536ee826da
tuned resolver balancing to keep the resolvers within 5MB per second of each other
2018-11-11 23:42:45 -08:00
0613a34845
The storage server would block the main thread when processing a single version with a large amount of data
2018-10-18 13:37:31 -07:00
636420abee
fix: if the disk queue adapter peek hangs for a while, switch to a peek from a different locality
2018-10-03 13:58:55 -07:00
28545e0f8d
multi cursors start a get more for the first 10 cursors to hide latency
2018-10-03 13:57:45 -07:00
a92fc911ac
do not spin on a failed storage server recruitment
2018-10-02 17:31:07 -07:00
200e65fe61
added a workload which tests killing an entire region, and recovering from the failure with data loss.
...
fix: we cannot pop the txs tag from remote logs until they have a full copy of the txnStateStore
fix: we have to modify all of history, we cannot stop after finding a local remote
2018-09-17 18:32:39 -07:00
6496a6d9c8
fix: start move keys will only move destination servers to become source servers if less than destination servers are healthy and the total number of sources is less than 2x the number of destinations
2018-08-31 12:43:14 -07:00
d7c01f0419
added a separate knob for tlog’s recoverMemoryLimit
2018-08-21 21:11:23 -07:00
7f7755165c
slowly send notifications to clients to clear the list of dead clients
2018-08-08 17:29:32 -07:00
be1a4d74c7
tlogs serve reads to log routers at a low priority, to prevent them from using all their resources catching up a remote dc that has been down for a long time
...
increase the amount of memory ratekeeper budgets for tlogs so that there is a gap after the spill threshold to prevent temporarily overshooting the budget
2018-08-04 10:31:30 -07:00
4fedd05506
added more yields to avoid slow tasks
2018-07-12 17:47:35 -07:00
392c73affb
fixed a few slow tasks
2018-07-12 14:06:59 -07:00
cd63c7a7cc
added a buffered cursor, which efficiently merges lots of peek cursors
2018-07-12 12:09:48 -07:00
ed7c744a8e
fix: we cannot lower max_read_transaction_life_versions too low when configured with remote logs, because it causes the log routers to fall too far behind
2018-07-09 16:55:33 -07:00
5a2cb3037b
merge 5.2 into 6.0
2018-07-08 20:14:06 -07:00
4dd18afb84
fix: we cannot make MAX_RECOVERY_VERSIONS lower than MAX_VERSIONS_IN_FLIGHT because we can mark a recovery as stalled before finishing the recovery, leading to an infinite loop of recoveries
2018-07-07 17:41:20 -07:00
d6c6e7d306
fix: do not attempt data movement to an unhealthy destination team
...
allow building more teams than desired if all teams are unhealthy
bestTeamStuck is an error in simulation again
2018-07-07 16:51:16 -07:00
cdafd542ee
fix: fixed a memory leak where leaderInfo notifications are not cleared out
2018-07-06 17:40:29 -07:00
7a12d3e130
added the (untested) ability to force a recovery to the remote datacenter, even if that results in data loss. If the DR lag is more than 1 week there could be potential data corruption if any primary storage servers are still alive.
2018-07-01 09:39:04 -04:00
1f02bdee0a
do not buggify future version delay, because remote storage servers will be delayed getting data so they need additional time
2018-06-29 11:29:22 -07:00
7e68bee692
update better machine classes first to give them a higher chance of becoming the next cluster controller
2018-06-29 01:11:59 -07:00
58c2f67ff6
checking outstanding requests can be CPU intensive, so rate limit checking requests
2018-06-27 23:02:08 -07:00
5fc8199abc
Swapped OkayFit and UnsetFit, because generally if machine classes are set on one machine they are set everywhere and it helps with wait_for_good_recruitment logic
...
wait_for_good_recruitment now requires that you have the desired count of each roll
remote recruitment is given a much longer wait_for_good_recruitment time interval, which does not start until enough remote machines have registered
2018-06-22 10:15:24 -07:00
678b4494f4
added logging for the datacenter version difference
2018-06-21 16:31:52 -07:00
b79feaddd3
added a hard memory limit to the TLog to prevent it from running out of memory. Because remote logs are not ratekeeper controlled this is their only protection
2018-06-18 17:22:40 -07:00
889889323e
The master will tell the cluster controller if it is going to take a long time to recruit new logs in its DC; the cluster controller can determine if the other DC would be better and recruit there.
...
The cluster controller will not switch to the other data center if remote logs are too far behind.
We will not recruit in DCs with negative priority.
2018-06-13 18:14:14 -07:00
372ed67497
Merge branch 'master' into feature-remote-logs
...
# Conflicts:
# fdbserver/DataDistribution.actor.cpp
# fdbserver/MasterProxyServer.actor.cpp
# fdbserver/TLogServer.actor.cpp
# fdbserver/TagPartitionedLogSystem.actor.cpp
2018-06-11 11:34:10 -07:00
4903df5ce9
fix: give time to detect failed servers before building teams
2018-06-10 20:21:39 -07:00
529d0497f1
Proxy going OOM when applying high volumes of writes to a proxy, particular in a sudden fashion before ratekeeper can control the workload.
...
Address this issue by proactively monitoring the memory used by commit batches and dropping requests if a certain memory limit is exceeded.
2018-06-01 15:21:40 -07:00
e8f6ad88f0
fix: tripled the smallStorageTarget to prevent simulations which do a lot of work from timing out
2018-05-07 17:26:44 -07:00
8cb8198250
fix: the e-brake should be buggified with ratekeeper storage limits to prevent simulation from running full blast into the e-brake resulting in simulation taking forever to complete (joshua timeouts)
2018-05-06 12:33:25 -07:00
12ef63b698
knobify replace contents bytes
2018-05-01 19:43:35 -07:00
9c8cb445d6
optimized the tlog to use a vector for tags instead of a map
2018-03-17 10:36:19 -07:00
ccd70fd005
The tlog uses the tags embedded in the message instead of a separate vector of locations
...
optimized remote tlog committing to avoid re-serializing the message
2018-03-16 16:47:05 -07:00
f2c804e14f
Reverting changes from merge of master into release-5.2 ( b25810711c). Note that we never intend to release master into release-5.2, but if we did we would need to revert this commit.
2018-03-06 10:15:04 -08:00
37a6a81634
Merge commit '7f6fc3e039c911cd84b8540f7f799fc38a1c1822' into feature-remote-logs
...
# Conflicts:
# fdbserver/workloads/RestartRecovery.actor.cpp
2018-02-23 12:33:28 -08:00
0bae9880f1
remove trailing whitespace from our copyright headers ; fixed formatting of python setup.py
2018-02-21 10:25:11 -08:00
42405c78a5
Merge commit '4038bd2fd968d88861f2cebd442ce511724816cb' into feature-remote-logs
...
# Conflicts:
# fdbserver/ClusterController.actor.cpp
# fdbserver/Knobs.cpp
2018-02-10 12:08:52 -08:00
c7b3be5b19
re-enabled better master exists
...
the cluster controller can choose a better data center for itself and let the workers know where the next cluster controller should be recruited
2018-02-09 16:48:55 -08:00
d0caffd339
fix: knob was set to incorrect value
2018-02-06 18:11:45 -08:00
b7dde88029
fix: the cluster controller did not consider the master sharing the same process as the cluster controller as bad in all needed locations
...
waited too long for good recruitment locations, which would add too much time to recoveries of clusters that do not use machine classes
2018-02-06 11:30:05 -08:00
21482a45e1
Merge branch 'master' into feature-remote-logs
...
# Conflicts:
# fdbserver/DBCoreState.h
# fdbserver/LogSystem.h
# fdbserver/LogSystemPeekCursor.actor.cpp
# fdbserver/TLogServer.actor.cpp
2018-01-14 13:40:24 -08:00
022df3b91b
backup and restore sometimes took too long in simulation
2018-01-09 17:26:42 -08:00
5ac4f73978
Merge branch 'release-5.1' into feature-remote-logs
...
# Conflicts:
# fdbclient/NativeAPI.actor.cpp
# fdbrpc/Locality.h
# fdbrpc/simulator.h
# fdbserver/ApplyMetadataMutation.h
# fdbserver/ClusterController.actor.cpp
# fdbserver/LogSystemPeekCursor.actor.cpp
# fdbserver/MasterProxyServer.actor.cpp
# fdbserver/SimulatedCluster.actor.cpp
# fdbserver/TLogServer.actor.cpp
# fdbserver/TagPartitionedLogSystem.actor.cpp
# fdbserver/WorkerInterface.h
# fdbserver/masterserver.actor.cpp
# flow/Net2.actor.cpp
# tests/fast/SidebandWithStatus.txt
# tests/rare/LargeApiCorrectnessStatus.txt
# tests/slow/DDBalanceAndRemoveStatus.txt
2018-01-05 11:33:42 -08:00
bb1297c686
Remove RkServerQueueInfo and RkTLogQueueInfo trace events, since this information is more or less already logged on the storage servers and tlogs. Update the quiet database check and magnesium to use the information from the logs and storage servers.
2017-11-14 12:59:42 -08:00
3a4078bdda
the keyservers shards are always a fixed large size
2017-10-27 11:52:11 -07:00
f997cb9038
Add a string knob to hold the Log directory, and write profiles to it.
...
This is the combination of two small changes.
1. Add support for a string knob type.
2. Change profiles to be written to the log directory instead of the working
directory.
We have three options of where to write files: the working directory, the data
directory, and the log directory.
The working directory may be set to a non-writable location, and likely
contains the fdb binaries. Allowing these files to be overwritten would likely
not be a wise idea.
The data directory hosts our sqlite b-trees. It would also be very unfortunate
if these were ever overwritten by an unfortunate profile name.
The log directory contains logs. Out of the three, these matter the least if
they disappear or become corrupted.
Thus, we write to the log directory.
2017-10-16 16:05:02 -07:00
15962cf079
Merge branch 'master' into feature-remote-logs
...
# Conflicts:
# fdbrpc/Locality.cpp
# fdbrpc/Locality.h
# fdbserver/ClusterController.actor.cpp
# fdbserver/ClusterRecruitmentInterface.h
# fdbserver/TLogServer.actor.cpp
# fdbserver/TagPartitionedLogSystem.actor.cpp
# fdbserver/WorkerInterface.h
# fdbserver/fdbserver.vcxproj.filters
# fdbserver/masterserver.actor.cpp
# fdbserver/worker.actor.cpp
# flow/error_definitions.h
2017-10-05 17:09:44 -07:00
0f8ff26029
Merge pull request #158 from bmuppana/master
...
<rdar://problem/34557380> Need a way to map real time to version
2017-09-27 17:56:42 -07:00
6a0b1d6808
Fixing PR comments
...
<rdar://problem/34557380> Need a way to map real time to version
2017-09-27 17:56:01 -07:00
Evan Tschannen