-
When running the crash tests, all of them fail quickly with:
```
WARNING: prefix_size is non-zero but memtablerep != prefix_hash
Verification failed for column family 0 key 0000000000000045000000…
-
**Why:**
Improve write performance (for multiple and single threads)
Allow parallelism when working with the memtable
**What :**
Inserting data into the memtable without sorting it first will el…
-
At the moment - there are several unit tests failing in 'main'.
4 of them are mentioned here:
https://github.com/speedb-io/speedb/issues/9#issuecomment-1164276865
Since those 4 are failing in Rocks…
-
MockMemTable members are not full initialized and have a garbage value that cause an assertion on each x runs
-
improve memtable switch latency with memtables that require large upfront allocation, such as as the hash-table-based memtable
### Expected behavior
### Actual behavior
### Steps to reproduce…
-
Some of our CI targets don't have bash by default, and it'd be wasteful to install it when it's not really needed. Bash also might not be available on every user machine, and since there's no real rea…
-
For various tasks (especially for the performance tests) we sometimes need to compare builds and tell how they are ordered in the git tree. The current version number is not enough because it's only c…
-
when running the stress test with filter uri, the CreateFilterPolicy method prints to stderr when bloom_bits > 0 the normal behavior:
"note: appending --bloom-bits (%f) to --filter-uri"
this is not…
-
#29 is the issue to implement the new filter.
The new filter needs documentation - Wiki page + benchmark results
This issue is to track the documentation.
-
### reason for the change:
purges are deletions of obsolete compaction files. by doing the cleanup of these files in the HIGH priority thread pool, the flushes are hindered. we’d like to avoid dist…