db_crashtest.py use cheaper settings

[ajkr]

Despite attempts to optimize db_stress setup phase (i.e.,
pre-OperateDb()) latency in #9470 and #9475, it still always took tens
of seconds. Since we still aren't able to setup a 100M key db_stress
quickly, we should reduce the number of keys. This PR reduces it 4x
while increasing value_size_mult 4x (from its default value of 8) so
that memtables and SST files fill at a similar rate compared to before this PR.

Also disabled bzip2 compression since we'll probably never use it and
I noticed many CI runs spending majority of CPU on bzip2 decompression.

[facebook-github-bot]

@ajkr has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator.

[siying]

I thought of removing bzip and thought that keeping it might be helpful to simulate the case where there is a large gap between compression speed among files. But I guess you are right that improving test speed can help us increase coverage and would be a far better trade-off.

[ajkr]

Thanks for the review! Yeah, after a db_stress run using bzip2 files, I was seeing several subsequent db_stress runs get killed before finishing setup phase because Get()/MultiGet() were taking too long to decompress. I think it wasn't an infinite cycle though - eventually there'd be a db_stress with options causing compaction using a different compression algorithm during the setup phase.