Commits · 212c57421bd20478cedf4e07ad01bca21815265a · core / tarantool

Jul 12, 2019

func: consolidate func_def checks in func_def_chekc · 212c5742
Konstantin Osipov authored 5 years ago
```
A follow up on #4182
```
212c5742

box: introduce Lua persistent functions · 200a492a

Closes #4182
Closes #4219
Needed for #1260

@TarantoolBot document
Title: Persistent Lua functions

Now Tarantool supports 'persistent' Lua functions.
Such functions are stored in snapshot and are available after
restart.
To create a persistent Lua function, specify a function body
in box.schema.func.create call:
e.g. body = "function(a, b) return a + b end"

A Lua persistent function may be 'sandboxed'. The 'sandboxed'
function is executed in isolated environment:
  a. only limited set of Lua functions and modules are available:
    -assert -error -pairs -ipairs -next -pcall -xpcall -type
    -print -select -string -tonumber -tostring -unpack -math -utf8;
  b. global variables are forbidden

Finally, the new 'is_deterministic' flag allows to mark a
registered function as deterministic, i.e. the function that
can produce only one result for a given list of parameters.

The new box.schema.func.create interface is:
box.schema.func.create('funcname', <setuid = true|FALSE>,
	<if_not_exists = true|FALSE>, <language = LUA|c>,
	<body = string ('')>, <is_deterministic = true|FALSE>,
	<is_sandboxed = true|FALSE>, <comment = string ('')>)

This schema change is also reserves names for sql builtin
functions:
    TRIM, TYPEOF, PRINTF, UNICODE, CHAR, HEX, VERSION,
    QUOTE, REPLACE, SUBSTR, GROUP_CONCAT, JULIANDAY, DATE,
    TIME, DATETIME, STRFTIME, CURRENT_TIME, CURRENT_TIMESTAMP,
    CURRENT_DATE, LENGTH, POSITION, ROUND, UPPER, LOWER,
    IFNULL, RANDOM, CEIL, CEILING, CHARACTER_LENGTH,
    CHAR_LENGTH, FLOOR, MOD, OCTET_LENGTH, ROW_COUNT, COUNT,
    LIKE, ABS, EXP, LN, POWER, SQRT, SUM, TOTAL, AVG,
    RANDOMBLOB, NULLIF, ZEROBLOB, MIN, MAX, COALESCE, EVERY,
    EXISTS, EXTRACT, SOME, GREATER, LESSER, SOUNDEX,
    LIKELIHOOD, LIKELY, UNLIKELY,
    _sql_stat_get, _sql_stat_push, _sql_stat_init, LUA

A new Lua persistent function LUA is introduced to evaluate
LUA strings from SQL in future.

This names could not be used for user-defined functions.

Example:
lua_code = [[function(a, b) return a + b end]]
box.schema.func.create('summarize', {body = lua_code,
		is_deterministic = true, is_sandboxed = true})
box.func.summarize
---
- aggregate: none
  returns: any
  exports:
    lua: true
    sql: false
  id: 60
  is_sandboxed: true
  setuid: false
  is_deterministic: true
  body: function(a, b) return a + b end
  name: summarize
  language: LUA
...
box.func.summarize:call({1, 3})
---
- 4
...

@kostja: fix style, remove unnecessary module dependencies,
add comments

200a492a

box: do not check state in case of reconnect · 77051a11

Mergen Imeev authored 5 years ago

Test box/net.box.test.lua checks state of the connection in case
of an error. It should be 'error_reconnect'. But, in cases where
testing was performed on a slow computer or in the case of a very
large load, it is possible that the connection status may change
from the 'error_reconnect' state to another state. This led to the
failure of the test. Since this check is not the main purpose of
the test, it is better to simply delete the check.

Closes #4335

77051a11

sql: move LIKE UConverter object to collation library · f9551660

Kirill Shcherbatov authored 5 years ago

Moved UConverter object to collation library. This is required
to get rid of sqlRegisterBuiltinFunctions function in further
patches.

Needed for #4113, #2200, #2233

f9551660

sql: replace bool is_derived_coll marker with flag · 4e523fa0

Kirill Shcherbatov authored 5 years ago

Introduce a new flag SQL_FUNC_DERIVEDCOLL for function that may
require collation to be applied on its result instead of separate
boolean variable. This is required to get rid of FuncDef in
further patches.

Needed for #4113, #2200, #2233

4e523fa0

sql: put analyze helpers to FuncDef cache · e1a7cb7f

Kirill Shcherbatov authored 5 years ago

Previously analyze functions refer to statically defined
service FuncDef context. We need to change this approach due we
going to rework the builtins functions machinery in following
patches.

Needed for #4113, #2200, #2233

e1a7cb7f

Jul 11, 2019

sql: enable SOUNDEX sql builtin · 873a3b30

Kirill Shcherbatov authored 5 years ago

Needed for #4182

@TarantoolBot document
Title: Introduce SOUNDEX sql function

The SOUNDEX function returns a 4-character code that represents
the sound of the words in the argument. The result can be
compared to the results of the SOUNDEX function of other strings.

The current SOUNDEX function supports only Latin strings.

@kostja: fix test tap count; remove optional invocation; remove
trailing spaces; fix alignment.

873a3b30

box/memtx: Skip tuple memory from coredump by default · 9d077bb4

Cyrill Gorcunov authored 5 years ago

Quoting feature request

 | Tarantool is Database and Application Server in one box.
 |
 | Appserver development process contains a lot of
 | lua/luajit-ffi/lua-c-extension code.
 |
 | Coredump is very useful in case when some part of appserver crashed.
 | If the reason is input - data from database is not necessary. If the reason
 | is output - data from database is already in snap/xlog files.
 |
 | Therefore consider core dumps without data enabled by default.

For info: the strip_core feature has been introduced in
549140b3

Closes #4337

@TarantoolBot document
Title: Document box.cfg.strip_core

When Tarantool runs under a heavy load the memory allocated
for tuples may be very huge in size and to eliminate this
memory from being present in `coredump` file the `box.cfg.strip_core`
parameter should be set to `true`.

The default value is `true`.

9d077bb4

sql: get rid of MATCH sql builtin · 4d936cb2

Kirill Shcherbatov authored 5 years ago

In relation with FuncDef cache rework we need to clean-up
builtins list. The MATCH fucntion is a stub that raises an error
and it could be dropped.

Needed for #4182

@kostja: make the patch actually pass the tests, remove
tap count change in e_expr.test.lua, since it's disabled and was
not run

4d936cb2

Fix travis script to install libiconv on macos using homebrew · f2876e63
Michael Filonenko authored 6 years ago

f2876e63

test: box/net.box test flaky fails on grep_log (#4330) · 9bde3406

avtikhon authored 5 years ago

box/net.box test flaky failed on grepping the log file
for 'ER_NO_SUCH_PROC' pattern on high load running hosts,
found that the issue can be resolved by updating the
grep_log to wait_log function to make able to wait the
needed message for some time.

[008] Test failed! Result content mismatch:
[008] --- box/net.box.result	Tue Jul  9 17:00:24 2019
[008] +++ box/net.box.reject	Tue Jul  9 17:03:34 2019
[008] @@ -1376,7 +1376,7 @@
[008]  ...
[008]  test_run:grep_log("default", "ER_NO_SUCH_PROC")
[008]  ---
[008] -- ER_NO_SUCH_PROC
[008] +- null
[008]  ...
[008]  box.schema.user.revoke('guest', 'execute', 'universe')
[008]  ---

Closes #4329

9bde3406

Add python environment for test-run (#4284) · 8dc6486c

Denis Ignatenko authored 5 years ago

Add python-dev and pip to container to install test-run dependency
Install test-run dependency from requirements.txt of test-run subrepo

8dc6486c

Add distribution info to box.info · 366466eb

Denis Ignatenko authored 5 years ago

There is compile time option PACKAGE in cmake to define
current build distribution info. For community edition
is it Tarantool by default. For enterprise it is
Tarantool Enterprise

There were no option to check distribution name in runtime.
This change adds box.info.package output for CE and TE.

366466eb

travis-ci: apt-get fails to update the APT repos (#4332) · 68dbd154

avtikhon authored 5 years ago

travis-ci APT repository update failed on Debian 10 (Buster)
with command 'apt-get update', like:

Get:1 http://deb.debian.org/debian buster InRelease [118 kB]
Get:2 http://security.debian.org/debian-security buster/updates InRelease [39.1 kB]
Get:3 http://deb.debian.org/debian buster-updates InRelease [46.8 kB]
Reading package lists... Done
N: Repository 'http://security.debian.org/debian-security buster/updates InRelease' changed its 'Version' value from '' to '10'
E: Repository 'http://security.debian.org/debian-security buster/updates InRelease' changed its 'Suite' value from 'testing' to 'stable'
N: This must be accepted explicitly before updates for this repository can be applied. See apt-secure(8) manpage for details.
N: Repository 'http://deb.debian.org/debian buster InRelease' changed its 'Version' value from '' to '10.0'
E: Repository 'http://deb.debian.org/debian buster InRelease' changed its 'Suite' value from 'testing' to 'stable'
N: This must be accepted explicitly before updates for this repository can be applied. See apt-secure(8) manpage for details.
E: Repository 'http://deb.debian.org/debian buster-updates InRelease' changed its 'Suite' value from 'testing-updates' to 'stable-updates'
N: This must be accepted explicitly before updates for this repository can be applied. See apt-secure(8) manpage for details.

The cause of the issue was:

According to Debian Project News published 1st of July, Debian 10 "Buster"
was scheduled to transition from testing to stable on 2019-07-06.
It looks like the transition was in fact performed as scheduled, and so the
testing distribution should now be catching up witn unstable, to eventually
become Debian 11 "Bullseye". You might be experiencing some temporary side
effects because of this transition of the Debian mirrors.
If you want to stay with Debian 10 "Buster", now would be a good time to switch
your /etc/apt/sources.list to use the release name buster instead of testing.
Otherwise, you'll soon be getting the raw bleeding-edge stuff from unstable,
and you might accidentally get a partial upgrade to proto-"Bullseye".
Also, this is a reminder for anyone using the word stable in their
/etc/apt/sources.list to consider whether to change it to stretch and stay with
the old version, or read the Release Notes and perform the recommended upgrade
steps.

To fix the issue it had to accept interactively the changes in the repositories
NOTE: apt instead of apt-get can accept the changes interactively

apt update -y

either accept only the needed changes for 'version' and 'suite'

apt-get update --allow-releaseinfo-change-version --allow-releaseinfo-change-suite

Seems that the only 'version' and 'suite' accept is better than blind
accept of the all changes.

Closes #4331

68dbd154

tarantoolctl: always initialize notify_socket (#4342) · 9f76bd86

Yaroslav Dynnikov authored 5 years ago

Notify socket used to be initialized during `box.cfg()`.
There is no apparent reason for that, because we can write tarantool
apps that don't use box api at all, but still leverage the event loop
and async operations.

This patch makes initialization of notify socket independent.
Instance can notify entering event loop even if box.cfg wasn't called.

Closes #4305

9f76bd86

sql: rfc for SQL and Lua functions · dc2e7a7b
Kirill Shcherbatov authored 5 years ago
```
Part of #4182
```
dc2e7a7b

sql: fix passing FP values to integer iterator · 8fac6972

Nikita Pettik authored 5 years ago

Before this patch it was impossible to compare indexed field of integer
type and floating point value. For instance:

CREATE TABLE t1(id INT PRIMARY KEY, a INT UNIQUE);
INSERT INTO t1 VALUES (1, 1);
SELECT * FROM t1 WHERE a = 1.5;
---
- error: 'Failed to execute SQL statement: Supplied key type of part 0 does not match
    index part type: expected integer'
...

That happened due to the fact that type casting mechanism (OP_ApplyType)
doesn't affect FP value when it is converted to integer. Hence, FP value
was passed to the iterator over integer field which resulted in error.
Meanwhile, comparison of integer and FP values is legal in SQL.  To cope
with this problem for each equality comparison involving integer field
we emit OP_MustBeInt, which checks whether value to be compared is
integer or not. If the latter, we assume that result of comparison is
always false and continue processing query.  For inequality constraints
we pass auxiliary flag to OP_Seek** opcodes to notify it that one of key
fields must be truncated to integer (in case of FP value) alongside with
changing iterator's type: a > 1.5 -> a >= 2.

Closes #4187

8fac6972

sql: remove redundant type derivation from QP · 0d5e757d

Nikita Pettik authored 5 years ago

Before value to be scanned in index search is passed to the iterator, it
is subjected to implicit type casting (which is implemented by
OP_ApplyType). If value can't be converted to required type,
user-friendly message is raised. Without this cast, type of iterator may
not match with type of key which in turn results in unexpected error.
However, array of types which is used to provide type conversions is
different from types of indexed fields: it is modified depending on
types of comparison's operands. For instance, when boolean field is
compared with blob value, resulting type is assumed to be scalar. In
turn, conversion to scalar is no-op. As a result, value with MP_BIN
format is passed to the iterator over boolean field. To fix that let's
remove this transformation of types. Moreover, it seems to be redundant.

Part of #4187

0d5e757d

sql: remove redundant check of space format from QP · 3570f366

Nikita Pettik authored 5 years ago

In SQL we are able to execute queries involving spaces only with formats.
Otherwise, at the very beginning of query compilation error is raised.
So, after that any checks of format existence are redundant.

3570f366

sql: fix antisymmetric boolean comparison in VDBE · db12efaf

Nikita Pettik authored 5 years ago

There are a few situations when booleans can be compared with values of
other types. To process them, we assume that booleans are always less
than numbers, which in turn are less than strings. On the other hand,
function which implements internal comparison of values -
sqlMemCompare() always returns 'less' result if one of values is boolean
and another one is not, ignoring the order of values. For instance:

... max (false, 'abc') -> 'abc'
... max ('abc', false) -> false

This patch fixes this misbehaviour making boolean values always less
than values of other types.

db12efaf

sqltester: don't convert booleans to 0/1 numerics · 7bd1dc4b

Nikita Pettik authored 5 years ago

When there were no booleans in SQL, to represent them numeric values 0
and 1 were involved. However, recently booleans have been introduced in
SQL, so values of result set can take boolean values. Hence, it makes
no sense to continue converting booleans to numeric, so we can use
directly booleans.

7bd1dc4b

sql: ANSI aliases for LENGTH() · 2885cf84

Mergen Imeev authored 5 years ago

This patch creates aliases CHARACTER_LENGTH() and CHAR_LENGTH()
for LENGTH(). These functions are added because they are described
in ANSI.

Closes #3929

@TarantoolBot document
Title: SQL functions CHAR_LENGTH() and CHARACTER_LENGTH()

The SQL functions CHAR_LENGTH() and CHARACTER_LENGTH() work the
same as the LENGTH() function. They take exactly one argument. If
an argument of type TEXT or can be cast to a TEXT value using
internal casting rules, these functions return the length of the
TEXT value that represents the argument. They throw an error if
the argument cannot be cast to a TEXT value.

2885cf84

Jul 09, 2019

swim: optimize struct swim_task layout · 31a26448

Vladislav Shpilevoy authored 5 years ago

Before the patch it was split in two parts by 1.5KB packet, and
in the constructor it was nullifying the whole volume. Obviously,
these were mistakes. The first problem breaks cache locality,
the second one flushes the cache.

31a26448

swim: pool IO tasks · 837e114e

Vladislav Shpilevoy authored 5 years ago

Before the patch each SWIM member had two preallocated task
objects, 3KB in total. It was a waste of memory, because network
load per member in SWIM is ~2 messages per round step regardless
of cluster size.

This patch moves the tasks to a pool, where they can be reused.
Even by different SWIM instances running on the same node.

837e114e

test: fix net.box occasional failure. Again · eb0cc50c

Serge Petrenko authored 5 years ago

The test regarding logging corrupted rows failed occasionally with
```
[016]  test_run:grep_log('default', 'Got a corrupted row.*')
[016]  ---
[016] -- 'Got a corrupted row:'
[016] +- null
[016]  ...
```
The logs then had
```
[010] 2019-07-06 19:36:16.857 [13046] iproto sio.c:261 !> SystemError writev(1),
called on fd 23, aka unix/:(socket), peer of unix/:(socket): Broken pipe
```
instead of the expected message.

This happened, because we closed a socket before tarantool could write a
greeting to the client, the connection was then closed, and execution
never got to processing the malformed request and thus printing the
desired message to the log.

To fix this, actually read the greeting prior to writing new data and
closing the socket.

Follow-up #4273

eb0cc50c

fio: introduce utime function · 6e393aca

Oleg Babin authored 5 years ago

Closes #4323

@TarantoolBot document
Title: fio.utime

fio.utime (filepath [, atime [, mtime]])
Set access and modification times of a file.
The first argument is the filename, the second argument (atime) is
the access time, and the third argument (mtime) is
the modification time. Both times are provided in seconds since the epoch.
If the modification time is omitted, the access time provided is used;
if both times are omitted, the current time is used.

6e393aca

txn: run on_rollback triggers on txn_abort · 6ac597db

Vladimir Davydov authored 5 years ago

When a memtx transaction is aborted on yield, it isn't enough to
rollback individual statements - we must also run on_rollback triggers,
otherwise changes done to the schema by an aborted DDL transaction will
be visible to other fibers until an attempt to commit it is made.

6ac597db

test: net.box: fix case re invalid msgpack warning · 0f9fdd72

Alexander V. Tikhonov authored 5 years ago

The test case has two problems that appear from time to time and lead to
flaky fails. Those fails are look as shown below in a test-run output.

 | Test failed! Result content mismatch:
 | --- box/net.box.result	Mon Jun 24 17:23:49 2019
 | +++ box/net.box.reject	Mon Jun 24 17:51:52 2019
 | @@ -1404,7 +1404,7 @@
 |  ...
 |  test_run:grep_log('default', 'ER_INVALID_MSGPACK.*')
 | ---
 | -- 'ER_INVALID_MSGPACK: Invalid MsgPack - packet body'
 | +- 'ER_INVALID_MSGPACK: Invalid MsgPack - packet length'
 | ...
 | -- gh-983 selecting a lot of data crashes the server or hangs the
 | -- connection

'ER_INVALID_MSGPACK.*' regexp should match 'ER_INVALID_MSGPACK: Invalid
MsgPack - packet body' log message, but if it is not in a log file at a
time of grep_log() call (just don't flushed to the file yet) a message
produced by another test case can be matched ('ER_INVALID_MSGPACK:
Invalid MsgPack - packet length'). The fix here is to match the entire
message and check for the message periodically during several seconds
(use wait_log() instead of grep_log()).

Another problem is the race between writing a response to an iproto
socket on a server side and closing the socket on a client end. If
tarantool is unable to write a response, it does not produce the warning
re invalid msgpack, but shows 'broken pipe' message instead. We need
first grep for the message in logs and only then close the socket on a
client. The similar problem (with another test case) is described in
[1].

[1]: https://github.com/tarantool/tarantool/issues/4273#issuecomment-508939695

Closes: #4311

Unverified

0f9fdd72

Jul 08, 2019

txn: fix execution order of commit triggers · 01343264

Vladimir Davydov authored 5 years ago

Both commit and rollback triggers are currently added to the list head.
As a result, they are both run in the reverse order. This is correct for
rollback triggers, because this matches the order in which statements
that added the triggers are rolled back, but this is wrong for commit
triggers. For example, suppose we create a space and then create an
index for it in the same transaction. We expect that on success we first
run the trigger that commits the space and only then the trigger that
commits the index, not vice versa. That said, reverse the order of
commit triggers in the scope of preparations for transactional DDL.

01343264

vinyl: don't sync WAL on space alter if not necessary · 27aba00b

Vladimir Davydov authored 5 years ago

Changes done to an altered space while a new index is being built or
the format is being checked are propagated via an on_replace trigger.
The problem is there may be transactions that started before the alter
request. Their working set can't be checked so we simply abort them.
We can't abort transactions that have reached WAL so we also call
wal_sync() to flush all pending WAL requests. This is a yielding
operation and we call it even if there's no transactions that need
to be flushed. As a result, vinyl space alter yields unconditionally,
even if the space is empty and there is no pending transactions
affecting it. This prevents us from implementing transactional DDL.
Let's call wal_sync() only if there's actually at least one pending
transaction affecting the altered space and waiting for WAL.

27aba00b

decimal: expose decimal type to lua. · 3ab387a4

Serge Petrenko authored 5 years ago

Add a decimal library to lua.

Part of #692

@TarantoolBot document
Title: Document decimal module in lua.

First of all, you have to require the package via
`decimal = require('decimal')`
Now you can construct decimals via `new` method.
Decimals may be constructed from lua numbers, strings, unsigned and
signed 64 bit integers.
Decimal is a fixed-point type with maximum 38 digits of precision. All
the calculations are exact, so, be careful when constructing decimals
from lua numbers: they may hold only 15 decimal digits of precision.
You are advised to construct decimals from strings, since strings
represent decimals exactly, and vice versa.

```
a = decimal.new(123e-7)
b = decimal.new('123.456')
c = decimal.new('123.456e2')
d = decimal.new(123ULL)
e = decimal.new(2)
```
The allowed operations are addition, subtraction, division,
multiplication and power. If at least one of the operands is decimal,
decimal operations are performed. The other operand may be either
decimal or string, containing a number representation, or a lua number.

Operations only fail on an overflow, i.e. when result exceeds 10^38 - 1.
This includes division by zero. In these cases an error `Operation
failed` is raised.
Underflow is also possible, when precision needed to store the exact
result exceeds 38 digits. Underflow is not an error. When an underflow
happens, the result is rounded to 38 digits of precision.

```
a = decimal.new(123e-7)
b = decimal.new('123.456')
c = decimal.new('123.456e2')
d = decimal.new(123ULL)
e = decimal.new(2)
```
```
tarantool> a + b
---
- '123.456012300000000'
...

tarantool> c - d
---
- '12222.6'
...

tarantool> c / b
---
- '100'
...

tarantool> d * d
---
- '15129'
...

tarantool> d ^ 2
---
- '15129'
...

tarantool> 2 ^ d
---
- '10633823966279326983230456482242756608'...

tarantool> e ^ d
---
- '10633823966279326983230456482242756608'
...
```
The following math functions are also supported:
log10, ln, exp, sqrt. When specified as
`decimal.opname()`, operations may be performed on
strings and lua numbers.
```
f = decimal.new(100)

tarantool> decimal.log10(f)
---
- '2'
...

tarantool> decimal.sqrt(f)
---
- '10'
...

tarantool> e2 = decimal.exp(2)
---
...

tarantool> decimal.ln(e2)
---
- '2.0000000000000000000000000000000000000'
...

There are also `abs` and `tostring` methods, and an unary minus
operator, which are pretty self-explanatory.

```
tarantool> a = decimal.new('-5')
---
...

tarantool> a
---
- '-5'
...

tarantool> decimal.abs(a)
---
- '5'
...

tarantool> -a
---
- '5'
...

tostring(a)
---
- '-5'
...

```

`decimal.precision`, `decimal.scale` and `decimal.round` :
The first two methods return precision, i.e. decimal digits in
number representation, and scale, i.e. decimal digits after the decimal
point in the number representation.
`decimal.round` rounds the number to the given scale.
```
tarantool> a = decimal.new('123.456789')
---
...

tarantool> decimal.precision(a)
---
- 9
...

tarantool> decimal.scale(a)
---
- 6
...

tarantool> decimal.round(a, 4)
---
- '123.4568'
...

```

Comparsions: `>`, `<`, `>=`, `<=`, `==` are also legal and work as
expected. You may compare decimals with lua numbers or strings. In that
case comparsion will happen after the values are converted to decimal
type.

3ab387a4

lua/utils: add a function to register FFI metatypes. · 4ca39537

Serge Petrenko authored 5 years ago

A ffi metatype has a CTypeID, which can be used to push cdata of the
type on the lua stack, and has an associated metatable, automatically
applied to every created member of the type.
This allows the behavior similar to pushing userdata and assigning a
metatable to it.

Needed for #692

4ca39537

decimal: fix string formatting on construction from double · f64481de

Serge Petrenko authored 5 years ago

Use printf "%g" option instead of "%f" to trim traling zeros in such
cases:
decimal_from_double(1) -> '1.000000000000000' -> decimal_from_string()
Now it should be
decimal_from_double(1) -> '1' ->decimal_from_string()

Follow-up 6d62c6c1

f64481de

decimal: diallow infinity and NaN entirely. · db27d470

Serge Petrenko authored 5 years ago

While arithmetic operations do not return infinities or NaNs, it is
possbile to construct an invalid decimal value from strings 'Infinity',
'NaN' and similar. Some decimal mathematic functions may also result in
an infinity, say, ln(0) yields '-Infinity'.
So, add checks that the number is not a NaN or infinity after each
operation, so that the operation either returns an error, or a valid
finite decimal number.

Follow-up 6d62c6c1

db27d470

decimal: fix ln hang on values between ~ 0.9 and 1.1 · e0d4a5dc

Serge Petrenko authored 5 years ago

Turns out decNumberLn hangs when result is subnormal, according to the
current context settings. To fix this, reset minimal allowed exponent
to a smaller value during the ln operation and round the result afterwards.

Follow-up 6d62c6c1

e0d4a5dc

vinyl: fix vy_range_update_compaction_priority hang · 75dc3e64

Vladimir Davydov authored 5 years ago

Under certain circumstances vy_slice_new() may create an empty slice,
e.g. on range split:

   |------------------ Slice ---------------|
                         |---- Run -----|
                     +
                  split key
   |---- Slice 1 ----||------ Slice 2 ------|
         ^^^^^^^
          Empty

vy_range_update_compaction_priority() uses the size of the last slice in
a range as a base for LSM tree level sizing. If the slice size happens
to be 0, it will simply hang in an infinite loop. Fix this potential
hang by using 1 if the last slice size is 0.

75dc3e64

alter: add a comment explaining the purpose of AlterSpaceLock · 1c548274
Konstantin Osipov authored 5 years ago

1c548274

Jul 06, 2019

gitlab-ci: merge test and deploy stages · b37959be

Alexander V. Tikhonov authored 5 years ago

Current results have some tests with flaky results, which
blocks the deploy stage - decided to merge deploy stage into
test stage temporary to fix it.

Follows up #4156

Unverified

b37959be

Jul 05, 2019

swim: push-pull dissemination · 3fb2b875

Vladislav Shpilevoy authored 5 years ago

SWIM in the original paper says, that dissemination time of an
event is O(log(N)), where N is size of the cluster. It is true,
when both ping and ack messages carry dissemination and
anti-entropy. Before this patch it wasn't so - only regular
pings were carrying something.

After this patch the SWIM module has true exponential
dissemination speed.

Closes #4253

3fb2b875

swim: speed-up empty payloads cluster bootstrap · f30309de

Vladislav Shpilevoy authored 5 years ago

One another place consuming most of the tests start up time is
useless dissemination of an empty payload, which can be skipped
in fact.

Consider a cluster of 300 nodes. Each one of them are
interconnected manually, and now a test wants to wait for a
stabilization, when there are no events. On such a cluster it
happens for ~200 round steps till there are no any single event.

This is not about big packets, or log() TTD. There may be a few
events, may be more, but when a test wants the cluster to be
clean, it needs to wait for all the events being done.

This patch abuses the fact, that empty payloads can be compared
for free, no any single memcmp. If both new and the old payload
are empty, then nothing to disseminate.

It could help in a real cluster too, if initially there are no
payloads.

Needed for #4253

f30309de