Huge Manifest size
- Dominant language
- C++
- Stars
- 39.4k
- Forks
- 8.2k
- PR merge metrics
- No merged PRs in 30d
Description
Hi,
We are using leveldb version 1.20. Ours is a java application which will do put and delete batch operations towards leveldb via jni. in one of our servers we have seen that manifest file size has been increased to 25GB. Can we know reason for this increase. Also if we execute repair db will our application be crashed due to large Manifest size. Do our application store that manifest file in memory and process the rdb operation? Or do we have any other impacts like extra RAM consumption when we execute rdb?
Please find below the settings of our leveldb and repair db:
LevelDB:
writebuffersize=64
cachesize=3072
compressiontype=SNAPPY
totaldbinstance=1
#Bloom filter policy bit count for db
filterpolicy.bitcount=100
RepairDB:
writebuffersize=4
compressiontype=SNAPPY
cachesize=4
#Bloom filter policy bit count for repair db
bitcount=10
Regards,
Krishna
Contributor guide
Assessment
This issue has not been assessed yet.