test: flaky box/cfg.test.lua test

[avtikhon]

Tarantool version:
Tarantool 2.7.0-48-gbdd5d5e9ef
Target: Linux-x86_64-RelWithDebInfo
Build options: cmake . -DCMAKE_INSTALL_PREFIX=/usr/local -DENABLE_BACKTRACE=ON
Compiler: /usr/bin/clang-11 /usr/bin/clang++-11
C_FLAGS: -Wno-unknown-pragmas -fexceptions -funwind-tables -fno-omit-frame-pointer -fno-stack-protector -fno-common -msse2 -fsanitize=address -fsanitize-blacklist=/builds/4BUsapPU/0/tarantool/tarantool/asan/asan.supp -std=c11 -Wall -Wextra -Wno-strict-aliasing -fsanitize=alignment,bool,bounds,builtin,enum,float-cast-overflow,float-divide-by-zero,function,integer-divide-by-zero,return,shift,unreachable,vla-bound -fno-sanitize-recover=alignment,bool,bounds,builtin,enum,float-cast-overflow,float-divide-by-zero,function,integer-divide-by-zero,return,shift,unreachable,vla-bound -Wno-char-subscripts -Wno-gnu-alignof-expression -Werror
CXX_FLAGS: -Wno-unknown-pragmas -fexceptions -funwind-tables -fno-omit-frame-pointer -fno-stack-protector -fno-common -msse2 -fsanitize=address -fsanitize-blacklist=/builds/4BUsapPU/0/tarantool/tarantool/asan/asan.supp -std=c++11 -Wall -Wextra -Wno-strict-aliasing -fsanitize=alignment,bool,bounds,builtin,enum,float-cast-overflow,float-divide-by-zero,function,integer-divide-by-zero,return,shift,unreachable,vla-bound -fno-sanitize-recover=alignment,bool,bounds,builtin,enum,float-cast-overflow,float-divide-by-zero,function,integer-divide-by-zero,return,shift,unreachable,vla-bound -Wno-char-subscripts -Wno-invalid-offsetof -Wno-gnu-alignof-expression -Werror

OS version:
Debian 10

Bug description:
https://gitlab.com/tarantool/tarantool/-/jobs/852378142#L4791
https://gitlab.com/tarantool/tarantool/-/jobs/852964472#L4296

artifacts.zip

results file checksum: 6562ddce4c4d8d42e2a81c7e452434b8

[007] box/cfg.test.lua                                                
[007] [Instance "cfg_test6"] Tarantool server failed to start
[007] 
[007] Last 15 lines of Tarantool Log file [Instance "cfg_test6"][/builds/4BUsapPU/0/tarantool/tarantool/test/var/007_box/cfg_test6.log]:
[007] Starting instance cfg_test6...
[007] Run console at unix/:/builds/4BUsapPU/0/tarantool/tarantool/test/var/007_box/cfg_test6.control
[007] started
[007] mkdir /builds/4BUsapPU/0/tarantool/tarantool/test/var/007_box/cfg_test6
[007] 2020-11-16 18:32:07.677 [18685] main/103/cfg_test6 C> Tarantool 2.7.0-48-gbdd5d5e9ef
[007] 2020-11-16 18:32:07.677 [18685] main/103/cfg_test6 C> log level 5
[007] 2020-11-16 18:32:07.678 [18685] main/103/cfg_test6 I> mapping 268435456 bytes for memtx tuple arena...
[007] 2020-11-16 18:32:07.678 [18685] main/103/cfg_test6 I> mapping 134217728 bytes for vinyl tuple arena...
[007] 2020-11-16 18:32:07.695 [18685] main/103/cfg_test6 I> instance uuid 8f8b0655-39ce-40de-ac43-9f426c68b664
[007] 2020-11-16 18:32:08.006 [18685] iproto/101/main I> binary: bound to unix/:(socket)
[007] 2020-11-16 18:32:08.007 [18685] main/103/cfg_test6 I> connecting to 1 replicas
[007] 2020-11-16 18:32:08.198 [18685] main/103/cfg_test6 C> failed to connect to 1 out of 1 replicas
[007] 2020-11-16 18:32:08.199 [18685] main/103/cfg_test6 replication.cc:744 E> ER_CFG: Incorrect value for option 'replication': failed to connect to one or more replicas
[007] 2020-11-16 18:32:08.199 [18685] main/103/cfg_test6 F> can't initialize storage: Incorrect value for option 'replication': failed to connect to one or more replicas
[007] 2020-11-16 18:32:08.199 [18685] main/103/cfg_test6 F> can't initialize storage: Incorrect value for option 'replication': failed to connect to one or more replicas
[007] 
[007] [Instance "box"] Failed to start tarantool instance "cfg_test6"
[007] [ fail ]

Steps to reproduce:

Optional (but very desirable):

• coredump
• backtrace
• netstat

[Totktonada]

cfg_test6.lua:

#!/usr/bin/env tarantool
os = require('os')

box.cfg{
    listen = os.getenv("LISTEN"),
    replication         = "admin:test-cluster-cookie@" .. os.getenv("LISTEN"),
    replication_connect_timeout = 0.1,
}

require('console').listen(os.getenv('ADMIN'))

It seems, something prevents connecting to itself during replication_connect_timeout. It seems, lifting the time budget will not break the test logic. Can you verify, whether it'll help with the flaky fail?