git-svn-id: file:///svn/toku/tokudb@50453 c7de825b-a66e-492c-adef-691d508d4ae1

refs #5710 not the right solution. needs to ignore secondary indexes and therefore get pushed to the handlerton


git-svn-id: file:///svn/toku/tokudb@50450 c7de825b-a66e-492c-adef-691d508d4ae1

git-svn-id: file:///svn/toku/tokudb@50428 c7de825b-a66e-492c-adef-691d508d4ae1

closes #5713 make mutex initializers macros (as posix requires) and clean up the if statement garbage


git-svn-id: file:///svn/toku/tokudb@50424 c7de825b-a66e-492c-adef-691d508d4ae1

git-svn-id: file:///svn/toku/tokudb@50422 c7de825b-a66e-492c-adef-691d508d4ae1

refs #5710 add FT_MSG_KEYVAL_BYTES_IN to measure the amount of key/value userdata ingested at the root. this value does not count message/xid overhead so it is a good denominator when calculating write amplification


git-svn-id: file:///svn/toku/tokudb@50404 c7de825b-a66e-492c-adef-691d508d4ae1

refs #5710 start new instrumentation work by cleaning out some commented out instrumentation in ft-ops


git-svn-id: file:///svn/toku/tokudb@50403 c7de825b-a66e-492c-adef-691d508d4ae1

refs #5654, grab parent's txn lock when moving rollback entries from child to parent when the child commits. This prevents a race condition with a hot indexer simultaneously adding something to the parent's entry

git-svn-id: file:///svn/toku/tokudb@50375 c7de825b-a66e-492c-adef-691d508d4ae1

git-svn-id: file:///svn/toku/tokudb@50367 c7de825b-a66e-492c-adef-691d508d4ae1

git-svn-id: file:///svn/toku/tokudb@50341 c7de825b-a66e-492c-adef-691d508d4ae1

git-svn-id: file:///svn/toku/tokudb@50336 c7de825b-a66e-492c-adef-691d508d4ae1

git-svn-id: file:///svn/toku/tokudb@50330 c7de825b-a66e-492c-adef-691d508d4ae1

git-svn-id: file:///svn/toku/tokudb@50329 c7de825b-a66e-492c-adef-691d508d4ae1

refs #5629 get rid of some unneeded byte flipping. stick to little endian witha little endian comparison function.


git-svn-id: file:///svn/toku/tokudb@50327 c7de825b-a66e-492c-adef-691d508d4ae1

git-svn-id: file:///svn/toku/tokudb@50326 c7de825b-a66e-492c-adef-691d508d4ae1

git-svn-id: file:///svn/toku/tokudb@50325 c7de825b-a66e-492c-adef-691d508d4ae1

closes #5669, increase the cachetable limit, it needs to be at least twice the size of the single pair that we have, otherwise, the evictor may sneak in when we are calling change_pair_attr, and catch the cachetable to have a size of 16, because we have added the new attr but not removed the old one

git-svn-id: file:///svn/toku/tokudb@50322 c7de825b-a66e-492c-adef-691d508d4ae1

git-svn-id: file:///svn/toku/tokudb@50321 c7de825b-a66e-492c-adef-691d508d4ae1

git-svn-id: file:///svn/toku/tokudb@50320 c7de825b-a66e-492c-adef-691d508d4ae1

git-svn-id: file:///svn/toku/tokudb@50319 c7de825b-a66e-492c-adef-691d508d4ae1

git-svn-id: file:///svn/toku/tokudb@50318 c7de825b-a66e-492c-adef-691d508d4ae1

git-svn-id: file:///svn/toku/tokudb@50317 c7de825b-a66e-492c-adef-691d508d4ae1

git-svn-id: file:///svn/toku/tokudb@50316 c7de825b-a66e-492c-adef-691d508d4ae1

git-svn-id: file:///svn/toku/tokudb@50314 c7de825b-a66e-492c-adef-691d508d4ae1

git-svn-id: file:///svn/toku/tokudb@50312 c7de825b-a66e-492c-adef-691d508d4ae1

git-svn-id: file:///svn/toku/tokudb@50311 c7de825b-a66e-492c-adef-691d508d4ae1

git-svn-id: file:///svn/toku/tokudb@50310 c7de825b-a66e-492c-adef-691d508d4ae1

git-svn-id: file:///svn/toku/tokudb@50309 c7de825b-a66e-492c-adef-691d508d4ae1

git-svn-id: file:///svn/toku/tokudb@50308 c7de825b-a66e-492c-adef-691d508d4ae1

refs #5702 remove the old perf_insert_multiple, add the new perf_iibench. it is currently very simple. the first DB is a unique index with an autoincrement key, the rest of the DBs are non-unique and have random keys.


git-svn-id: file:///svn/toku/tokudb@50304 c7de825b-a66e-492c-adef-691d508d4ae1

fixes #5701 The assert half correct, half over-active. When the manager's mutex is held it is okay to transition an existing locktree's reference count from 0 to 1. If the lock isn't held, it's a race, since one thread could be destroying the zero-ref count locktree when you're incrementing the count. The failure occurred with the manager's mutx held (so the assert was overactive) and it's a layering-pain to assert that the manager's mutex is held otherwise, so just get rid of it.


git-svn-id: file:///svn/toku/tokudb@50297 c7de825b-a66e-492c-adef-691d508d4ae1

git-svn-id: file:///svn/toku/tokudb@50277 c7de825b-a66e-492c-adef-691d508d4ae1

git-svn-id: file:///svn/toku/tokudb@50207 c7de825b-a66e-492c-adef-691d508d4ae1

git-svn-id: file:///svn/toku/tokudb@50177 c7de825b-a66e-492c-adef-691d508d4ae1

git-svn-id: file:///svn/toku/tokudb@50176 c7de825b-a66e-492c-adef-691d508d4ae1

git-svn-id: file:///svn/toku/tokudb@50174 c7de825b-a66e-492c-adef-691d508d4ae1

git-svn-id: file:///svn/toku/tokudb@50172 c7de825b-a66e-492c-adef-691d508d4ae1

(an hour here)


git-svn-id: file:///svn/toku/tokudb@50170 c7de825b-a66e-492c-adef-691d508d4ae1

refs #5671 closes #5662 pass long join timeout for stress tests, fix num_ptquery/num_update reporting


git-svn-id: file:///svn/toku/tokudb@50169 c7de825b-a66e-492c-adef-691d508d4ae1

git-svn-id: file:///svn/toku/tokudb@50168 c7de825b-a66e-492c-adef-691d508d4ae1