Commit Graph

169 Commits

Author SHA1 Message Date
Amnon Heiman
f0d2df3d15 StorageProxy.java: Add view write metrics
nodetool proxyhistograms command look for the view write metric.

While we do not report that metric yet, we still want the command to
succeed.

After this patch:
$ nodetool proxyhistograms
proxy histograms
Percentile       Read Latency      Write Latency      Range Latency   CAS Read Latency  CAS Write Latency View Write Latency
                     (micros)           (micros)           (micros)           (micros)           (micros)           (micros)
50%                    326.00             110.00             424.50               0.00               0.00               0.00
75%                   1253.00             193.25             877.75               0.00               0.00               0.00
95%                   2935.90            1007.25            5182.55               0.00               0.00               0.00
98%                   3100.00            1040.60            5492.00               0.00               0.00               0.00
99%                   3100.00            1058.00            5492.00               0.00               0.00               0.00
Min                     34.00               9.00              36.00               0.00               0.00               0.00
Max                   3100.00            1058.00            5492.00               0.00               0.00               0.00

See scylladb/scylla#4470

Signed-off-by: Amnon Heiman <amnon@scylladb.com>
Message-Id: <20190514063316.28040-1-amnon@scylladb.com>
2019-07-21 19:20:37 +03:00
Amnon Heiman
9dae28e2f0 ColumnFamilyStore: finishLocalSampling should respect count limit
When calling nodetool toppartitions with size limit, finishLocalSampling
should respect that and limit the number of the results.

Example:
$ nodetool toppartitions -k 2 keyspace1 standard1 20
WRITES Sampler:
  Cardinality: ~2 (256 capacity)
  Top 2 partitions:
	Partition                Count       +/-
	38333032394d4f4d5030         4         3
	4e353937383137503330         4         3

READS Sampler:
  Cardinality: ~2 (256 capacity)
  Top 2 partitions:
	Nothing recorded during sampling period...

Fixes #66

Signed-off-by: Amnon Heiman <amnon@scylladb.com>
2019-06-23 15:31:18 +03:00
Calle Wilund
5f974bc2bb ColumnFamilyStore: Propapgate exception cause in sampling wait
Message-Id: <20190311082942.3310-1-calle@scylladb.com>
2019-05-05 18:10:37 +03:00
Calle Wilund
da21305989 StorageService: Include the arguments in "upgrade" call.
Message-Id: <20190219133431.29009-1-calle@scylladb.com>
2019-02-27 10:33:43 +02:00
Amnon Heiman
27313ee2c4 ColumnFamilyStore: Add an implementation for table sampling
This patch adds the implementation for begin and finish local sampling
of a column family.

There is a difference in the implementation of Cassandra API and Scylla.

In Cassandra and the JMX an external source start and stop the sampling.

In Scylla, a single API call start the sampling and return with the
result. In Scylla the API call always return sampling of the read and of
the writes.

To bridge the difference, the begin sampling command will use a Future
when calling the API. The finish method will wait for the future to end.

Because of the different implementation, it is possible that two
consecutive calls will be made to start sampling one for the read and
one for the write, similarly, two calls will be made to finish for read
and write.

The implementation would ignore the second call to start and will
store the result, so the second call to finish will be served from the
stored result.

Note, that the use of future is only for safety, the way we expect it to
work, the caller to the begin sampling will sleep anyhow while waiting
for the result.

To avoid breaking the MBean compatibility we piggyback the duration on
top of the sampler string.

If no duration is given, a default duration will be taken, this is also
just as a precaution, we will modify the nodetool implementation to
pass that information.

There is a known issue with cardinality, that will need to be addressed.
Also we return a value in the raw column to match what Cassandra JMX
returns, but it's a duplication of the partition key.

See scylladb/scylla#2811

Signed-off-by: Amnon Heiman <amnon@scylladb.com>
Message-Id: <20190128143505.5241-1-amnon@scylladb.com>
2019-02-03 12:40:04 +02:00
Calle Wilund
ca3fa8de20 scylla-jmx: Fix tablemetricsobjectname breakage
Fixes #57

The usage of TableMetricsObjectName-yada-yada relies on translating the
"fake" objectname to a canonical one on remote
publication/serialization. However, the implementation of
ObjectName.getInstance has changed in JDK (micro) updates so it no
longer applies overridable methods -> wrong name published.

Fix by doing explicit ObjectName instansiation.
Message-Id: <20181023132005.23099-1-calle@scylladb.com>
2018-10-23 16:30:29 +03:00
Calle Wilund
c6aee9f63e scylla-jmx: Add "PendingTasksByTableName" gauge to CompactionMetrics
Required by origin 3.11 nodetool.

Message-Id: <20180801084545.23239-1-calle@scylladb.com>
2018-08-01 14:25:06 +03:00
Calle Wilund
9c3ac3e547 scylla-jmx: Update JMX interfaces to origin 3.11
Almost 100% null implementations, which is ok for most purposes
currently used by scylla. Some of these new calls (like dropped
mutations etc) should perhaps however be implemented.

Tested with the nodetool dtests. So sparsely.

Needed when/if scylla-tools-java is upgraded to origin 3.11,
otherwise noodtool breaks.

Message-Id: <20180730113741.14952-1-calle@scylladb.com>
2018-07-30 15:47:43 +03:00
Piotr Jastrzebski
455f5717ea Introduce and use TableMetricObjectName
This is a new extention of ObjectName that uses less memory.

TableMetricNameFactory and AllTableMetricNameFactory can
create it instead of regular ObjectName to save memory.

It is possible to save memory because each name created by
TableMetricNameFactory (or AllTableMetricNameFactory) shares
most of its data with other names created by the same factory
and there's no need to create multiple copies.

Signed-off-by: Piotr Jastrzebski <piotr@scylladb.com>
2018-05-12 19:08:37 +02:00
Piotr Jastrzebski
2c48bab91a Use JmxMBeanServer instead of MBeanServer
JmxMBeanServer is a concrete implementation of a MBeanServer.
We want to use it directly because we need to bypass calls to
JmxMBeanServer.registerMBean and JmxMBeanServer.unregisterMBean.
They take ObjectName as parameter, copy it using
ObjectName.getInstance(ObjectName) and pass it to registerMBean
and unregisterMBean of JmxMBeanServer.getMBeanServerInterceptor().
We want to avoid this copy and pass the ObjectName argument directly
to JmxMBeanServer.getMBeanServerInterceptor().

To do that this patch:
1. changes all MBeanServer variables to JmxMBeanServer
2. creates JmxMBeanServer in APIBuilder making sure accessing
   interceptors is allowed
3. makes sure that JmxMBeanServer.getMBeanServerInterceptor().registerMBean
   is called wherever JmxMBeanServer.registerMBean was called
4. makes sure that JmxMBeanServer.getMBeanServerInterceptor().unregisterMBean
   is called whenever JmxMBeanServer.unregisterMBean was called

Next patch will use different ObjectName implementation that will
use less memory and this patch is crucial because without it every ObjectName
is transformed with ObjectName.getInstance which turns the object into
a regular ObjectName.

Signed-off-by: Piotr Jastrzebski <piotr@scylladb.com>
2018-05-12 18:35:18 +02:00
Duarte Nunes
55abaa1bc8 StorageService: Allow querying the view build status
Signed-off-by: Duarte Nunes <duarte@scylladb.com>
Message-Id: <20180327002342.11494-1-duarte@scylladb.com>
2018-04-03 14:43:27 +03:00
Amnon Heiman
4e4589ba6f FailureDetector: check that states is not null before use
When a node is part of a cluster but is down (like in the situation where
a cluster is taken down and up again but not all nodes are up). There is
no application_state information for that node.

This patch check that the information exists before using it to prevent
null pointer exception.

After this patch, a call to nodetool gossipinfo would return the
available information without failing.

See scylladb/scylla#3330

Signed-off-by: Amnon Heiman <amnon@scylladb.com>
Message-Id: <20180329115345.29357-1-amnon@scylladb.com>
2018-03-29 15:18:48 +03:00
Glauber Costa
2b1ed89ec3 fix auto_compaction request strings
Column family is not being passed, so the requests fail to reach the
correct endpoint.

Signed-off-by: Glauber Costa <glauber@scylladb.com>
2017-08-21 22:55:31 -04:00
Calle Wilund
e954db8444 TableMetrics: bugfix: local metrics registry is a proxy. It must proxy
Refs scylladb/scylla#2340 (trunk/1.7)

Must proxy "register" call, otherwise unregistration of mbeans
will instead try to double-register. Code for this somehow fell away.

Message-Id: <1494417610-9720-1-git-send-email-calle@scylladb.com>
2017-05-28 14:02:32 +03:00
Tomasz Grabiec
8176d5729f Bring back old forceKeyspaceCompaction() overload
nodetool from scylla-tools-java still uses it. Currently `nodetool compact` fails like this:

error: forceKeyspaceCompaction(java.lang.String, [Ljava.lang.String;)
-- StackTrace --
java.lang.NoSuchMethodException: forceKeyspaceCompaction(java.lang.String, [Ljava.lang.String;)

Fixes scylladb/scylla#2261

Probably broken by 3e146845b4

Message-Id: <1491470483-6147-1-git-send-email-tgrabiec@scylladb.com>
2017-04-06 12:59:52 +03:00
Amnon Heiman
0c541d73e7 MetricsRegistry: Solving empty histograms in nodetool
This patch fixes two issues with the histogram implementation:
* Need to all update before trying to read values from the histogram.
* The histogram values return from the API in microseconds and not nano.

See Scylladb/scylla#2155

Signed-off-by: Amnon Heiman <amnon@scylladb.com>
Message-Id: <20170315130725.22261-1-amnon@scylladb.com>
2017-03-20 12:33:45 +02:00
Amnon Heiman
0c7afef8f4 Keep tokensEndPointMap sorted by the API order
This patch change the sorting of tokensEndpointMap so it will use the
order returned by the API.

See Scylladb/scylla#1945

Signed-off-by: Amnon Heiman <amnon@scylladb.com>
2016-12-28 12:48:03 +02:00
Calle Wilund
6f916b9d8e scylla-jmx: Add dummy "compactionId" to compation info
To keep nodetool happy. Should maybe add actual Ids to compations.
Message-Id: <1482335118-9595-2-git-send-email-calle@scylladb.com>
2016-12-22 14:17:38 +02:00
Calle Wilund
85c3293ef1 scylla-jmx: Missing getter in MBean interface.
It got lost somewhere.
Message-Id: <1482335118-9595-1-git-send-email-calle@scylladb.com>
2016-12-22 14:17:13 +02:00
Calle Wilund
ae6a000807 ColumnFamilyStore: Remove compaction parameter API usage
Do manual mangling of in/out data in JMX instead. Saves on
controversy over more or less pointless API additions.
2016-11-01 09:44:17 +00:00
elcallio
434ce947b0 Code formatting + source cleanup (eclipse) 2016-10-24 11:43:52 +00:00
elcallio
f4f3c44dc1 Rework StreamManager 2016-10-24 11:43:52 +00:00
Calle Wilund
4ed049739a Storage service: Fix 3.x style notifications (repair) 2016-10-24 11:43:51 +00:00
elcallio
4ec7d58249 Rework service.* beans 2016-10-24 11:43:51 +00:00
elcallio
fec8b44942 Rework MessagingService 2016-10-24 11:43:51 +00:00
Calle Wilund
3fe9cfc232 EndpointSnitchInfo: Fix getRack/DC host handling
I.e. our localhost might be (and probably is) different from scyllas
"fb::broadcast", and if not, try to get numerical asap.
2016-10-24 11:43:51 +00:00
elcallio
21a343d003 Rework EnpointSnitchInfo 2016-10-24 11:43:51 +00:00
elcallio
80762eb60a Rework gms beans 2016-10-24 11:43:51 +00:00
elcallio
e49b4ef322 Rework CompactionManager 2016-10-24 11:43:51 +00:00
elcallio
1470b37193 Rework CommitLog 2016-10-24 11:43:51 +00:00
elcallio
e55863e375 Rework ColumnFamilyStore 2016-10-24 11:43:51 +00:00
elcallio
cd9deafc51 Rework all org.apache.cassandra.metrics types to new style
I.e. bind only JMX object via registry.
2016-10-24 11:43:51 +00:00
elcallio
319dadb79c Add TableMetrics - c3 version of ColumnFamilyMetrics
Using new, slimmer, metrics binding
2016-10-24 11:43:51 +00:00
elcallio
a44c18c621 Add metric/mbean base types + metrics JMX object factory 2016-10-24 11:43:51 +00:00
Calle Wilund
3e146845b4 StorageService: update to c3 compat
Note: some calls that are not (yet) applicable to scylla are 
unimplemented.
2016-10-24 11:43:51 +00:00
Calle Wilund
b4e483b179 StorageProxy: update to c3 compat 2016-10-24 11:43:51 +00:00
Calle Wilund
de28e68532 GCInspector: Add SuppressWarnings("restriction") 2016-10-24 11:43:51 +00:00
Calle Wilund
3a4adcb676 CacheService: update to c3 compat 2016-10-24 11:43:51 +00:00
Calle Wilund
85e1b07544 MessagingService: update to c3 compat
Note: c3 adds configurable size threshold counting of messages sent, 
dividing info "large"/"small" partitions (+gossiper). Message bulk 
queries in v3 mbean reflects this. 

Scylla does not (yet?) have such a threshold divider, so this is 
highly incomplete and just delegates to old apis that "sort-of" fit.
2016-10-24 11:43:51 +00:00
Calle Wilund
f4759f05e7 EndpointSnitchInfo: update to c3 compat 2016-10-24 11:43:51 +00:00
Calle Wilund
68ce437b03 Gossiper: update to c3 compat 2016-10-24 11:43:51 +00:00
Calle Wilund
b7a6554ee9 FailureDetector: update to c3 compat 2016-10-24 11:43:51 +00:00
Calle Wilund
3efcd5103b CompactionManager: update to c3 compat 2016-10-24 11:43:51 +00:00
Calle Wilund
39e4cd8f3f CommitLog: update to c3 compat 2016-10-24 11:43:51 +00:00
Calle Wilund
85b39d7fbe ColumnFamilyStore: update to c3 compat
Note: some calls still unimplemented
2016-10-24 11:43:51 +00:00
Amnon Heiman
1e4edeb858 MessagingService: Move to APITimer and drop the pulling
With the change to APITimer there is no longer a need to periodically
pull the API.

The verb will be register on the object initialization and will be
updated whenever they are been used.

Signed-off-by: Amnon Heiman <amnon@scylladb.com>
2016-05-17 13:04:55 +03:00
Amnon Heiman
7756f4751a DroppedMessageMetrics: Change APISettableMeter to APIMeter
The APIMeter replaces the APISettableMeter as the Meter implementation.

Signed-off-by: Amnon Heiman <amnon@scylladb.com>
2016-05-17 12:27:36 +03:00
Amnon Heiman
74d062851c CompactionMetrics: Use APITimer
This change the Timer from Timer to APITimer.

Signed-off-by: Amnon Heiman <amnon@scylladb.com>
2016-05-17 12:25:51 +03:00
Amnon Heiman
78bb2ef25a CacheService: Do not get non existing caches metrics
With the change in the meter implementation, retrieving a non existing
metrics would take time.

For this, the CacheService would mark caches that are not supported with
null url, so the metrics will be register but will return 0 for all
request (instead of going to the API that will return 0).

Signed-off-by: Amnon Heiman <amnon@scylladb.com>
2016-05-17 12:20:35 +03:00
Amnon Heiman
8a73d6a840 CacheMetrics: Switch to APITimer
CacheMetrics is a general counter that is used for all possible caches.
For caches that we do not support, there is no need to go and fetch
their values.

When moving to the APITimer each such request will take longer (the
value will not be available as it use it) now it will be possible to
supply a null as a url which would cause to the metrics to return 0 for
all counters.

Signed-off-by: Amnon Heiman <amnon@scylladb.com>
2016-05-17 11:32:59 +03:00