Migrate runtime pytest.skip() calls across 34 files to use the typed
skip_env() wrapper from test.pylib.skip_types.
These sites skip at runtime because a required feature, config option,
library version, build mode, or runtime topology is not available.
Also fixes 'raise pytest.skip(...)' in test_audit.py — skip_env()
already raises internally, so the explicit raise was incorrect.
Each file gains one new import:
from test.pylib.skip_types import skip_env
If the keyspace is migrating, it reports the intended and actual storage
mode for each node.
Signed-off-by: Nikos Dragazis <nikolaos.dragazis@scylladb.com>
The endpoint is the following:
POST /storage_service/vnode_tablet_migrations/keyspaces/{keyspace}/finalization
When called, it issues a `finalize_migration` topology request and waits
for its completion.
Signed-off-by: Nikos Dragazis <nikolaos.dragazis@scylladb.com>
The endpoint is the following:
POST /storage_service/vnode_tablet_migrations/node/storage_mode?intended_mode={tablets,vnodes}
This endpoint is part of the vnodes-to-tablets migration process and
controls a node's intended_storage_mode in system.topology. The storage
mode represents the node-local data distribution model, i.e., how data
are organized across shards. The node will apply the intended storage
mode to migrating tables upon next restart by resharding their SSTables
(either on vnode boundaries if intended_mode=tablets, or with the static
sharder if intended_mode=vnodes).
Note that this endpoint controls the intended_storage_mode of the local
node only. This has the nice benefit that once the API call returns, the
change has not only been committed to group0 but also applied to the
local node's state machine. This guarantees that the change is part of
the node's local copy upon next restart; no additional read barrier is
needed.
Signed-off-by: Nikos Dragazis <nikolaos.dragazis@scylladb.com>
The endpoint is the following:
POST /storage_service/vnode_tablet_migrations/keyspaces/{keyspace}
Its purpose is to start the migration of a whole keyspace from vnodes to
tablets.
When called, Scylla will synchronously create a tablet map for each
table in the specified keyspace. The tablet maps of all tables are
identical and they mirror the vnode layout; they contain one tablet per
vnode and each tablet uses the same replica hosts and token boundaries
as the corresponding vnode.
The only difference from vnodes lies in the sharding approach. Tablets
are assigned to a single shard - using a round-robin strategy in this
patch - whereas vnodes are distributed evenly across all shards. If the
tablet count per shard is low and tablet sizes are uneven, or some
shards have more tablets than others, performance may degrade during the
migration process. For example, a cluster with i8g.48xlarge (192 vCPUs),
256 vnodes per node and RF=3 will have 256 * 3 / 192 vCPUs = 4 tablet
replicas per shard during the migration. One additional tablet or a
double-sized tablet would cause 25% overcommit.
Signed-off-by: Nikos Dragazis <nikolaos.dragazis@scylladb.com>
Add a segment_set member to replica::compaction_group that manages the
logstor segments that belong to the compaction group, similarly to how
it manages sstables. Add also a separator buffer in each compaction
group.
When writing a mutation to a compaction group, the mutation is written
to the active segment and to the separator buffer of the compaction
group, and when the separator buffer is flushed the segment is added to
the compaction_group's segment set.
add barrier operation that forces switch of the active segment and
separator, and waits for all existing segments to close and all
separators to flush.
With migration to the pytest, file descriptors will be hanged during the whole life of the process. Previously it was not an issue, because test.py was executing only one file with Popen, so descriptors will be freed with process done. With new approach they are blocked. This will allow to eliminate this.
Fix issue when we had issue with getting cluster and then trying to set it dirty while it None.
Put cluster to the pool only if it was created
It is observed that:
repair - repair[667d4a59-63fb-4ca6-8feb-98da49946d8b]: Failed to update
system.repair_history table of node d27de212-6f32-4649ad76-a9ef1165fdcb:
seastar::rpc::remote_verb_error (repair[667d4a59-63fb-4ca6-8feb-98da49946d8b]: range (minimum
token,maximum token) is not in the format of (start, end])
This is because repair checks the end of the range to be repaired needs
to be inclusive. When small_table_optimization is enabled for regular
repair, a (minimum token,maximum token) will be used.
To fix, we can relax the check of (start, end] for the min max range.
Fixes#27220Closesscylladb/scylladb#27357
This series allows an operator to reset 'cleanup needed' flag if he already cleaned up the node, so that automatic cleanup will not do it again. We also change 'nodetool cleanup' back to run cleanup on one node only (and reset 'cleanup needed' flag in the end), but the new '--global' option allows to run cleanup on all nodes that needed it simultaneously.
Fixes https://github.com/scylladb/scylladb/issues/26866
Backport to all supported version since automatic cleanup behaviour as it is now may create unexpected by the operator load during cluster resizing.
Closesscylladb/scylladb#26868
* https://github.com/scylladb/scylladb:
cleanup: introduce "nodetool cluster cleanup" command to run cleanup on all dirty nodes in the cluster
cleanup: Add RESTful API to allow reset cleanup needed flag
97ab3f6622 changed "nodetool cleanup" (without arguments) to run
cleanup on all dirty nodes in the cluster. This was somewhat unexpected,
so this patch changes it back to run cleanup on the target node only (and
reset "cleanup needed" flag afterwards) and it adds "nodetool cluster
cleanup" command that runs the cleanup on all dirty nodes in the
cluster.
This commits adds a tests checking various scenarios of restoring
via load and stream with primary_replica_only and a scope specified.
The tests check that in a few topologies, a mutation is replicated
a correct amount of times given primary_replica_only and that
streaming happens according to the scope rule passed.
Signed-off-by: Robert Bindar <robert.bindar@scylladb.com>
Recently (#26231) there was added a test to check that several API
endpoints, that return tokens and corresponding replica nodes, are
consistent with tablet map. This patch adds one more API endpoint to the
validation -- the /storage_service/tokens_endpoint one.
The extention is pretty straightforward, but the new endpoint returns
back a single (primary) replica for a token, so the test check is
slightly modified to account for that.
Signed-off-by: Pavel Emelyanov <xemul@scylladb.com>
Closesscylladb/scylladb#26580
Problems addressed by this PR
* Missing barrier before cleanup: If a node was bootstrapped before cleanup, some request coordinators could still be in `write_both_read_new` and send stale requests to replicas being cleaned up.
* Sessions not drained before cleanup: We lacked protection against stale streaming or repair operations.
* `sstable_vnodes_cleanup_fiber()` calling `flush_all_tables()` under group0 lock: This caused SCT test failures (see [this comment](https://github.com/scylladb/scylladb/issues/25333#issuecomment-3298859046) for details).
* Issues with `storage_proxy::start_write()` used by `sstable_vnodes_cleanup_fiber`:
* The result of `start_write()` was not held during `abstract_write_response_handler::apply_locally`, so coordinator-local writes were not properly awaited.
* Synchronization was racy — `start_write()` was not atomic with the fence check, allowing stale writes to sneak in if `fence_version` changed in between.
* It waited for all writes, including local tables and tablet-based tables, which is redundant because `sstable_vnodes_cleanup_fiber` does not apply to them.
* It also waited for writes with versions greater than the current `fence_version`, which is unnecessary.
Fixesscylladb/scylladb#26150
backport: this PR fixes several issues with the vnodes cleanup procedure, but it doesn't seem they are critical enough to deserve backporting
Closesscylladb/scylladb#26315
* https://github.com/scylladb/scylladb:
test_automatic_cleanup: add test_cleanup_waits_for_stale_writes
test_fencing: fix due to new version increment
test_automatic_cleanup: clean it up
storage_proxy: wait for closing sessions in sstable cleanup fiber
storage_proxy: rename await_pending_writes -> await_stale_pending_writes
storage_proxy: use run_fenceable_write
storage_proxy: abstract_write_response_handler: apply_locally: extract post fence check
storage_proxy: introduce run_fenceable_write
storage_proxy: move update_fence_version from shared_token_metadata
storage_proxy: fix start_write() operation scope in apply_locally
storage_proxy: move post fence check into handle_write
storage_proxy: move fencing into mutate_counter_on_leader_and_replicate
storage_proxy::handle_read: add fence check before get_schema
storage_service: rebrand cleanup_fiber to vnodes_cleanup_fiber
sstable_cleanup_fiber: use coroutine::parallel_for_each
storage_service: sstable_cleanup_fiber: move flush_all_tables out of the group0 lock
topology_coordinator: barrier before cleanup
topology_coordinator: small start_cleanup refactoring
global_token_metadata_barrier: add fenced flag
Remove redundant imports and variables. Extract cleanup_all
function. Add logs. Remove pytest.mark.prepare_3_racks_cluster --
the test doesn't actually need a 3 node cluster, one initial
node is enough.
The universalasync.wrap function doesn't preserve the
type information, which confuses the VS Code Pylance
plugin and makes code navigation hard.
In this commit we fix the problem by adding a typed
wrapped around universalasync.wrap.
Fixes: scylladb/scylladb#26639
This PR contains various improvements in the recovery procedure
tests, mostly `test_raft_recovery_user_data`:
- decreasing the running time,
- some simplifications,
- making sure group 0 majority is lost when expected.
These are not critical test changes, so no need to backport.
Closesscylladb/scylladb#26442
* github.com:scylladb/scylladb:
test: assert that majority is lost in some tests of the recovery procedure
test: rest_client: add timeout support for read_barrier
test: test_raft_recovery_user_data: lose majority when killing one dc
test: test_raft_recovery_user_data: shutdown driver sessions
test: test_raft_recovery_user_data: use a separate driver connection for the write workload
test: test_raft_recovery_user_data: send ALTER KEYSPACE to any node
test: test_raft_recovery_user_data: bring failure_detector_timeout_in_ms back to 20 s
test: test_raft_recovery_user_data: speed up replace operations
test: stop/start servers concurrently in the recovery procedure tests
test_two_tablets_concurrent_repair_and_migration_repair_writer_level waits
for the first node that logs info about repair_writer using asyncio.wait.
The done group is never awaited, so we never learn about the error.
The test itself is incorrect and the log about repair_writer is never
printed. We never learn about that and tests finishes successfully
after 10 minutes timeout.
Fix the test:
- disable hinted handoff;
- repair tablets of the whole table:
- new table is added so that concurrent migration is possible;
- use wait_for_first_completed that awaits done group;
- do some cleanups.
Remove nightly mark.
Fixes: #26148.
Closesscylladb/scylladb#26209
Sometimes it's convenient to use slashes in injection names,
for example my_component/my_method/my_condition. Without quote()
we get 'handler not found' error from Scylla.
This patch introduces a new `incremental_mode` parameter to the tablet
repair REST API, providing more fine-grained control over the
incremental repair process.
Previously, incremental repair was on and could not be turned off. This
change allows users to select from three distinct modes:
- `regular`: This is the default mode. It performs a standard
incremental repair, processing only unrepaired sstables and skipping
those that are already repaired. The repair state (`repaired_at`,
`sstables_repaired_at`) is updated.
- `full`: This mode forces the repair to process all sstables, including
those that have been previously repaired. This is useful when a full
data validation is needed without disabling the incremental repair
feature. The repair state is updated.
- `disabled`: This mode completely disables the incremental repair logic
for the current repair operation. It behaves like a classic
(pre-incremental) repair, and it does not update any incremental
repair state (`repaired_at` in sstables or `sstables_repaired_at` in
the system.tablets table).
The implementation includes:
- Adding the `incremental_mode` parameter to the
`/storage_service/repair/tablet` API endpoint.
- Updating the internal repair logic to handle the different modes.
- Adding a new test case to verify the behavior of each mode.
- Updating the API documentation and developer documentation.
Fixes#25605Closesscylladb/scylladb#25693
The central idea of incremental repair is to allow repair participants
to select and repair only a portion of the dataset to speed up the
repair process. All repair participants must utilize an identical
selection method to repair and synchronize the same selected dataset.
There are two primary selection methods: time-based and file-based. The
time-based method selects data within a specified time frame. It is
versatile but it is less efficient because it requires reading all of
the dataset and omitting data beyond the time frame. The file-based
method selects data from unrepaired SSTables and is more efficient
because it allows the entire SSTable to be omitted. This document patch
implements the file-based selection method.
Incremental repair will only be supported for tablet tables; it will not
be supported for vnode tables. On one hand, the legacy vnode is less
important to support. On the other hand, the incremental repair for
vnode is much harder to implement. With vnodes, a SSTalbe could contain
data for multiple vnode ranges. When a given vnode range is repaired,
only a portion of the SSTable is repaired. This complicates the
manipulation of SSTables significantly during both repair and
compaction. With tablets, an entire tablet is repaired so that a
sstable is either fully repaired or not repaired which is a huge
simplification.
This patch uses the repaired_at from sstables::statistics component to
mark a sstable as repaired. It uses a virtual clock as the repair
timestamp, i.e., using a monotonically increasing number for the
repaired_at field of a SSTable and sstables_repaired_at column in
system.tablets table. Notice that when a sstable is not repaired, the
repaired_at field will be set to the default value 0 by default. The
being_repaired in memory field of a SSTable is used to explicitly mark
that a SSTable is being selected. The following variables are used for
incremental repair:
The repaired_at on disk field of a SSTable is used.
- A 64-bit number increases sequentially
The sstables_repaired_at is added to the system.tablets table.
- repaired_at <= sstables_repaired_at means the sstable is repaired
The being_repaired in memory field of a SSTable is added.
- A repair UUID tells which sstable has participated in the repair
Initial test results:
1) Medium dataset results
Node amount: 3
Instance type: i4i.2xlarge
Disk usage per node: ~500GB
Cluster pre-populated with ~500GB of data before starting repairs job.
Results for Repair Timings:
The regular repair run took 210 mins.
Incremental repair 1st run took 183 mins, 2nd and 3rd runs took around 48s
The speedup is: 183 mins / 48s = 228X
2) Small dataset results
Node amount: 3
Instance type: i4i.2xlarge
Disk usage per node: ~167GB
Cluster pre-populated with ~167GB of data before starting the repairs job.
Regular repair 1st run took 110s, 2nd and 3rd runs took 110s.
Incremental repair 1st run took 110 seconds, 2nd and 3rd run took 1.5 seconds.
The speedup is: 110s / 1.5s = 73X
3) Large dataset results
Node amount: 6
Instance type: i4i.2xlarge, 3 racks
50% of base load, 50% read/write
Dataset == Sum of data on each node
Dataset Non-incremental repair (minutes)
1.3 TiB 31:07
3.5 TiB 25:10
5.0 TiB 19:03
6.3 TiB 31:42
Dataset Incremental repair (minutes)
1.3 TiB 24:32
3.0 TiB 13:06
4.0 TiB 5:23
4.8 TiB 7:14
5.6 TiB 3:58
6.3 TiB 7:33
7.0 TiB 6:55
Fixes#22472Closesscylladb/scylladb#24291
* github.com:scylladb/scylladb:
replica: Introduce get_compaction_reenablers_and_lock_holders_for_repair
compaction: Move compaction_reenabler to compaction_reenabler.hh
topology_coordinator: Make rpc::remote_verb_error to warning level
repair: Add metrics for sstable bytes read and skipped from sstables
test.py: Disable incremental for test_tombstone_gc_for_streaming_and_repair
test.py: Add tests for tablet incremental repair
repair: Add tablet incremental repair support
compaction: Add tablet incremental repair support
feature_service: Add TABLET_INCREMENTAL_REPAIR feature
tablet_allocator: Add tablet_force_tablet_count_increase and decrease
repair: Add incremental helpers
sstable: Add being_repaired to sstable
sstables: Add set_repaired_at to metadata_collector
mutation_compactor: Introduce add operator to compaction_stats
tablet: Add sstables_repaired_at to system.tablets table
test: Fix drain api in task_manager_client.py
The following tests are added for tablet incremental repair:
- Basic incremental repair
- Basic incremental repair with error
- Minor compaction and incremental repair
- Major compaction and incremental repair
- Scrub compaction and incremental repair
- Cleanup/Upgrade compaction and incremental repair
- Tablet split and incremental repair
- Tablet merge and incremental repair
In the ScyllaMetrics `get` function, when requesting the value for a
specific shard, it is expected to return the sum of all values of
metrics for that shard that match the labels.
However, it would return the value of the first matching line it finds
instead of summing all matching lines.
For example, if we have two lines for one shard like:
some_metric{scheduling_group_name="compaction",shard="0"} 1
some_metric{scheduling_group_name="sl:default",shard="0"} 2
The result of this call would be 1 instead of 3:
get('some_metric', shard="0")
We fix this to sum all matching lines.
The filtering of lines by labels is fixed to allow specifying only some
of the labels. Previously, for the line to match the filter, either the
filter needs to be empty, or all the labels in the metric line had to be
specified in the filter parameter and match its value, which is
unexpected, and breaks when more labels are added.
We also simplify the function signature and the implementation - instead
of having the shard as a separate parameter, it can be specified as a
label, like any other label.
This change adds the --scope option to nodetool refresh.
Like in the case of nodetool restore, you can pass either of:
* node - On the local node.
* rack - On the local rack.
* dc - In the datacenter (DC) where the local node lives.
* all (default) - Everywhere across the cluster.
as scope.
The feature is based on the existing load_and_stream paths, so it
requires passing --load-and-stream to the refresh command.
Also, it is not compatible with the --primary-replica-only option.
Signed-off-by: Robert Bindar <robert.bindar@scylladb.com>
Closesscylladb/scylladb#23861
Use universalasync library to make test.py async code compatible
with synchronous code of dtest/ccm
Also, copied unmodified error_example_test.py from dtest as an example.
Run the test in `dev` mode only.
Add an API call which estimates the effectiveness of possible
compression config changes.
This can be used to make an informed decision about whether to
change the compression method, without actually recompressing
any SSTables.
Add an API call which will retrain the SSTable compression dictionary
for a given table.
Currently, it needs all nodes to be alive to succeed. We can relax this later.
Set true to wait for the repair to complete. Set false to skip waiting
for the repair to complete. When the option is not provided, it defaults
to false.
It is useful for management tool that wants the api to be async.
Fixes#22418Closesscylladb/scylladb#22436
The following is observed in pytest:
1) node1, stream master, tried to pull data from node3
2) node3, stream follower, found node1 restarted
3) node3 killed the rpc stream
4) node1 did not get the stream session failure message from node3. This
failure message was supposed to kill the stream plan on node1. That's the
reason node1 failed the stream session much later at "2024-08-19 21:07:45,539".
Note, node3 failed the stream on its side, so it should have sent the stream
session failure message.
```
$ cat node1.log |grep f890bea0-5e68-11ef-99ae-e5bca04385fc
INFO 2024-08-19 20:24:01,162 [shard 0:strm] stream_session - [Stream #f890bea0-5e68-11ef-99ae-e5bca04385fc] Executing streaming plan for Tablet migration-ks-index-0 with peers={127.0.34.3}, master
ERROR 2024-08-19 20:24:01,190 [shard 1:strm] stream_session - [Stream #f890bea0-5e68-11ef-99ae-e5bca04385fc] Failed to handle STREAM_MUTATION_FRAGMENTS (receive and distribute phase) for ks=ks, cf=cf, peer=127.0.34.3: seastar::nested_exception: seastar::rpc::stream_closed (rpc stream was closed by peer) (while cleaning up after seastar::rpc::stream_closed (rpc stream was closed by peer))
WARN 2024-08-19 21:07:45,539 [shard 0:main] stream_session - [Stream #f890bea0-5e68-11ef-99ae-e5bca04385fc] Streaming plan for Tablet migration-ks-index-0 failed, peers={127.0.34.3}, tx=0 KiB, 0.00 KiB/s, rx=484 KiB, 0.18 KiB/s
$ cat node3.log |grep f890bea0-5e68-11ef-99ae-e5bca04385fc
INFO 2024-08-19 20:24:01,163 [shard 0:strm] stream_session - [Stream #f890bea0-5e68-11ef-99ae-e5bca04385fc] Executing streaming plan for Tablet migration-ks-index-0 with peers=127.0.34.1, slave
INFO 2024-08-19 20:24:01,164 [shard 1:strm] stream_session - [Stream #f890bea0-5e68-11ef-99ae-e5bca04385fc] Start sending ks=ks, cf=cf, estimated_partitions=2560, with new rpc streaming
WARN 2024-08-19 20:24:01,187 [shard 0: gms] stream_session - [Stream #f890bea0-5e68-11ef-99ae-e5bca04385fc] Streaming plan for Tablet migration-ks-index-0 failed, peers={127.0.34.1}, tx=633 KiB, 26506.81 KiB/s, rx=0 KiB, 0.00 KiB/s
WARN 2024-08-19 20:24:01,188 [shard 0:strm] stream_session - [Stream #f890bea0-5e68-11ef-99ae-e5bca04385fc] stream_transfer_task: Fail to send to 127.0.34.1:0: seastar::rpc::stream_closed (rpc stream was closed by peer)
WARN 2024-08-19 20:24:01,189 [shard 0:strm] stream_session - [Stream #f890bea0-5e68-11ef-99ae-e5bca04385fc] Failed to send: seastar::rpc::stream_closed (rpc stream was closed by peer)
WARN 2024-08-19 20:24:01,189 [shard 0:strm] stream_session - [Stream #f890bea0-5e68-11ef-99ae-e5bca04385fc] Streaming error occurred, peer=127.0.34.1
```
To be safe in case the stream fail message is not received, node1 could fail
the stream plan as soon as the rpc stream is aborted in the
stream_mutation_fragments handler.
Fixes#20227Closesscylladb/scylladb#21960
Before this change, it was possible to change non-liveupdatable config
parameter without process restart. This erroneous behavior not only
contradicts the documentation but is potentially dangerous, as various
components theoretically might not be prepared for a change of
configuration parameter value without a restart. The issue came from
a fact that liveupdatability verification check was skipped for default
configuration parameters (those without its initial values
in configuration file during process start).
This change:
- Introduce _initialization_completed member in config_file
- Set _initialization_completed=true when config file is processed on
server start
- Verify config_file's initialization status during config update - if
config_file was initialized, prevent from further changes of
non-liveupdatable parameters
- Implement ScyllaRESTAPIClient::get_config() that obtains a current
value of given configuration parameter via /v2/config REST API
- Implement test to confirm that only liveupdatable parameters are
changed when SIGHUP is sent after configuration file change
Function set_initialization_completed() is called only once in main.cc,
and the effect is expected to be visible in all shards, as a side effect
of cfg->broadcast_to_all_shards() that is called shortly after. The same
technique was already used for enable_3_1_0_compatibility_mode() call.
Fixesscylladb/scylladb#5382
No backport - minor fix.
Closesscylladb/scylladb#22655
* github.com:scylladb/scylladb:
test: SIGHUP doesn't change non-liveupdatable configuration
test: implement ScyllaRESTAPIClient::get_config()
config: prevent SIGHUP from changing non-liveupdatable parameters
config: remove unused set_value_on_all_shards(const YAML::Node&)
