Explains how to use the guts utility to retrieve tiering
statistics.
You can run the /opt/mapr/bin/guts utility
to get granular information on ongoing offloads and recalls including:
- The number of objects that are offloaded to and recalled from the tier
- The number of reads and writes on HPE Ezmeral Data Fabric Database
- The number of reads and writes on filesystem
Syntax
/opt/mapr/bin/guts <argument>:<options>
Arguments
| Argument |
Description |
mastgateway |
Refers to operations on the MAST Gateway node. See Usage for information on the syntax for running the guts command with this argument. |
fstier |
Refers to operations on filesystem node. See Usage for information on the syntax for running the guts command with this argument. |
Options
| Option |
Description |
all |
Statistics for all operations. |
db |
Statistics for MAST Gateway operations currently running on HPE Ezmeral Data Fabric Database. |
mfsops |
Statistics for MAST Gateway operations on filesystem. |
none |
Specifies the column(s) to exclude from the output. |
tier |
Statistics for MAST Gateway operations on the storage tier. |
Usage
- MAST Gateway Node
-
/opt/mapr/bin/guts mastgateway:all
/opt/mapr/bin/guts mastgateway:db
/opt/mapr/bin/guts mastgateway:tier
/opt/mapr/bin/guts mastgateway:mfsops
/opt/mapr/bin/guts mastgateway:none
- filesystem Node
-
/opt/mapr/bin/guts fstier:all
/opt/mapr/bin/guts fstier:none
Note: These commands might show statistics for several other fields. To skip, use
none with the components whose fields you do not wish to retrieve. For
example, to retrieve statistics for only the mastgateway tier, run the following command:
/opt/mapr/bin/guts allocator:none btree:none cache:none cleaner:none client:none cpu:none db:none dbrepl:none disk:none fs:none fstier:none gateway:none io:none kv:none log:none mastgateway:tier net:none nfs:none rpc:none ssd:none streams:none time:none vcd:none
Output
mastgateway:tier
-
objP |
Number of objects (whose maximum size is 8MB for cold-tier or whose size is
computed based on the erasure coding scheme for warm-tier) that were offloaded
to the storage tier. |
objG |
Number of objects (whose size is up to 1 MB for cold-tier and whose size is
computed based on the erasure coding scheme for warm-tier) that were recalled
from the storage tier. |
objD |
Number of deletions on the tier. |
objPM |
Amount of data (in MB) offloaded per second to the storage tier. |
objGM |
Amount of data (in MB) recalled per second from the storage tier. |
mastgateway:db
-
tdbP |
Number of puts on HPE Ezmeral Data Fabric Database
tables. |
tdbG |
Number of gets on HPE Ezmeral Data Fabric Database
tables. |
tdbD |
Number of deletes on HPE Ezmeral Data Fabric Database
tables. |
mastgateway:mfsops
-
moR |
Number of read requests from client to the MAST Gateway service to
read from cache volumes. |
mp |
Number of filesystem
purge requests sent by the MAST Gateway service. |
mrw |
Number of filesystem
reads from the MAST Gateway service to perform modify/write
operation. |
moRM |
Amount of filesystem
reads (in MB) sent by the MAST Gateway service to read offloaded
data. |
mrwM |
Amount of recall writes (in MB) sent by the MAST Gateway
service. |
fstier:all
-
tp |
Number of blocks (of 64KB) purged during an offload
operation. |
trw |
Number of blocks (of 64KB) written during recall of offloaded
data. |
trr |
Number of blocks (of 64KB) read during read of offloaded
data. |
twr |
Number of blocks (of 64KB) recalled for partial overwrites. |