these unused includes were identifier by clang-include-cleaner. after
auditing these source files, all of the reports have been confirmed.
please note, because quite a few source files relied on
`utils/to_string.hh` to pull in the specialization of
`fmt::formatter<std::optional<T>>`, after removing
`#include <fmt/std.h>` from `utils/to_string.hh`, we have to
include `fmt/std.h` directly.
Signed-off-by: Kefu Chai <kefu.chai@scylladb.com>
Since we manage ip to id mapping directly in gossiper now we need to
load the mapping on boot. We already do it anyway, but only due to a bug
which checks raft topology mode config before it is set, so the code
thinks that it is in the gossiper mode and loads peers table into the
gossiper and token metadata. Fix the bug and load peers into the gossiper
only since token metadata is managed by raft.
The series also removes address map related test that no longer checks
anything and replace it with unit test.
It also adds the dc/rack check to "join node" rpc. The check is done
during shadow round now, but for it to work it requires dc/rack to be
propagated through the gossiper and we want to eventually drop it.
Ref: scylladb/scylladb#21777
* 'load-peers' of https://github.com/gleb-cloudius/scylla:
topology coordinator: reject replace request if topology does not match
gossiper: fix the logic of shadow_round parameter
storage_service: do not add endpoint to the gossiper during topology loading.
storage_service: load peers into gossiper on boot in raft topology mode
storage_service: set raft topology change mode before using it in join_cluster
locator: drop inet_address usage to figure out per dc/rack replication
test: drop test_old_ip_notification_repro.py
test: address_map: check generation handling during entry addition
We still have a number of issues to be solved for views with tablets.
Until they are fixed, we should prevent users from creating them,
and use the vnode-based views instead.
This patch prepares the feature for enabling views with tablets. The
feature is disabled by default, but currently it has no effect.
After all tests are adjusted to use the feature, we should depend
on the feature for deciding whether we can create materialized views
in tablet-enabled keyspaces.
The unit tests are adjusted to enable this feature explicitly, and it's
also added to the scylla sstable tool config - this tool treats all
tables as if they were tablet-based (surprisingly, with SimpleStrategy),
so for it to work on views, the new feature must be enabled.
Refs scylladb/scylladb#21832Closesscylladb/scylladb#21833
Information about the number of shares per service level will be stored
in an additional column in the service levels table, which is managed
through group0. We will need the feature to make sure that all nodes in
the cluster know about the new column before any node starts applying
group0 commands the would touch the new column.
This feature also serves a role for the legacy service levels
implementation that uses system_distributed for storage: after all nodes
are upgraded to support workload prioritization, one of the nodes will
perform a schema change operation and will add the new column.
The test no longer test anything since the address map is updated much
earlier now by the gossiper itself, not by the notifiers. The
functionality is tested by a unit test now.
We already ignore a gossiper entries with host id equal to local host id
in raft mode since those entries are just outdated entries since before
ip change. The same logic applies to gossiper mode as well though, so do
the same in both modes.
Fixes: scylladb/scylladb#21930
Message-ID: <Z20kBZvpJ1fP9WyJ@scylladb.com>
This is a forward port (from scylla-enterprise) of additional compression options (zstd, dictionaries shared across messages) for inter-node network traffic. It works as follows:
After the patch, messaging_service (Scylla's interface for all inter-node communication)
compresses its network traffic with compressors managed by
the new advanced_rpc_compression::tracker. Those compressors compress with lz4,
but can also be configured to use zstd as long as a CPU usage limit isn't crossed.
A precomputed compression dictionary can be fed to the tracker. Each connection
handled by the tracker will then start a negotiation with the other end to switch
to this dictionary, and when it succeeds, the connection will start being compressed using that dictionary.
All traffic going through the tracker is passed as a single merged "stream" through dict_sampler.
dictionary_service has access to the dict_sampler.
On chosen nodes (in the "usual" configuration: the Raft leader), it uses the sampler to maintain
a random multi-megabyte sample of the sampler's stream. Every several minutes,
it copies the sample, trains a compression dictionary on it (by calling zstd's
training library via the alien_worker thread) and publishes the new dictionary
to system.dicts via Raft's write_mutation command.
This update triggers (eventually) a callback on all nodes, which feeds the new dictionary
to advanced_rpc_compression::tracker, and this switches (eventually) all inter-node connections
to this dictionary.
Closesscylladb/scylladb#22032
* github.com:scylladb/scylladb:
messaging_service: use advanced_rpc_compression::tracker for compression
message/dictionary_service: introduce dictionary_service
service: make Raft group 0 aware of system.dicts
db/system_keyspace: add system.dicts
utils: add advanced_rpc_compressor
utils: add dict_trainer
utils: introduce reservoir_sampling
utils: introduce alien_worker
utils: add stream_compressor
now that we are allowed to use C++23. we now have the luxury of using
`std::ranges::to`.
in this change, we:
- replace `boost::copy_range` to `std::ranges::to`
- remove unused `#include` of boost headers
Signed-off-by: Kefu Chai <kefu.chai@scylladb.com>
Closesscylladb/scylladb#21880
Adds glue which causes the contents of system.dicts to be sent
in group 0 snapshots, and causes a callback to be called when
system.dicts is updated locally. The callback is currently empty
and will be hooked up to the RPC compressor tracker in one of the
next commits.
The series contains small fixes to the gossiper one of which fixes#21930. Others I noticed while debugged the issue.
Fixes: scylladb/scylladb#21930Closesscylladb/scylladb#21956
* github.com:scylladb/scylladb:
gossiper: do not reset _just_removed_endpoints in non raft mode
gossiper: do not send echo message to yourself
gossiper: do not call apply for the node's old state
"
This series converts repair, streaming and node_ops (and some parts of
alternator) to work on host ids instead of ips. This allows to remove
a lot of (but not all) functions that work on ips from effective
replication map.
CI: https://jenkins.scylladb.com/job/scylla-master/job/scylla-ci/13830/
Refs: scylladb/scylladb#21777
"
* 'gleb/move-to-host-id-more' of github.com:scylladb/scylla-dev:
locator: topology: remove no longer use get_all_ips()
gossiper: change get_unreachable_nodes to host ids
locator: drop no longer used ip based functions from effective replication map and friends
test: move network_topology_strategy_test and token_metadata_test to use host id based APIs
replica/database: drop usage of ip in favor of host id in get_keyspace_local_ranges
replica/mutation_dump: use host ids instead of ips
alternator: move ttl to work with host ids instead of ips
storage_service: move node_ops code to use host ids instead of host ips
streaming: move streaming code to use host ids instead of host ips
repair: move repair code to use host ids instead of host ips
gossiper: add get_unreachable_host_ids() function
locator: topology: add more function that return host ids to effective replication map
locator: add more function that return host ids to effective replication map
When sending by ID we should check that we do not translate our old
address to our ID and sending locally. mark_alive should not be called
with node's old ip anyway.
Add migration_task_info column to system.tablets. Set migration_task_info
value on migration request if the feature is enabled in the cluster.
Reflect the column content in tablet_metadata.
Currently truncating a table works by issuing an RPC to all the nodes which call `database::truncate_table_on_all_shards()`, which makes sure that older writes are dropped.
It works with tablets, but is not safe. A concurrent replication process may bring back old data.
This change makes makes TRUNCATE TABLE a topology operation, so that it excludes with other processes in the system which could interfere with it. More specifically, it makes TRUNCATE a global topology request.
Backporting is not needed.
Fixes#16411Closesscylladb/scylladb#19789
* github.com:scylladb/scylladb:
docs: docs: topology-over-raft: Document truncate_table request
storage_proxy: fix indentation and remove empty catch/rethrow
test: add tests for truncate with tablets
storage_proxy: use new TRUNCATE for tablets
truncate: make TRUNCATE a global topology operation
storage_service: move logic of wait_for_topology_request_completion()
RPC: add truncate_with_tablets RPC with frozen_topology_guard
feature_service: added cluster feature for system.topology schema change
system.topology_requests: change schema
storage_proxy: propagate group0 client and TSM dependency
In the cleanup commit a840949ea0
a regression was introduced that caused backward incompatible changes
in the gossiper application state name strings.
In the e486e0f759 the value
`application_state::CDC_STREAMS_TIMESTAMP` was changed to
`application_state::CDC_GENERATION_ID`, but the name string
"CDC_STREAMS_TIMESTAMP" was kept for backward compatibility.
The cleanup commit a840949ea0 however
changed the name string to "CDC_GENERATION_ID" by ommission (not noticing
the difference) which caused backward incompatible change.
There is also another case found of "IGNOR_MSB_BITS" (that has a typo -
missing the "E" in "IGNORE") to "IGNORE_MSB_BITS", which also needs to
be reverted back to keep the backward compatibility.
Fixes: scylladb/scylladb#21811Closesscylladb/scylladb#21813
The goal of merge is to reduce the tablet count for a shrinking table. Similar to how split increases the count while the table is growing. The load balancer decision to merge is implemented today (came with infrastructure introduced for split), but it wasn't handled until now.
Initial tablet count is respected while the table is in "growing mode". For example, the table leaves it if there was a need to split above the initial tablet count. After the table leaves the mode, the average size can be trusted to determine that the table is shrinking. Merge decision is emitted if the average tablet size is 50% of the target. Hysteresis is applied to avoid oscillations between split and merges.
Similar to split, the decision to merge is recorded in tablet map's resize_type field with the string "merge". This is important in case of coordinator failover, so new coordinator continues from where the old left off.
Unlike split, the preparation phase during merge is not done by the replica (with split compactions), but rather by the coordinator by co-locating sibling tablets in the same node's shard. We can define sibling tablets as tablets that have contiguous range and will become one after merge. The concept is based on the power-of-two constraint and token contiguity. For example, in a table with 4 tablets, tablets of ids 0 and 1 are siblings, 2 and 3 are also siblings.
The algorithm for co-locating sibling tablets is very simple. The balancer is responsible for it, and it will emit migrations so that "odd" tablet will follow the "even" one. For example, tablet 1 will be migrated to where tablet 0 lives. Co-location is low in priority, it's not the end of the world to delay merge, but it's not ideal to delay e.g. decommission or even regular load balancing as that can translate into temporary unbalancing, impacting the user activities. So co-location migrations will happen when there is no more important work to do.
While regular balancing is higher in priority, it will not undo the co-location work done so far. It does that by treating co-located tablets as if they were already merged. The load inversion convergence check was adjusted so balancer understand when two tablets are being migrated instead of one, to avoid oscillations.
When balancer completes co-location work for a table undergoing merge, it will put the id of the table into the resize_plan, which is about communicating with the topology coordinator that a table is ready for it. With all sibling tablets co-located, the coordinator can resize the tablet map (reduce it by a factor of 2) and record the new map into group0. All the replicas will react to it (on token metadata update) by merging the storage (memtable(s) + sstables) of sibling tablets into one.
Fixes#18181.
system test details:
test: https://github.com/pehala/scylla-cluster-tests/blob/tablets_split_merge/tablets_split_merge_test.py
yaml file: https://github.com/pehala/scylla-cluster-tests/blob/tablets_split_merge/test-cases/features/tablets/tablets-split-merge-test.yaml
instance type: i3.8xlarge
nodes: 3
target tablet size: 0.5G (scaled down by 10, to make it easier to trigger splits and merges)
description: multiple cycles of growing and shrinking the data set in order to trigger splits and merges.
data_set_size: ~100G
initial_tablets: 64, so it grew to 128 tablets on split, and back to 64 on merge.
latency of reads and writes that happened in parallel to split and merge:
```
$ for i in scylla-bench*; do cat $i | grep "Mode\|99th:\|99\.9th:"; done
Mode: write
99.9th: 3.145727ms
99th: 1.998847ms
99.9th: 3.145727ms
99th: 2.031615ms
Mode: read
99.9th: 3.145727ms
99th: 2.031615ms
99.9th: 3.145727ms
99th: 2.031615ms
Mode: write
99.9th: 3.047423ms
99th: 1.933311ms
99.9th: 3.047423ms
99th: 1.933311ms
Mode: read
99.9th: 3.145727ms
99th: 1.900543ms
99.9th: 3.145727ms
99th: 1.900543ms
Mode: write
99.9th: 5.079039ms
99th: 3.604479ms
99.9th: 35.389439ms
99th: 25.624575ms
Mode: write
99.9th: 3.047423ms
99th: 1.998847ms
99.9th: 3.047423ms
99th: 1.998847ms
Mode: read
99.9th: 3.080191ms
99th: 2.031615ms
99.9th: 3.112959ms
99th: 2.031615ms
```
Closesscylladb/scylladb#20572
* github.com:scylladb/scylladb:
docs: Document tablet merging
tests/boost: Add test to verify correctness of balancer decisions during merge
tests/topology_experimental_raft: Add tablet merge test
service: Handle exception when retrying split
service: Co-locate sibling tablets for a table undergoing merge
gms: Add cluster feature for tablet merge
service: Make merge of resize plan commutative
replica: Implement merging of compaction groups on merge completion
replica: Handle tablet merge completion
service: Implement tablet map resize for merge
locator: Introduce merge_tablet_info()
service: Rename topology::transition_state::tablet_split_finalization
service: Respect initial_tablet_count if table is in growing mode
service: Wire migration_tablet_set into the load balancer
locator: Add tablet_map::sibling_tablets()
service: Introduce sorted_replicas_for_tablet_load()
locator/tablets: Extend tablet_replica equality comparator to three-way
service: Introduce alias to per-table candidate map type
service: Add replication constraint check variant for migration_tablet_set
service: Add convergence check variant for migration_tablet_set
service: Add migration helpers for migration_tablet_set
service/tablet_allocator: Introduce migration_tablet_set
service: Introduce migration_plan::add(migrations_vector)
locator/tablets: Introduce tablet_map::for_each_sibling_tablets()
locator/tablets: Introduce tablet_map::needs_merge()
locator/tablets: Introduce resize_decision::initial_decision()
locator/tablets: Fix return type of three-way comparison operators
service: Extract update of node load on migrations
service: Extract converge check for intra-node migration
service: Extract erase of tablet replicas from candidate list
scripts/tablet-mon: Allow visualization of tablet id
In the current scenario, if during startup, a node crashes after initiating gossip and before joining group0,
then it keeps floating in the gossiper forever because the raft based gossiper purging logic is only effective
once node joins group0. This orphan node hinders the successor node from same ip to join cluster since it collides
with it during gossiper shadow round.
This commit intends to fix this issue by adding a background thread which periodically checks for such orphan entries in
gossiper and removes them.
A test is also added in to verify this logic. This test fails without this background thread enabled, hence
verifying the behavior.
Fixes: scylladb/scylladb#20082Closesscylladb/scylladb#21600
This patch adds a feature serive which protects the system.topology
schema change against situations where clusters are incompletely
upgraded to new a version and could be rolled back.
The reason we need it is that tablet merge can only be finalized
when the cluster agrees on the feature, otherwise unpatched
nodes would fail to handle merge finalization, potentially
crashing.
Signed-off-by: Raphael S. Carvalho <raphaelsc@scylladb.com>
"
This rather large patch series moves storage proxy and some adjacent
services (like migration manager) to use host ids to identify nodes rather
than ips. Messaging service gains a capability to address nodes by host
ids (which allows dropping translations from topology coordinator code
that worked on host ids already) and also makes sure that a node with
incorrect host id will reject a message (can happen during address
changes).
The series gets rid of the raft address map completely and replaces it with
the gossiper address map which is managed by the gossiper since translation
is now done in the layer below raft.
Fixes: scylladb/scylladb#6403
perf-simple-query -- smp 1 -m 1G output
Before:
enable-cache=1
Running test with config: {partitions=10000, concurrency=100, mode=read, frontend=cql, query_single_key=no, counters=no}
Disabling auto compaction
Creating 10000 partitions...
64336.82 tps ( 63.1 allocs/op, 0.0 logallocs/op, 14.1 tasks/op, 41291 insns/op, 24485 cycles/op, 0 errors)
62669.58 tps ( 63.1 allocs/op, 0.0 logallocs/op, 14.1 tasks/op, 41277 insns/op, 24695 cycles/op, 0 errors)
69172.12 tps ( 63.1 allocs/op, 0.0 logallocs/op, 14.2 tasks/op, 41326 insns/op, 24463 cycles/op, 0 errors)
56706.60 tps ( 63.1 allocs/op, 0.0 logallocs/op, 14.1 tasks/op, 41143 insns/op, 24513 cycles/op, 0 errors)
56416.65 tps ( 63.1 allocs/op, 0.0 logallocs/op, 14.1 tasks/op, 41186 insns/op, 24851 cycles/op, 0 errors)
throughput: mean=61860.35 standard-deviation=5395.48 median=62669.58 median-absolute-deviation=5153.75 maximum=69172.12 minimum=56416.65
instructions_per_op: mean=41244.62 standard-deviation=76.90 median=41276.94 median-absolute-deviation=58.55 maximum=41326.19 minimum=41142.80
cpu_cycles_per_op: mean=24601.35 standard-deviation=167.39 median=24512.64 median-absolute-deviation=116.65 maximum=24851.45 minimum=24462.70
After:
enable-cache=1
Running test with config: {partitions=10000, concurrency=100, mode=read, frontend=cql, query_single_key=no, counters=no}
Disabling auto compaction
Creating 10000 partitions...
65237.35 tps ( 63.1 allocs/op, 0.0 logallocs/op, 14.2 tasks/op, 40733 insns/op, 23145 cycles/op, 0 errors)
59283.09 tps ( 63.1 allocs/op, 0.0 logallocs/op, 14.1 tasks/op, 40624 insns/op, 23948 cycles/op, 0 errors)
70851.03 tps ( 63.1 allocs/op, 0.0 logallocs/op, 14.1 tasks/op, 40625 insns/op, 23027 cycles/op, 0 errors)
70549.61 tps ( 63.1 allocs/op, 0.0 logallocs/op, 14.1 tasks/op, 40650 insns/op, 23266 cycles/op, 0 errors)
68634.96 tps ( 63.1 allocs/op, 0.0 logallocs/op, 14.1 tasks/op, 40622 insns/op, 22935 cycles/op, 0 errors)
throughput: mean=66911.21 standard-deviation=4814.60 median=68634.96 median-absolute-deviation=3638.40 maximum=70851.03 minimum=59283.09
instructions_per_op: mean=40650.89 standard-deviation=47.55 median=40624.60 median-absolute-deviation=27.11 maximum=40733.37 minimum=40622.33
cpu_cycles_per_op: mean=23264.16 standard-deviation=402.12 median=23145.29 median-absolute-deviation=237.63 maximum=23947.96 minimum=22934.59
CI: https://jenkins.scylladb.com/job/scylla-master/job/scylla-ci/13531/
SCT (longevity-100gb-4h with nemesis_selector: ['topology_changes']): https://jenkins.scylladb.com/view/staging/job/scylla-staging/job/gleb/job/move-to-host-id/3/
Tested mixed cluster manually.
"
* 'gleb/move-to-host-id-v2' of github.com:scylladb/scylla-dev: (55 commits)
group0: drop unused field from replace_info struct
test: rename raft_address_map_test to address_map_test and move if from raft tests
raft_address_map: remove raft address map
topology coordinator: do not modify expire state for left/new nodes any more in raft address map
topology coordinator: drop expiring entries in gossiper address map on error injections since raft one is no longer used
group0: drop raft address map dependency from raft_rpc
group0: move raft_ticker_type definition from raft_address_map.hh
storage_service: do not update raft address map on gossiper events
group0: drop raft address map dependency from raft_server_with_timeouts
group0: move group0 upgrade code to host ids
repair: drop raft address map dependency
group0: remove unused raft address map getter from raft_group0
group0: drop raft address map from group0_state_machine dependency since it is not used there any more
group0: remove dependency on raft address map from group0_state_id_handler
gossiper: add get_application_state_ptr that searches by host_id
gossiper: change get_live_token_owners to return host ids
view: move view building to host id
hints: use host id to send hints
storage_proxy: remove id_vector_to_addr since it is no longer used
db: consistency_level: change is_sufficient_live_nodes to work on host ids
...
now that we are allowed to use C++23. we now have the luxury of using
`std::views::transform`.
in this change, we:
- replace `boost::adaptors::transformed` with `std::views::transform`
- use `fmt::join()` when appropriate where `boost::algorithm::join()`
is not applicable to a range view returned by `std::view::transform`.
- use `std::ranges::fold_left()` to accumulate the range returned by
`std::view::transform`
- use `std::ranges::fold_left()` to get the maximum element in the
range returned by `std::view::transform`
- use `std::ranges::min()` to get the minimal element in the range
returned by `std::view::transform`
- use `std::ranges::equal()` to compare the range views returned
by `std::view::transform`
- remove unused `#include <boost/range/adaptor/transformed.hpp>`
- use `std::ranges::subrange()` instead of `boost::make_iterator_range()`,
to feed `std::views::transform()` a view range.
to reduce the dependency to boost for better maintainability, and
leverage standard library features for better long-term support.
this change is part of our ongoing effort to modernize our codebase
and reduce external dependencies where possible.
limitations:
there are still a couple places where we are still using
`boost::adaptors::transformed` due to the lack of a C++23 alternative
for `boost::join()` and `boost::adaptors::uniqued`.
Signed-off-by: Kefu Chai <kefu.chai@scylladb.com>
Closesscylladb/scylladb#21700
RPCs from old nodes will still use old format so translation will be
used in this case. The change is backwards compatible thanks to RPC
extensibility.
The parameter was needed when nodes were addressed by IP, so during
replace with the same IP a new node had to "hide" itself from the
cluster to not get accidentally confused with the old node. Now,
when nodes are addressed by host id the situation is impossible.
The function checks if a node with provided id is alive. If it fails to map id to ip
or there is no state for the ip found the node is considered to be dead.
`coroutine::parallel_for_each` accepts both a range and a pair of
iterators. let's use the former when appropriate. it is simpler this way.
Signed-off-by: Kefu Chai <kefu.chai@scylladb.com>
Closesscylladb/scylladb#21684
More gossiper cleanups that accumulated since the previous one.
* 'gleb/more-gossip-cleanup-v2' of github.com:scylladb/scylla-dev:
gossiper: replace milliseconds with seconds where appropriate
gossiper: simplify failure_detector_loop loop a bit
gossiper: use fmt library to format time
gossiper: drop on_success callback from mutate_live_and_unreachable_endpoints
gossiper: remove code duplication between shadow round and regular path when state is applied
gossiper: remove remnants of old shadow round
gossiper: fix indentation after the last patch
gossiper: co-routinize do_shadow_round
