The subscriber write goroutine would drop points if the write load
was higher than it could process. This could happen with a just
a few writers to the server.
Instead, process the channel with multiple writers to avoid dropping
writes so easily. This also adds some config options to control how
large the channel buffer is as well as how many goroutines are started.
Fixes#7330
The tags passed into Statistics() calls are not supposed to be modified.
The balanceWriter in subscribers tried to modify them triggering a panic
because they can be nil.
The vet checks for some files did not pass for go 1.7. As part of a
preliminary start to making go 1.7 work with this software, go vet
should pass.
Also updated the gogo/protobuf dependency which fixed the code generator
to work with go 1.7 too. Ran `go generate` on the entire repository to
ensure every file was up to date.
When we refactored expvar, the cmdline and memstats sections were not
readded to the output. This adds it back if they can be found inside of
`expvar`.
It also stops trying to sort the output of the statistics so they get
returned faster. JSON doesn't need them to be sorted and it causes
enough latency problems that sorting them hurts performance.
Instead of having the parser set the defaults, the command will set the
defaults so that the constants for that are actually used. This way we
can also identify which things the user provided and which ones we are
filling with default values.
This allows the meta client to be able to make smarter decisions when
determining if the user requested a conflict or if the requested
capabilities match with what is currently available. If you just say
`CREATE DATABASE WITH NAME myrp`, the user doesn't really care what the
duration of the retention policy is and just wants to use the default.
Now, we can use that information to determine if an existing retention
policy would conflict with what the user requested rather than returning
an error if a default value ever gets changed since the meta client
command can communicate intent more easily.
Previously, we implicitly added a newline and had to add one to the
number of bytes transmitted because we added that byte. That was removed
at some point and the metric was not updated to record the correct
value.
The query killing functionality depends on the ResponseWriter exposing a
CloseNotify method. Since we wrap the http.ResponseWriter, the new
struct does not have that method and the HTTP handler would skip past
calling that method.
Instead of duplicating `Flush()` and `CloseNotify()` for every response
formatter, we will unify all of that under a single struct and create
formatters instead.
Also, fixes a bug where the header information from a query would not be
returned until some other data was returned with it because of
buffering and another bug in the gzipResponseWriter that wouldn't flush
the actual underlying ResponseWriter.
The query can be uploaded from a file using `multipart/form-data` and
setting the file name to `q`. An example of using curl to execute an
async query would be:
curl -F "q=@database.iql" -F "async=true" http://localhost:8086/query
It will return a 204 No Content as long as the query is accepted
(immediate errors will be returned, but not individual errors with
specific queries). The only way to kill the query is by using the task
manager.
CSV doesn't offer a way to separate different sheets from each other and
it doesn't really have a standard format. We separate sheets with a
newline so they can be imported into something like Excel or LibreOffice
more easily.
The number of columns for each sheet is inferred from the first returned
row in each statement since they should all be the same.
Instead of having the parser set the defaults, the command will set the
defaults so that the constants for that are actually used. This way we
can also identify which things the user provided and which ones we are
filling with default values.
This allows the meta client to be able to make smarter decisions when
determining if the user requested a conflict or if the requested
capabilities match with what is currently available. If you just say
`CREATE DATABASE WITH NAME myrp`, the user doesn't really care what the
duration of the retention policy is and just wants to use the default.
Now, we can use that information to determine if an existing retention
policy would conflict with what the user requested rather than returning
an error if a default value ever gets changed since the meta client
command can communicate intent more easily.
This commit limits queries to only process one shard at a time.
However, within a shard, multiple series can still be processed in
parallel. Shard iterators are lazily instantiated during query
execution to limit the amount of memory a given query uses.
The `Content-Type` header is meant to say what the content type of the
request is supposed to be and `Accept` is used to ask for a specific
content type. I messed this up and used `Content-Type` instead of
`Accept`. This works with `GET` requests accidentally, but `POST`
requests stop working.
According to the HTTP standard, a lack of authentication credentials or
incorrect authentication credentials should send back a 401
(Unauthorized) with a `WWW-Authenticate` header with a challenge that
can be used to authenticate. This is because a 401 status should be sent
when an authentication attempt can be retried by the browser.
The 403 (Forbidden) status code should be sent when authentication
succeeded, but the user does not have the necessary authorization.
Previously, the server would always send a 401 status code.
Truncate the time interval output of the monitor service to be on even
time intervals rather than on every minute based on the start time. This
normalizes the output from the monitor service.
Similar to #1339 .
Fix getClientVersion() on setups where the admin interface resides on some
other path than root path due to some reverse proxy setup.
Normalize the output for the various help options so they all follow the
same format and display all relevant options.
Removing some of the unused config options from the configuration file
and updating the help documentation. Removing some remaining references
to clustering within the open source version.
The tsdb package had a substantial amount of dead code related to the
old query engine still in there. It is no longer used, so it was removed
since it was left unmaintained. There is likely still more code that is
the same, but wasn't found as part of this code cleanup.
influxql has dead code show up because of the code generation so it is
not included in this pruning.
The highest time represented by a nanosecond needs to be used for an
exclusive range, so the maximum time needs to be one less than the
possible maximum number of nanoseconds representable by an int64 so that
we don't lose a point at that one time.
Previously worked in the open source version because the timestamp used
for finding a shard would be truncated by the retention policy so the
lookup time didn't run into this edge case because it didn't rest on the
truncation boundary. Since that point didn't really belong in that shard
group and was placed there by mistake, it's best to fix this bug since
the timestamp used to create the shard group should be capable of
retrieving it.
changes the httpd log lines from this:
[httpd] 2016/06/08 14:06:39 ::1 - - [08/Jun/2016:14:06:39 +0100] POST /write?consistency=any&db=telegraf&precision=s&rp= HTTP/1.1 204 0 - InfluxDBClient d6aa01fc-2d79-11e6-8024-000000000000 2.751391ms
to this:
[httpd] ::1 - - [08/Jun/2016:14:06:39 +0100] "POST /write?consistency=any&db=telegraf&precision=s&rp= HTTP/1.1" 204 0 "-" "InfluxDBClient" d6aa01fc-2d79-11e6-8024-000000000000 2751
So it changes a few things:
1. Remove the logger timestamp at the beginning which isn't very relevant anyways
2. adds quotes around "METHOD URI PROTOCOL", because this is part of the
common log format.
3. adds quotes around "AGENT" and "REFERRER" because this is part of the
"combined" log format.
4. Puts the response time in integer microseconds, because this is
consistent with apache's %D config mod option.
Compared with CLF, our logs now look like this:
[httpd] %{COMMON_LOG_FORMAT} "<agent>" "<referrer>" <request_uuid> <response_time_µs>
For reference, see:
https://en.wikipedia.org/wiki/Common_Log_Formathttp://httpd.apache.org/docs/current/mod/mod_log_config.html
The task manager now acts as its own statement executor so that a custom
statement executor can perform custom actions for KillQueryStatement and
ShowQueriesStatement.
The graphite service will attempt to create the retention policy and use
it. If the retention policy doesn't exist, it will be created with the
default options.
Fixes#5655.
This allows us to add additional options to ExecuteQuery without
creating parameter bloat.
Removing the unused Series structs. Their necessity was removed by a
previous commit, but the structs were not removed yet.
Add another type of interrupt iterator that monitors the interrupt
channel and calls `Close()` on the iterator when the interrupt happens.
It will primarily be used for asynchronously closing the ReaderIterator,
but it will only close the read side of the connection properly. More
work needs to be done to allow closing the write side efficiently.
The default retention policy name is changed to "autogen" instead of
"default" since it ends up being ambiguous when we tell a user to check
the default retention policy, it is uncertain if we are referring to the
default retention policy (which can be changed) or the retention policy
with the name "default".
Now the automatically generated retention policy name is "autogen".
The default retention policy is now also configurable through the
configuration file so an administrator can customize what they think
should be the default.
Fixes#3733.
Use the latest documentation by default if the server version can't be
found. If it can be found, update the documentation link to that
specific version so it always points at the exact documentation relevant
for the server version.
Fixes#5906.
The HTTPS configuration for the httpd service only had an option to
specify the certificate file and the same file would be used for both
the certificate and private key file (they could be concatenated
together).
This adds an additional option to specify the files differently from
each other while still allowing the previous behavior. If only
`https-certificate` is specified, the httpd service will try to load the
private key from the `https-certificate` file. If a separate
`https-private-key` file is specified, the private key will be loaded
from there instead.
Fixes#1310.
The parser can be passed a map of keys to literal values to be replaced
into the query. Parameters are preceded by a dollar sign (`$`). If a
parameter key is missing, an error is thrown by the parser.
Fixes#2926.
When authenticating a request, check that an admin user exists instead
of checking for len(users) > 0. This prevents getting stuck with no
admin user and being unable to create one.
The http connection limit is for any HTTP operation and is independent
of the other connection limits. It should be set to a higher value than
the query limit. The difference between this and the query limit is it
will close out the connection immediately without any further
processing.
This is the equivalent of the `max_connections` option in PostgreSQL.
Also removes some unused config options from the cluster config.
Fixes#6559.
In order to follow REST a bit more carefully, all write operations
should go through a POST in the future. We still allow read operations
through either GET or POST (similar to the Graphite /render endpoint),
but write operations will trigger a returned warning as part of the JSON
response and will eventually return an error.
Also updates the Golang client libraries to always use POST instead of
GET.
Fixes#6290.
It can sometimes be difficult to determine if a query submitted by
the UI has been accepted for execution.
In particular, if a query that returns empty results is followed
by a query that takes a long time, then the green success message
from the first query may be misleadingly displayed to the user when,
in fact, the second query is still in progress.
To address this, we always hide the query error and success divs
before the query is submitted. Previously, just the results were
cleared.
Secondly, if the user re-runs a query expecting slightly different results,
it can also be unclear whether the second attempt to execute the command
is simply redisplaying the results of the previous query or the results of the
resubmission, this is particularly true if the queries involved
have short execution times.
To support the ability to distinguish these two cases, we have any attempt
to use the history arrow also clear the results and status divs. This
reduces the ambiguity about whether the next results display is, in fact,
the result of executing a new query.
Signed-off-by: Jon Seymour <jon@wildducktheories.com>
Sanitizing is now done through pattern matching rather than parsing the
query and replacing the password in the query. This prevents
accidentally redacting the wrong part of a query and revealing what the
password is through association.
Fixes#3883.
This has various benefits:
- Users embedding InfluxDB within other Go programs can specify a different logger / prefix easily.
- More consistent with code used elsewhere in InfluxDB (e.g. services, other `run.Server.*` fields, etc).
- This is also more efficient, because it means `executeQuery` no longer allocates a single `*log.Logger` each time it is called.
The deprecated message is now attached to a new attribute returned with
the results. This message can then be read by clients to warn a user
about upcoming changes to the query engine.
The `influx` client has already been modified to read this message and
print it out for every format except CSV.
The first warning message is a deprecated message about removing `IF NOT
EXISTS` from `CREATE DATABASE`.
The message will also be printed to the server log.
Fixes#5707.
The QueryExecutor had a lot of dead code made obsolete by the query
engine refactor that has now been removed. The TSDBStore interface has
also been cleaned up so we can have multiple implementations of this
(such as a local and remote version).
A StatementExecutor interface has been created for adding custom
functionality to the QueryExecutor that may not be available in the open
source version. The QueryExecutor delegate all statement execution to
the StatementExecutor and the QueryExecutor will only keep track of
housekeeping. Implementing additional queries is as simple as wrapping
the cluster.StatementExecutor struct or replacing it with something
completely different.
The PointsWriter in the QueryExecutor has been changed to a simple
interface that implements the one method needed by the query executor.
This is to allow different PointsWriter implementations to be used by
the QueryExecutor. It has also been moved into the StatementExecutor
instead.
The TSDBStore interface has now been modified to contain the code for
creating an IteratorCreator. This is so the underlying TSDBStore can
implement different ways of accessing the underlying shards rather than
always having to access each shard individually (such as batch
requests).
Remove the show servers handling. This isn't a valid command in the open
source version of InfluxDB anymore.
The QueryManager interface is now built into QueryExecutor and is no
longer necessary. The StatementExecutor and QueryExecutor split allows
task management to much more easily be built into QueryExecutor rather
than as a separate struct.
A change to the admin UI prevented the success message being displayed
for empty results. This change restores the original behaviour for
this case.
Signed-off-by: Jon Seymour <jon@wildducktheories.com>
FormValue() would attempt to parse the body of a request when the
content-type is set to `application/x-www-form-urlencoded`. The write
handler never wants url-encoded forms, and should only ever check the
URL for query parameters.
Fixes#6061
It's possible for a single query to send multiple results that get
aggregated in the HTTP handler. If an earlier result passed in data and
a later result had an error, the error would be ignored.
Now an error for a statement will overwrite any previous results for
that statement.
Partially fixes#6094.
Prior to this when passing the same query and CQ name in a CREATE
CONTINUOUS QUERY command an error would be returned. This means the
command was not behaving in a similar way to other commands.a
Now when running the command with the same CQ name and query string no
error will be returned. Note, this change does not parse the query, it
simply compares a normalised query string to the existing one on the CQ.
Partially addresses #6094.
Previously, when creating a retention policy only the name was
considered when deciding if the policy already existed. This meant that
adding a second policy with the same name but different duration or
replica factor returned the original policy and no error.
This commit fixes that and ensures that name, duration and replica
factor are all considered.
Allows configuration of shard group duration at database creation, and retention
policy create/alter time.
Query examples:
```
CREATE DATABASE testdb WITH DURATION 90d SHARD DURATION 30m NAME rp_testdb
CREATE RETENTION POLICY rp_testdb2 ON testdb DURATION INF REPLICATION 1 SHARD DURATION 30m
ALTER RETENTION POLICY rp_testdb2 ON testdb SHARD DURATION 1h
```
This can be useful with long duration retention policies with lots of data, where
you can split into smaller shards to relieve memory pressure.
This seems to have been an oversight since all of the response writers
are supposed to implement this interface, but the gzipResponseWriter
didn't implement this interface for some reason.
The currently running queries can be listed with the command
`SHOW QUERIES` and it will display the current commands that have been
run, the database they were run against, and how long they have been
running.
Go 1.4.3 was a security release that also created a strange edge-case
that caused connections to not be kept alive and reused when Close()
is called on the Body of the request. Close() hasn't been required on
the Body of a request for some time, so there is no harm is not calling
it anymore.
... by extracting the db/rp from the given path.
Now that the code has "standardized" on extracting db/rp this way, the
ShardLocation struct is no longer necessary and thus has been removed.
We're back on the previous style of passing the path and walPath to
NewShard.
This commit updates tsdb.Shard to contain a ShardConfig and updates
tsdb.Store to directly reference a map of tsdb.Shard rather than the
previous tsdb.shardLocation abstraction.
Previously, CQs with the same name would be stored in the last run map
the same way. This caused only one of the CQs to run because after the
first one ran it would update the last run time for all CQs with the
same name.
Add the database name to the CQ ID in the last run map to differentiate
between CQs in different databases.
Fixes#5814.
Fixes#5612, #5573 and #5518.
Using the MetaExecuter, queries that need to run on both data nodes
and optionally the meta store will be executed across all data nodes
in the cluster.
Fixes#5680.
When dropping a data node, the following will now happen on the
Meta Store.
1) If any shards no longer have any owners (because the data node
being dropped is the only owner), they will be reassigned a
new owner from within their respective shard group.
2) If a shard group no longer has any shards/data nodes, they will
be marked as deleted.
When a shard is being assigned a new owner a data node with the fewest
number of shards in the shard group will be selected as the new owner.
Finally, checking the validity of a data node's ID now happens in the
Meta store, rather than in the state machine.
Jason Wilder
Nathaniel Cook
Mike Glazer
Edd Robinson
Joe LeGasse
Jason Wilder
Jonathan A. Sternberg
Todd Persen
Dan Dascalescu
Ben Johnson
kun
Cameron Sparr
David Norton
Mikael Taina
Cole Mackenzie
simon
Jack Zampolin
Rubycut
Joe Farrell
minkikim89
Nathaniel Cook
Cory LaNou
Todd Persen
Seif Lotfy
Vidhuran Harichandra Babu
Jon Seymour
Stephen Gutekanst
Jonathan A. Sternberg
Aaron Knister
Chris Ramón
Gunnar
Valeriy Simonov
gunnaraasen
Regan Kuchan
mvadu
Tait Clarridge
Mark Rushakoff
Mark Rushakoff