This commit removes the FAILED state for the IndexAuditTrail so that we always try to keep starting
the service. Previously, on any exception during startup we moved to a failed state and never tried
to start again. The users only option was to restart the node. This was problematic in the case of
large clusters as there could be common timeouts of cluster state listeners that would cause the
startup of this service to fail.
Additionally, the logic in the IndexAuditTrail to update the template on the current cluster has
been removed and replaced with the use of the TemplateUpgradeService. However, we still need to
maintain the ability to determine if a template on a remote cluster should be PUT. To avoid always
PUTing the template, the version field has been added so it only needs to be PUT once on upgrade.
Finally, the default queue size has been increased as this is another common issue that users hit
with high traffic clusters.
relates elastic/x-pack-elasticsearch#2658
Original commit: elastic/x-pack-elasticsearch@27e2ce7223
* Remove usage of Settings inside SqlSettings
Also hook client timeouts to the backend
Set UTC as default timezone when using CSV
As the JVM timezone changes, make sure to pin it to UTC since this is what the results are computed against
Original commit: elastic/x-pack-elasticsearch@3e7aad8c1f
For the purpose of getting this API consumed by our UI, returning
overall buckets that match the job's largest `bucket_span` can
result in too much data. The UI only ever displays a few buckets
in the swimlane. Their span depends on the time range selected and
the screen resolution, but it will only ever be a relatively
low number.
This PR adds the ability to aggregate overall buckets in a user
specified `bucket_span`. That `bucket_span` may be equal or
greater to the largest job's `bucket_span`. The `overall_score`
of the result overall buckets is the max score of the
corresponding overall buckets with a span equal to the job's
largest `bucket_span`.
The implementation is now chunking the bucket requests
as otherwise the aggregation would fail when too many buckets
are matching.
Original commit: elastic/x-pack-elasticsearch@981f7a40e5
This adds all of the security tests I think SQL is going to need for the initial release. SQL is still missing an entire scenario though: SSL enabled. Either way, this removes some `NOCOMMIT`s in `qa/sql/security`. Adding the SSL testing can come later.
Original commit: elastic/x-pack-elasticsearch@851620b606
Add security tests for SQL's CLI and JDBC features. I do this by factoring out all the "actions" from the existing REST tests into an interface and implement it for REST, CLI, and JDBC. This way we can share the same audit log assertions across tests and we can be sure that the REST, CLI, and JDBC tests cover all the same use cases.
Original commit: elastic/x-pack-elasticsearch@82ff66a520
This is *way* faster because we don't have to wait for the audit
events from previous test runs to drain into the index. And we
don't have to wait for the index's refresh cycle. We have to parse
the log lines which is a bit more brittle but it feels worth it
at this point.
Original commit: elastic/x-pack-elasticsearch@4b1758fc32
This commit replaces the REST test that the global checkpoint sync
action runs successfully as a privileged user. The test needs to be
replaced because it has a small race condition. Namely, the check that
the post-operation global checkpoint sync was successful could run
before the sync finishes running. To address this, we replace the REST
test with a test where we have a little more control and can assert busy
to avoid this race from failing the test.
Relates elastic/x-pack-elasticsearch#2749
Original commit: elastic/x-pack-elasticsearch@ea585b843c
Firstly, data in H2 is now stored in TIMESTAMP WITH TIME ZONE since H2
does not allow a global TZ to be set and picks the JVM TZ when a record
is read.
JdbcAssert is now aware of this allows TIMESTAMP with TZ == TIMESTAMP
Discovered a serious bug in DateTimeFunction - unfortunately date
histogram is not useful except for year since most extract functions
avoid ordering which a histogram preserves.
Thus most DTF are now terms aggs with scripting.
Improved a bug that caused duplicate functions to not be detected because
of aliasing.
Moved some datetime tests to CSV but the aggs tests now are in sync with
H2
Fixed bug that caused arithmetic on aggs to not be properly resolved by
splitting the processor definition tree to aggName (unresolved) and
aggPath (resolved)
Original commit: elastic/x-pack-elasticsearch@e75ada68f1
It was disabled because the CLI didn't work with security but
we've since fixed that so we can enable it.
Original commit: elastic/x-pack-elasticsearch@8d9b5ad89b
We weren't returning errors correctly from the server
or catching them correctly in the CLI. This fixes that
and adds simple integration tests.
Original commit: elastic/x-pack-elasticsearch@259da0da6f
After a write operation on an index, a post-operation global checkpoint
sync fires. Previously, this action fired on the same user as executed
the write action. If the user did not have priviledges for this action,
the global checkpoint sync would fail. With an upstream change in core,
this action now fires as the system user. This commit adds a test that
create a user that has minimal write permissions on an index, but none
that would imply it could execute the global checkpoint sync. This then
serves as a test that the upstream change to fire the global checkpoint
sync as the system user is correct. This test must run as a mulit-node
test so that a replica is a assigned so that the global checkpoint sync
fires in the first place. This test does indeed fail without the
upstream change, and passes with it.
Relates elastic/x-pack-elasticsearch#2744
Original commit: elastic/x-pack-elasticsearch@bf7e771756
* Improve JDBC communication
Jdbc HTTP client uses only one url for messages and relies on / for ping
Fixed ES prefix being discarded (missing /)
Add HEAD handler for JDBC endpoint
Original commit: elastic/x-pack-elasticsearch@389f82262e
We put the CLI in unix mode so if we send it
`\r\n` (the default in windows) then it'll
spit out extra "you are on a line continuation"
characters (`|`). Instead, we can use `\n`
directly and everything works.
I've also added a timeout to the reads from the
CLI because it makes the tests easier to debug.
Original commit: elastic/x-pack-elasticsearch@69f69f4092
This commit fixes indentation in certgen.bash, adds a check on cluster
health in bootstrap_password.bash and fixes a bug in xpack.bash
Original commit: elastic/x-pack-elasticsearch@d6847f6640
SQL was cleaning up the audit logs *after* each test
but this switches it to cleaning up the audit logs
*before* each test. This is faster because we can
generate a lot of audit logs before the first test.
Original commit: elastic/x-pack-elasticsearch@71d8f76667
Adds the GET overall_buckets API.
The REST end point is: GET
/_xpack/ml/anomaly_detectors/job_id/results/overall_buckets
The API returns overall bucket results. An overall bucket
is a summarized bucket result over multiple jobs.
It has the `bucket_span` of the longest job's `bucket_span`.
It also has an `overall_score` that is the `top_n` average of the
max anomaly scores per job.
relates elastic/x-pack-elasticsearch#2693
Original commit: elastic/x-pack-elasticsearch@ba6061482d
Do not execute bind on on the LDAP reader thread
Each LDAP connection has a single associated thread, executing the handlers for async requests; this is managed by the LDAP library. The bind operation is blocking for the connection. It is a deadlock to call bind, if on the LDAP reader thread for the same connection, because waiting for the bind response blocks the thread processing responses (for this connection).
This will execute the bind operation (and the subsequent runnable) on a thread pool after checking for the conflict above.
Closes: elastic/x-pack-elasticsearch#2570, elastic/x-pack-elasticsearch#2620
Original commit: elastic/x-pack-elasticsearch@404a3d8737
Since elastic/elasticsearch#26878, array and list of settings are
internally represented as actual lists. This makes filtering works
as expected when it comes to filter out arrays/lists.
The packaging tests used to check the presence of the XPack SSL
certificated_authorities setting which should have always been filtered.
By fixing the filtering of settings, elastic/elasticsearch#26878 broke
this packaging test.
This commit changes this test so that it does not expect certificated_authorities
setting to exist in the Nodes Info response.
relates elastic/x-pack-elasticsearch#2688
Original commit: elastic/x-pack-elasticsearch@cb299186b8
Now that we have fetch size working consistently we should randomize
the fetch size that we use in the tests to detect any errors caused
by strange fetch sizes.
Original commit: elastic/x-pack-elasticsearch@2c41fb5309
The true purpose of this test is to introduce another test alongside
the original, so that the test suite passes even if the other test
is skipped due to the assumption it makes about `build.snapshot`.
Original commit: elastic/x-pack-elasticsearch@709d7a5dc5
* Switch `ResultSet#getFetchSize` from returning the *requested*
fetch size to returning the size of the current page of results.
For scrolling searches without parent/child this'll match the
requested fetch size but for other things it won't. The nice thing
about this is that it lets us tell users a thing to look at if
they are wondering why they are using a bunch of memory.
* Remove all the entire JDBC action and implement it on the REST
layer instead.
* Create some code that can be shared between the cli and jdbc
actions to properly handle errors and request deserialization so
there isn't as much copy and paste between the two. This helps
because it calls out explicitly the places where they are different.
* I have not moved the CLI REST handle to shared code because
I think it'd be clearer to make that change in a followup PR.
* There is now no more need for constructs that allow iterating
all of the results in the same request. I have not removed these
because I feel that that cleanup is big enough to deserve its own
PR.
Original commit: elastic/x-pack-elasticsearch@3b12afd11c
Column type autodetect of integer types is broken in JDBC CSV library when it is used in tr-TR locale. The library is using toLowerCase() calls with default locale, which causes it to convert autodetected type name "Int" to lowercase "ınt" in tr-TR locale and not recognize it as an int type afterwards.
This commit adds a temporary workaround that makes the prevents that test from failing by specifying explicit column types in all tests where integer columns are present.
Original commit: elastic/x-pack-elasticsearch@86ca2acd8c
Release tests were introduced that sets the `build.snapshot`
system property to `false` to mimic release builds. This invalidates
the hardcoded license signatures that were signed against the
integration test pub/priv keys. This commit modifies the
license-validation assertions to assume the test is running against
those test keys, and will be skipped/ignored when these assertions
fail (which should only occur with `build.snapshot=true`)
Original commit: elastic/x-pack-elasticsearch@871704a3af
Builds on elastic/x-pack-elasticsearch#2403 to move all of sql's integration testing into
qa modules with different running server configurations. The
big advantage of this is that it allows us to test the cli and
jdbc with security present.
Creating a project that depends on both cli and jdbc and the
server has some prickly jar hell issues because cli and jdbc
package their dependencies in the jar. This works around it
in a few days:
1. Include only a single copy of the JDBC dependencies with
careful gradle work.
2. Do not include the CLI on the classpath at all and instead
run it externally.
I say "run it externally" rather than "fork it" because Elasticsearch
tests aren't allowed to fork other processes. This is forbidden
by seccomp on linux and seatbelt on osx and cannot be explicitly
requested like additional security manager settings. So instead
of forking the CLI process directly the tests interact with a test
fixture that isn't bound by Elasticsearch's rules and *can* fork
it.
This forking of the CLI has a nice side effect: it forces us to
make sure that things like security and connection strings other
than `localhost:9200` work. The old test could and did work around
missing features like that. The new tests cannot so I added the
ability to set the connection string. Configuring usernames and
passwords was also not supported but I did not add support for
that, only created the failing test and marked it as `@AwaitsFix`.
Original commit: elastic/x-pack-elasticsearch@560c6815e3
This commit adds back the ability to disable TLS on the transport layer and also disables TLS by
default to restore the 5.x behavior. The auto generation of key/cert and bundled CA certificate
have also been removed.
Relates elastic/x-pack-elasticsearch#2463
Original commit: elastic/x-pack-elasticsearch@abc66ec67d
Instead of throwing and catching an exception for invalid
indices this returns *why* they are invalid in a convenient
object form that can be thrown as an exception when the index
is required or the entire index can be ignored when listing
indices.
Original commit: elastic/x-pack-elasticsearch@f45cbce647
This integrates SQL's metadata calls with security by creating
`SqlIndicesAction` and routing all of SQL's metadata calls through
it. Since it *does* know up from which indices it is working against
it can be an `IndicesRequest.Replaceable` and integrate with the
existing security infrastructure for filtering indices.
This request is implemented fairly similarly to the `GetIndexAction`
with the option to read from the master or from a local copy of
cluster state. Currently SQL forces it to run on the local copy
because the request doesn't properly support serialization. I'd
like to implement that in a followup.
Original commit: elastic/x-pack-elasticsearch@15f9512820
On Windows a named pipe server must call ConnectNamedPipe() before using
a named pipe. However, if the client has already connected then this
function returns a failure code, with detailed error code
ERROR_PIPE_CONNECTED. The server must check for this, as it means the
connection will work fine. The Java test that emulates what the C++
would do in production did not have this logic.
This was purely a test problem. The C++ code used in production already
does the right thing.
relates elastic/x-pack-elasticsearch#2359
Original commit: elastic/x-pack-elasticsearch@e162887f28
The changes made for elastic/x-pack-elasticsearch#2369 showed that the ML security tests were seriously
weakened by the decision to grant many "minimal" privileges to all users
involved in the tests. A better solution is to override the auth header
such that a superuser runs setup actions and assertions that work by
querying raw documents in ways that an end user wouldn't. Then the ML
endpoints can be called with the privileges provided by the ML roles and
nothing else.
Original commit: elastic/x-pack-elasticsearch@4de42d9e54
Implementation details of ML endpoints should be performed using the
internal client, so that the end user only requires permissions for
the public ML endpoints and does not need to know how they are
implemented. This change fixes some instances where this rule was
not adhered to.
Original commit: elastic/x-pack-elasticsearch@01c8f5172c
* Add support for authz checks at on shard requests
* Add Rest Tests for authorization
* Bulk security - Only reject individual items, rather than a whole shard
* Sync with core change
* Grant "delete" priv in ML smoketest
This role had index and+bulk privileges but it also needs delete (in order to delete ML model-snapshots)
Original commit: elastic/x-pack-elasticsearch@830e89e652
This shuffles all of SQL's QA tests into the `qa/sql` directory, moving
some shared resources into the new `qa:sql` project. It also rigs up
testing of the rest SQL interface in all the sql qa configurations:
without security, with security, and against multiple nodes.
I've had to make some modifications to how we handle the audit log
because it has gotten pretty slow. If these modifications turn out to
not be fast enough then I'll change the test to querying the log files
and drop the audit log index entirely but the index seems to be holding
out for now.
Original commit: elastic/x-pack-elasticsearch@ff3b5a74c1
This commit removes some extraneous trailing newlines from
keystore.bash, the packaging test cases for the interaction between
installing X-Pack and the keystore.
Original commit: elastic/x-pack-elasticsearch@86250ecfbc
Scrolling was only implemented for the `SqlAction` (not jdbc or cli)
and it was implemented by keeping request state on the server. On
principle we try to avoid adding extra state to elasticsearch where
possible because it creates extra points of failure and tends to
have lots of hidden complexity.
This replaces the state on the server with serializing state to the
client. This looks to the user like a "next_page" key with fairly
opaque content. It actually consists of an identifier for the *kind*
of scroll, the scroll id, and a base64 string containing the field
extractors.
Right now this only implements scrolling for `SqlAction`. The plan
is to reuse the same implementation for jdbc and cli in a followup.
This also doesn't implement all of the required serialization.
Specifically it doesn't implement serialization of
`ProcessingHitExtractor` because I haven't implemented serialization
for *any* `ColumnProcessors`.
Original commit: elastic/x-pack-elasticsearch@a8567bc5ec
`authc.token.enabled` is true unless `http.ssl.enabled` is `false` and `http.enabled` is `true`.
* TokenService default enabled if HTTP_ENABLED == false
* Fixed tests that need TokenService explicitly enabled
* [DOC] Default value for `xpack.security.authc.token.enabled`
Original commit: elastic/x-pack-elasticsearch@bd154d16eb
As there are two indices to upgrade for watcher, it makes a lot of sense
to also have two upgrade checks.
There is one upgrader for the watches index, which deletes
old templates, adds the new one before and then does the reindexing.
Same for the triggered watches index.
This also means, that there will be two entries popping up in the kibana
UI.
Note: Each upgrade check checks if the other index (for the .watches
upgrade check the triggered watches index and vice versa) is already
upgraded and only if that is true, watcher is restarted.
relates elastic/x-pack-elasticsearch#2238
Original commit: elastic/x-pack-elasticsearch@2c92040ed6
Today we require a pre-shared key to use the token service. Beside the
additional setup step it doesn't allow for key-rotation which is a major downside.
This change adds a TokenService private ClusterState.Custom that is used to distribute
the keys used to encrypt tokens. It also has the infrastructur to add automatic key
rotation which is not in use yet but included here to illustrate how it can work down
the road.
This is considered a prototype and requires additioanl integration testing. Yet, it's fully
BWC with a rolling / full cluster restart from a previous version (also from 5.6 to 6.x)
since if the password is set it will just use it instead of generating a new one.
Once we implement the automatic key rotation via the clusterstate we need to ensure that we are
fully upgraded before we do that.
Also note that the ClusterState.Custom is fully transient and will never be serialized to disk.
Original commit: elastic/x-pack-elasticsearch@1ae22f5d41
Today we require the `bootstrap.password` to be present in the keystore in order to
bootstrap xpack. With the addition of `keystore.seed` we have a randomly generated password
per node to do the bootstrapping. This will improve the initial user experience significantly
since the user doesn't need to create a keystore and add a password, they keystore is created
automatically unless already present and is always created with this random seed.
Relates to elastic/elasticsearch#26253
Original commit: elastic/x-pack-elasticsearch@5a984b4fd8
Adapts audit logging to actions that delay getting index access control until the action is started. The audit log will contain an entry for the action itself starting without any associated indices because the indices are not yet known. The audit log will also contain an entry for every time the action resolved security for a set of indices. Since sql resolves indices one at a time it will contain an entry per index.
All of this customization is entirely in the security code. The only SQL change in this PR is to add audit logging support to the integration test.
Original commit: elastic/x-pack-elasticsearch@539bb3c2a8
The rest test waited for the watch to run in the background, but there
were no guarantees that this really happened. Also it waited for five
seconds, instead of just executing the watch manually.
relates elastic/x-pack-elasticsearch#2255
Original commit: elastic/x-pack-elasticsearch@56765a649e
These tests have repeating but not reproducible failures,
where the stash is filled with a second PUT operation and the
watcher stats response does not match. Setting the log to trace
should shed some light on this.
As the smoke tests are only four tests this will not lead to a
log explosion.
Relates elastic/x-pack-elasticsearch#1513, elastic/x-pack-elasticsearch#1874
Original commit: elastic/x-pack-elasticsearch@5832dc7990
Adds a test that shows *how* SQL fails to address an index with two types
to the full cluster restart tests. Because we're writing this code
against 7.0 don't actually execute the test, but we will execute it when
we merge to 6.x and it *should* work.
Original commit: elastic/x-pack-elasticsearch@b536e9a142
This adds support for field level security to SQL by creating a new type of flow for securing requests that look like sql requests. `AuthorizationService` verifies that the user can execute the request but doesn't check the indices in the request because they are not yet ready. Instead, it adds a `BiFunction` to the context that can be used to check permissions for an index while servicing the request. This allows requests to cooperatively secure themselves. SQL does this by implementing filtering on top of its `Catalog` abstraction and backing that filtering with security's filters. This minimizes the touch points between security and SQL.
Stuff I'd like to do in followups:
What doesn't work at all still:
1. `SHOW TABLES` is still totally unsecured
2. `DESCRIBE TABLE` is still totally unsecured
3. JDBC's metadata APIs are still totally unsecured
What kind of works but not well:
1. The audit trail doesn't show the index being authorized for SQL.
Original commit: elastic/x-pack-elasticsearch@86f88ba2f5
This commit re-enables the OpenLDAP tests that were previously running against a one-off instance
in AWS but now run against a vagrant fixture. There were some IntegTests that would run against the
OpenLDAP instance randomly but with this change they no longer run against OpenLDAP. This is ok as
the functionality that is tested by these has coverage elsewhere.
relates elastic/x-pack-elasticsearch#1823
Original commit: elastic/x-pack-elasticsearch@ac9bc82297
SQL relies on being able to fetch information about fields from
the cluster state and it'd be disasterous if that information
wasn't available. This should catch that.
Original commit: elastic/x-pack-elasticsearch@1a62747332
Running the sql rest action test inside the server caused a dependency
loop which was failing the build.
Original commit: elastic/x-pack-elasticsearch@43283671d8
Replacing integration tests with rest tests and unit tests, thus removing integration tests that require start of a new cluster. Removing unused testing methods
Original commit: elastic/x-pack-elasticsearch@265966d80c
Some standalone tests set the max script compilation limit. However,
this setting is now set in the main cluster setup in core so it is no
longer needed here. This commit removes these obviated settings in
standalone tests.
Original commit: elastic/x-pack-elasticsearch@089328c8d7
The full cluster restart tests are currently geared towards the 5.6 -> 6.0 upgrade and have some
issues when the versions are changed to 6.x -> 7.0. One issue is a real code issue in that the
security code always expects the mappings to have the same version as the version of the node, but
we no longer update the mappings on the security index during a rolling upgrade. We know look at
the index format to determine if the index is up to date.
Original commit: elastic/x-pack-elasticsearch@14c1c72ff6
The command line flag --path.conf was removed yet this test was still
using it. This commit replaces the usage of this flag in this test with
the new mechanism.
Original commit: elastic/x-pack-elasticsearch@c37050894c
Add some basic security testing/integration.
The good news:
1. Basic security now works. Users without access to an index can't run sql queries against it. Without this change they could.
2. Document level security works! At least so far as I can tell.
The work left to do:
1. Field level security doesn't work properly. I mean, it kind of works in that the field's values don't leak but it just looks like they all have null values.
2. We will need to test scrolling.
3. I've only added tests for the rest sql action. I'll need to add tests for jdbc and the CLI as well.
4. I've only added tests for `SELECT` and have ignored stuff like `DESCRIBE` and `SHOW TABLES`.
Original commit: elastic/x-pack-elasticsearch@b9909bbda0
This commit improves the error message in 6x if the security index has
not been upgraded, and warns the user that the native realm will not be
functional until the upgrade API is run.
Original commit: elastic/x-pack-elasticsearch@710b7634b4
the upgrade API is lacking some functionality in a special case,
where triggered_watches exists, but .watches does not. This
deletes the triggered watches index manually until we integrated
this properly in the upgrade API to fix the tests
Original commit: elastic/x-pack-elasticsearch@e9d1b0d35d
This commit adds the upgrade API functionality and script for security.
It also enables previously muted tests that would fail due to the lack
of security upgrade features in testing cluster restarts and old
security index backward compatibility.
Original commit: elastic/x-pack-elasticsearch@4abe9f1263
The current tests were only able to finish successfully, if the earlier
tests were run as well, you could not run the restart test in isolation.
This commit ensures an upgrade is executed if needed.
Original commit: elastic/x-pack-elasticsearch@616ebbd6eb
Otherwise we might run into race conditions that prevent a useful
start up again.
Those tests can be massively improved (no need to run against the real
master node), once the watcher BWC compatible stats are in.
relates elastic/x-pack-elasticsearch#2004
Original commit: elastic/x-pack-elasticsearch@52ca77809c
This is related to elastic/x-pack-elasticsearch#1217. This commit requires that the elastic password
be bootstrapped for the user to be authenticated. As a result it removes
the special "setup" mode that allowed the user to be authenticated from
localhost.
Additionally, this commit updates the tests to work with this
functionality.
Original commit: elastic/x-pack-elasticsearch@d0d5d697a7
When a user asks for upgrade information for all indices and there are no indices in the cluster, upgrade assistance should just return an empty response indicating that no indices require upgrade or reindexing. This commit also reverts the temporary fix in WatchBackwardsCompatibilityIT tests that was added as a workaround for this issue.
Original commit: elastic/x-pack-elasticsearch@2ea9707867
This allows for messages to be returned, and distinguishes between 4 different results:
- I have authenticated the user
- I don't know how to authenticate that user. Try another realm.
- I tried to authenticate the user, but failed. Try another realm.
- I tried to authenticate the user, but failed. Fail the authentication attempt.
Original commit: elastic/x-pack-elasticsearch@f796949cfb
When there are data or master nodes in the cluster, that are older
than ES 6.0 alpha3, then watcher will only start on the master node.
Changed all transport actions to be master node actions, as there is
already a method to decide to run locally, which we can piggyback on.
Original commit: elastic/x-pack-elasticsearch@65cecb6d69
Jay Modi
Costin Leau
Nik Everett
Costin Leau
Dimitris Athanasiou
Simon Willnauer
Jason Tedor
Tanguy Leroux
Igor Motov
David Roberts
Albert Zaharovits
Yannick Welsch
Boaz Leskes
Nik Everett
Tal Levy
Tim Brooks
David Kyle
Tim Vernum
Michael Basnight
Alexander Reelsen
Ryan Ernst
Ali Beyad
jaymode