-
Notifications
You must be signed in to change notification settings - Fork 10
v3.4 Byte Ranges Read and Update
Partial write/read and append operations performance is the subject of interest also in some cases. To configure the partial/append operations it's necessary to specify the byte ranges to work with somehow. There are two ways - use random byte ranges and to specify the fixed byte ranges. The random byte ranges are specified by the count of the arbitrary ranges selected randomly by internal algorithm. There's also RFC 7233 specification describing how to specify the fixed byte ranges.
-
Effective only if load type is set to update or read.
-
Random byte range count should not be more than maximum for the particular data item size used. (For details please refer to this spec)
-
It's not allowed to specify both random and fixed byte ranges simultaneously. The fixed byte ranges configuration will be used in this case only.
- Note that combining multiple fixed ranges is also possible and is a valid case.
- The fixed ranges configuration will be used if both fixed and random ranges are configured (invalid case but should not lead to failure).
Byte Ranges Configuration | Effect |
---|---|
No byte ranges configured | Read/Overwrite the entire data item. |
A random count N | Random Byte Ranges Read or Update with new data, N ranges per request. N should be > 0. |
Fixed value: "N-" | Read/Overwrite the part of the data item with the same data starting from the position of N bytes to the end of the data item. N should be less than data item size. |
Fixed value: "-N" | Read/Overwrite last N bytes of the data item. N should be less than data item size. |
Fixed value: "N1-N2" | Read/Overwrite the part of the data tiem with the same data in the range of N1-N2 bytes. Note that according RFC 7233 the N2 is included in the range. N1 should be not more than N2. N2 may be more than data item size. |
Fixed value: "-N-" | Append N bytes to the data item using the same data source. |
Random ranges read example:
java -jar mongoose.jar
--read
--item-data-ranges-random=2
--item-input-file=items.csv
...
Random ranges update example:
java -jar mongoose.jar
--update
--item-data-ranges-random=2
--item-input-file=items2update.csv
--item-output-file=items_updated.csv
...
Partial read of the data items from 2KB (2048th byte, the 1st byte after 2048 bytes) to the end:
java -jar mongoose.jar
--read
--item-data-ranges-fixed=2KB-
--item-input-file=items.csv
...
Overwrite the data items from the second byte (including it) to the end:
java -jar mongoose.jar
--update
--item-data-ranges-fixed=1-
--item-input-file=items2overwrite_tail2KBs.csv
--item-output-file=items_with_overwritten_tails.csv
...
Read the last 1234 bytes of the data items:
java -jar mongoose.jar
--read
--item-data-ranges-fixed=-1234
--item-input-file=items.csv
...
Overwrite the last 1234 bytes of the data items:
java -jar mongoose.jar
--update
--item-data-ranges-fixed=-1234
--item-input-file=items2overwrite_tail2KBs.csv
--item-output-file=items_with_overwritten_tails.csv
...
Partially read the data items each in the range from 2KB to 5KB + 1:
java -jar mongoose.jar
--read
--item-data-ranges-fixed=2KB-5KB
--item-input-file=items.csv
...
Overwrite the data items in the range from 2KB to 5KB + 1:
java -jar mongoose.jar
--update
--item-data-ranges-fixed=2KB-5KB
--item-input-file=items2overwrite_range.csv
--item-output-file=items_overwritten_in_the_middle.csv
...
Append 16KB to the data items:
java -jar mongoose.jar
--update
--item-data-ranges-fixed=-16KB-
--item-input-file=items2append_16KB_tails.csv
--item-output-file=items_appended.csv
...
Partially read the data items using multiple fixed ranges configuration:
java -jar mongoose.jar
--read
--item-data-ranges-fixed=0-1KB,2KB-5KB,8KB-
--item-input-file=items.csv
...
- Update/Read load type should be used to use the feature: --update/--read or --load-type=update/--load-type=read
- Fixed byte ranges should be specified using the --item-data-ranges-fixed= configuration parameter.
- Overview
- Deployment
- User Guide
- Troubleshooting
- Reference