This is the following squashed patches that were reverted.
I will fix the known issues with some follow-on patches.
======================================================================
IMPALA-4835: Part 1: simplify I/O mgr mem mgmt and cancellation
In preparation for switching the I/O mgr to the buffer pool, this
removes and cleans up a lot of code so that the switchover patch starts
from a cleaner slate.
* Remove the free buffer cache (which will be replaced by buffer pool's
own caching).
* Make memory limit exceeded error checking synchronous (in anticipation
of having to propagate buffer pool errors synchronously).
* Simplify error propagation - remove the (ineffectual) code that
enqueued BufferDescriptors containing error statuses.
* Document locking scheme better in a few places, make it part of the
function signature when it seemed reasonable.
* Move ReturnBuffer() to ScanRange, because it is intrinsically
connected with the lifecycle of a scan range.
* Separate external ReturnBuffer() and internal CleanUpBuffer()
interfaces - previously callers of ReturnBuffer() were fudging
the num_buffers_in_reader accounting to make the external interface work.
* Eliminate redundant state in ScanRange: 'eosr_returned_' and
'is_cancelled_'.
* Clarify the logic around calling Close() for the last
BufferDescriptor.
-> There appeared to be an implicit assumption that buffers would be
freed in the order they were returned from the scan range, so that
the "eos" buffer was returned last. Instead just count the number
of outstanding buffers to detect the last one.
-> Touching the is_cancelled_ field without holding a lock was hard to
reason about - violated locking rules and it was unclear that it
was race-free.
* Remove DiskIoMgr::Read() to simplify the interface. It is trivial to
inline at the callsites.
This will probably regress performance somewhat because of the cache
removal, so my plan is to merge it around the same time as switching
the I/O mgr to allocate from the buffer pool. I'm keeping the patches
separate to make reviewing easier.
Testing:
* Ran exhaustive tests
* Ran the disk-io-mgr-stress-test overnight
======================================================================
IMPALA-4835: Part 2: Allocate scan range buffers upfront
This change is a step towards reserving memory for buffers from the
buffer pool and constraining per-scanner memory requirements. This
change restructures the DiskIoMgr code so that each ScanRange operates
with a fixed set of buffers that are allocated upfront and recycled as
the I/O mgr works through the ScanRange.
One major change is that ScanRanges get blocked when a buffer is not
available and get unblocked when a client returns a buffer via
ReturnBuffer(). I was able to remove the logic to maintain the
blocked_ranges_ list by instead adding a separate set with all ranges
that are active.
There is also some miscellaneous cleanup included - e.g. reducing the
amount of code devoted to maintaining counters and metrics.
One tricky part of the existing code was the it called
IssueInitialRanges() with empty lists of files and depended on
DiskIoMgr::AddScanRanges() to not check for cancellation in that case.
See IMPALA-6564/IMPALA-6588. I changed the logic to not try to issue
ranges for empty lists of files.
I plan to merge this along with the actual buffer pool switch, but
separated it out to allow review of the DiskIoMgr changes separate from
other aspects of the buffer pool switchover.
Testing:
* Ran core and exhaustive tests.
======================================================================
IMPALA-4835: Part 3: switch I/O buffers to buffer pool
This is the final patch to switch the Disk I/O manager to allocate all
buffer from the buffer pool and to reserve the buffers required for
a query upfront.
* The planner reserves enough memory to run a single scanner per
scan node.
* The multi-threaded scan node must increase reservation before
spinning up more threads.
* The scanner implementations must be careful to stay within their
assigned reservation.
The row-oriented scanners were most straightforward, since they only
have a single scan range active at a time. A single I/O buffer is
sufficient to scan the whole file but more I/O buffers can improve I/O
throughput.
Parquet is more complex because it issues a scan range per column and
the sizes of the columns on disk are not known during planning. To
deal with this, the reservation in the frontend is based on a
heuristic involving the file size and # columns. The Parquet scanner
can then divvy up reservation to columns based on the size of column
data on disk.
I adjusted how the 'mem_limit' is divided between buffer pool and non
buffer pool memory for low mem_limits to account for the increase in
buffer pool memory.
Testing:
* Added more planner tests to cover reservation calcs for scan node.
* Test scanners for all file formats with the reservation denial debug
action, to test behaviour when the scanners hit reservation limits.
* Updated memory and buffer pool limits for tests.
* Added unit tests for dividing reservation between columns in parquet,
since the algorithm is non-trivial.
Perf:
I ran TPC-H and targeted perf locally comparing with master. Both
showed small improvements of a few percent and no regressions of
note. Cluster perf tests showed no significant change.
Change-Id: I3ef471dc0746f0ab93b572c34024fc7343161f00
Reviewed-on: http://gerrit.cloudera.org:8080/9679
Reviewed-by: Tim Armstrong <tarmstrong@cloudera.com>
Tested-by: Tim Armstrong <tarmstrong@cloudera.com>
Support row_regex and other lines for the subset and superset verifiers,
which previously assumed that lines in the actual and expected had to
match exactly.
Use in test_stats_extrapolation to make the test more robust to
irrelevant changes in the explain plan.
Testing:
Manually modified a superset and a subset test to check that tests fail
as expected.
Change-Id: Ia7a28d421c8e7cd84b14d07fcb71b76449156409
Reviewed-on: http://gerrit.cloudera.org:8080/10155
Reviewed-by: Tim Armstrong <tarmstrong@cloudera.com>
Tested-by: Impala Public Jenkins <impala-public-jenkins@cloudera.com>
Revert "IMPALA-6585: increase test_low_mem_limit_q21 limit"
This reverts commit 25bcb258df.
Revert "IMPALA-6588: don't add empty list of ranges in text scan"
This reverts commit d57fbec6f6.
Revert "IMPALA-4835: Part 3: switch I/O buffers to buffer pool"
This reverts commit 24b4ed0b29.
Revert "IMPALA-4835: Part 2: Allocate scan range buffers upfront"
This reverts commit 5699b59d0c.
Revert "IMPALA-4835: Part 1: simplify I/O mgr mem mgmt and cancellation"
This reverts commit 65680dc421.
Change-Id: Ie5ca451cd96602886b0a8ecaa846957df0269cbb
Reviewed-on: http://gerrit.cloudera.org:8080/9480
Reviewed-by: Dan Hecht <dhecht@cloudera.com>
Tested-by: Impala Public Jenkins
This is the final patch to switch the Disk I/O manager to allocate all
buffer from the buffer pool and to reserve the buffers required for
a query upfront.
* The planner reserves enough memory to run a single scanner per
scan node.
* The multi-threaded scan node must increase reservation before
spinning up more threads.
* The scanner implementations must be careful to stay within their
assigned reservation.
The row-oriented scanners were most straightforward, since they only
have a single scan range active at a time. A single I/O buffer is
sufficient to scan the whole file but more I/O buffers can improve I/O
throughput.
Parquet is more complex because it issues a scan range per column and
the sizes of the columns on disk are not known during planning. To
deal with this, the reservation in the frontend is based on a
heuristic involving the file size and # columns. The Parquet scanner
can then divvy up reservation to columns based on the size of column
data on disk.
I adjusted how the 'mem_limit' is divided between buffer pool and non
buffer pool memory for low mem_limits to account for the increase in
buffer pool memory.
Testing:
* Added more planner tests to cover reservation calcs for scan node.
* Test scanners for all file formats with the reservation denial debug
action, to test behaviour when the scanners hit reservation limits.
* Updated memory and buffer pool limits for tests.
* Added unit tests for dividing reservation between columns in parquet,
since the algorithm is non-trivial.
Perf:
I ran TPC-H and targeted perf locally comparing with master. Both
showed small improvements of a few percent and no regressions of
note. Cluster perf tests showed no significant change.
Change-Id: Ic09c6196b31e55b301df45cc56d0b72cfece6786
Reviewed-on: http://gerrit.cloudera.org:8080/8966
Reviewed-by: Tim Armstrong <tarmstrong@cloudera.com>
Tested-by: Impala Public Jenkins
Introduces a new TBLPROPERTY for controlling stats
extrapolation on a per-table basis:
impala.enable.stats.extrapolation=true/false
The property key was chosen to be consistent with
the impalad startup flag --enable_stats_extrapolation
and to indicate that the property was set and is used
by Impala.
Behavior:
- If the property is not set, then the extrapolation
behavior is determined by the impalad startup flag.
- If the property is set, it overrides the impalad
startup flag, i.e., extrapolation can be explicitly
enabled or disabled regardless of the startup flag.
Testing:
- added new unit tests
- code/hdfs run passed
Change-Id: Ie49597bf1b93b7572106abc620d91f199cba0cfd
Reviewed-on: http://gerrit.cloudera.org:8080/9139
Reviewed-by: Alex Behm <alex.behm@cloudera.com>
Tested-by: Impala Public Jenkins
This change enables clustering by default. IMPALA-2521 introduced the
'clustered' hint which inserts a local sort by the partitioning columns
to a query plan. The hint is only effective for HDFS and Kudu tables.
Like before, the 'noclustered' hint prevents clustering. If a table has
ordering columns defined, the 'noclustered' hint is ignored and we
issue a warning.
This change removes some tests that were added specifically to test
that clustering can be enabled using the 'clustered' hint. It changes
some tests to use the 'noclustered' hint to make sure that clustering
can be disabled. It also adds tests to make sure that we cover the
'noclustered' case properly.
Cherry-picks: not for 2.x.
Change-Id: Idbf2368cf4415e6ecfa65058daf6ff87ef62f9d9
Reviewed-on: http://gerrit.cloudera.org:8080/9153
Reviewed-by: Lars Volker <lv@cloudera.com>
Tested-by: Impala Public Jenkins
Adds the TABLESAMPLE clause for COMPUTE STATS.
Syntax:
COMPUTE STATS <table> TABLESAMPLE SYSTEM(<number>) [REPEATABLE(<number>)]
Computes and replaces the table-level row count and total file size,
as well as all table-level column statistics. Existing partition-level
row counts are not modified.
The TABLESAMPLE clause can be used to limit the scanned data volume to
a desired percentage. When sampling, the unmodified results of the
COMPUTE STATS queries are sent to the CatalogServer. There, the stats
are extrapolated before storing them into the HMS so as not to confuse
other engines like Hive/SparkSQL which may rely on the shared HMS
fields being accurate.
Limitations
- Only works for HDFS tables
- TABLESAMPLE is not supported for COMPUTE INCREMENTAL STATS
- TABLESAMPLE requires --enable_stats_extrapolation=true
Changes to EXPLAIN
The stored statistics from the HMS are more clearly displayed under
a 'stored statistics' section. Example:
00:SCAN HDFS [functional.alltypes, RANDOM]
partitions=24/24 files=24 size=478.45KB
stored statistics:
table: rows=7300 size=478.45KB
partitions: 24/24 rows=7300
columns: all
Testing:
- added new functional tests
- core/hdfs run passed
Change-Id: I7f3e72471ac563adada4a4156033a85852b7c8b7
Reviewed-on: http://gerrit.cloudera.org:8080/8136
Reviewed-by: Alex Behm <alex.behm@cloudera.com>
Tested-by: Impala Public Jenkins
Before this change, the per-host reservation size was computed
by the Planner. However, scheduling happens after planning,
so the Planner must assume that all fragments run on all
hosts, and the reservation size is likely much larger than
it needs to be.
This moves the computation of the per-host reservation size
to the BE where it can be computed more precisely. This also
includes a number of plan/profile changes.
Change-Id: Idbcd1e9b1be14edc4017b4907e83f9d56059fbac
Reviewed-on: http://gerrit.cloudera.org:8080/7630
Reviewed-by: Matthew Jacobs <mj@cloudera.com>
Tested-by: Impala Public Jenkins
This moves away from the PipelinedPlanNodeSet approach of enumerating
sets of concurrently-executing nodes because unions would force
creating many overlapping sets of nodes. The new approach computes
the peak resources during Open() and the peak resources between Open()
and Close() (i.e. while calling GetNext()) bottom-up for each plan node
in a fragment. The fragment resources are then combined to produce the
query resources.
The basic assumptions for the new resource estimates are:
* resources are acquired during or after the first call to Open()
and released in Close().
* Blocking nodes call Open() on their child before acquiring
their own resources (this required some backend changes).
* Blocking nodes call Close() on their children before returning
from Open().
* The peak resource consumption of the query is the sum of the
independent fragments (except for the parallel join build plans
where we can assume there will be synchronisation). This is
conservative but we don't synchronise fragment Open() and Close()
across exchanges so can't make stronger assumptions in general.
Also compute the sum of minimum reservations. This will be useful
in the backend to determine exactly when all of the initial
reservations have been claimed from a shared pool of initial reservations.
Testing:
* Updated planner tests to reflect behavioural changes.
* Added extra resource requirement planner tests for unions, subplans,
pipelines of blocking operators, and bushy join plans.
* Added single-node plans to resource-requirements tests. These have
more complex plan trees inside a single fragment, which is useful
for testing the peak resource requirement logic.
Change-Id: I492cf5052bb27e4e335395e2a8f8a3b07248ec9d
Reviewed-on: http://gerrit.cloudera.org:8080/7223
Reviewed-by: Tim Armstrong <tarmstrong@cloudera.com>
Tested-by: Impala Public Jenkins
This is similar to the single-node execution optimisation, but applies
to slightly larger queries that should run in a distributed manner but
won't benefit from codegen.
This adds a new query option disable_codegen_rows_threshold that
defaults to 50,000. If fewer than this number of rows are processed
by a plan node per impalad, the cost of codegen almost certainly
outweighs the benefit.
Using rows processed as a threshold is justified by a simple
model that assumes the cost of codegen and execution per row for
the same operation are proportional. E.g. if x is the complexity
of the operation, n is the number of rows processed, C is a
constant factor giving the cost of codegen and Ec/Ei are constant
factor giving the cost of codegen'd and interpreted execution and
d, then the cost of the codegen'd operator is C * x + Ec * x * n
and the cost of the interpreted operator is Ei * x * n. Rearranging
means that interpretation is cheaper if n < C / (Ei - Ec), i.e. that
(at least with the simplified model) it makes sense to choose
interpretation or codegen based on a constant threshold. The
model also implies that it is somewhat safer to choose codegen
because the additional cost of codegen is O(1) but the additional
cost of interpretation is O(n).
I ran some experiments with TPC-H Q1, varying the input table size, to
determine what the cut-over point where codegen was beneficial was.
The cutover was around 150k rows per node for both text and parquet.
At 50k rows per node disabling codegen was very beneficial - around
0.12s versus 0.24s. To be somewhat conservative I set the default
threshold to 50k rows. On more complex queries, e.g. TPC-H Q10, the
cutover tends to be higher because there are plan nodes that process
many fewer than the max rows.
Fix a couple of minor issues in the frontend - the numNodes_
calculation could return 0 for Kudu, and the single node optimization
didn't handle the case where for a scan node with conjuncts, a limit
and missing stats correctly (it considered the estimate still valid.)
Testing:
Updated e2e tests that set disable_codegen to set
disable_codegen_rows_threshold to 0, so that those tests run both
with and without codegen still.
Added an e2e test to make sure that the optimisation is applied in
the backend.
Added planner tests for various cases where codegen should and shouldn't
be disabled.
Perf:
Added a targeted perf test for a join+agg over a small input, which
benefits from this change.
Change-Id: I273bcee58641f5b97de52c0b2caab043c914b32e
Reviewed-on: http://gerrit.cloudera.org:8080/7153
Reviewed-by: Tim Armstrong <tarmstrong@cloudera.com>
Tested-by: Impala Public Jenkins
The main idea of this patch is to use table stats to
extrapolate the row counts for new/modified partitions.
Existing behavior:
- Partitions that lack the row count stat are ignored
when estimating the cardinality of HDFS scans. Such
partitions effectively have an estimated row count
of zero.
- We always use the row count stats for partitions that
have one. The row count may be innaccurate if data in
such partitions has changed significantly.
Summary of changes:
- Enhance COMPUTE STATS to also store the total number
of file bytes in the table.
- Use the table-level row count and file bytes stats
to estimate the number of rows in a scan.
- A new impalad startup flag is added to enable/disable
the extrapolation behavior. The feature is disabled by
default. Note that even with the feature disabled,
COMPUTE STATS stores the file bytes so you can enable
the feature without having to run COMPUTE STATS again.
Testing:
- Added new FE unit test
- Added new EE test
Change-Id: I972c8a03ed70211734631a7dc9085cb33622ebc4
Reviewed-on: http://gerrit.cloudera.org:8080/6840
Reviewed-by: Alex Behm <alex.behm@cloudera.com>
Tested-by: Impala Public Jenkins
