41a3e6c853
add write throttling
2020-07-28 03:49:47 +00:00
0df2a8d014
better code style
2020-07-18 01:48:58 +00:00
9d0d189cc8
better serialize; TransactionOption::clear patch
2020-07-15 22:39:21 +00:00
eb44ae0e86
finish local shard estimation
2020-07-15 16:08:00 +00:00
a310faf9d1
solve some code reviews
2020-07-14 17:19:55 +00:00
d512170cd8
add clear cost estimation
2020-07-14 00:18:52 +00:00
c06d24497f
send commitCostEstimation to ratekeeper
2020-07-13 17:18:13 +00:00
6374cd96c9
add commit cost estimation
2020-07-10 23:07:29 +00:00
bb3d4b6b89
Add a bunch of TEST macros and some other little things
2020-05-04 10:11:36 -07:00
ede69c5141
Serialize throttle limits with expiration expressed as a duration to avoid clock differences.
2020-04-22 14:15:35 -07:00
d5fb4d26fe
Send tag throttle updates from ratekeeper to proxy only when they change
2020-04-21 16:33:56 -07:00
6619a1a36a
Rename transaction tag map.
2020-04-17 09:06:45 -07:00
2b66dcd24a
Some more refactoring. Reduce what is sent from RK->MP->clients
2020-04-17 08:07:01 -07:00
0fba8c47be
Checkpoint: Ratekeeper sets absolute limits for tag throttles and enforces them by distributing requests to proxies, who distribute them to clients.
...
A few refactorings.
2020-04-16 14:43:22 -07:00
29b2c2f3aa
Add hash to StringRef. Use unordered maps for storing tags. Create some helpful typedefs.
2020-04-10 12:54:59 -07:00
a1d8623e5f
Various changes to the throttling scheme, including most notably that clients now enforce the throttles with info they receive from the proxy.
2020-04-07 16:28:09 -07:00
26b7e02d4c
Some initial work to support tagging transactions and passing them around.
2020-03-20 11:23:11 -07:00
96aaa31a6c
Compiling on clang again
2019-05-13 14:15:23 -07:00
69fa3d3903
fixed compilation issues after rebase
2019-05-13 14:15:23 -07:00
0fb6a03c07
First round of review comment fixes for PR#1307
2019-03-19 11:29:19 -07:00
99d521ef4f
Monitor Ratekeeper and DataDistributor to use stateless processes
...
Since Ratekeeper and DataDistributor are no longer running with Master, they
might be running with stateful processes before a new Master becomes alive,
which is undesirable.
This PR adds a monitoring of both Ratekeeper and DataDistributor at Cluster
Controller -- if Master runs on a stateless class and RK/DD runs at a worse
class, then RK/DD will be killed. I.e., RK/DD should be running at their own
classes or on the same stateless process as Master. After restart, RK/DD should
be running at a better process class.
2019-03-14 15:00:57 -07:00
2b0139670e
Fix review comment for PR 1176
2019-03-12 12:02:30 -07:00
dc129207a9
Minor fix after rebase.
2019-03-07 13:16:20 -08:00
d52ff738c0
Fix merge conflicts during rebase.
2019-03-07 13:16:20 -08:00
e6ac3f7fe8
Minor fix on ratekeeper work registration.
2019-03-07 13:16:20 -08:00
3c86643822
Separate Ratekeeper from data distribution.
...
Add a new role for ratekeeper.
Remove StorageServerChanges from data distribution.
Ratekeeper monitors storage servers, which borrows the idea from
DataDistribution.
2019-03-07 13:16:20 -08:00
Xiaoxi Wang