frequently accessed. This is why, its important that data read from disk is quickly written to the cache. Any delay when reading from disk to cache can result in a high incidence of cache misses, as key data objects may not be present in the cache. Likewise, latencies when writing cache updates to the disk can result in significant data loss in the event of the database server crashes. It is therefore imperative that administrators are promptly notified if any slowness is observed in reads from disk to cache or writes from cache to disk. The Mongo IO test provides these useful alerts to administrators!

This test monitors reads from and writes to disk, and reports the time taken by the server to perform these read/write activities. In the process, the test proactively alerts administrators to potential latencies in reading from and/or writing to disk.

Measurement	Description	Measurement Unit	Interpretation
Pages_read_frm_dsk_cache	Indicates the number of pages read from disk to cache during the last measurement period.	Number
Pages_rd_frm_dsk_Ch_Time	Indicates the total time taken to read pages from disk to cache during the last measurement period.	Seconds
Disk_read_latency	Indicates the time taken to perform a single read from the disk to the cache.	Seconds/read	A low value is desired for this measure. A high value implies that it is taking a long time for a single page to be read from the disk and written to the cache. This can indicate slowness in reading from disk to cache, which in turn can result in cache misses.
Pages_wrte_frm_dsk_cache	Indicates the number of pages written from cache to disk during the last measurement period.	Number
Pages_wt_frm_dsk_ch_time	Indicates the total time taken to write pages from the cache into the disk during the last measurement period.	Seconds
Disk_write_latency	Indicates the time taken to perform a single write to the disk from the cache.	Seconds/write	A low value is desired for this measure. A high value could imply that the server is slow when writing cache updates to disk. Ideally, the data in the cache and disk should be in sync. Slowness in writing to disk can cause the data in disk to be out of sync with the data in cache. In such a situation, significant data loss may occur when disaster strikes.