[Tarantool-patches] [PATCH v1] test: block qsync_snapshots.test.lua

[Tarantool-patches] [PATCH v1] test: block qsync_snapshots.test.lua

[Alexander V. Tikhonov]

Found that after qsync_snapshots.test.lua test reproducer:

  test_run = require('test_run').new()
  engine = test_run:get_cfg('engine')

  box.schema.user.grant('guest', 'replication')
  test_run:cmd('create server replica with rpl_master=default,\
                                         script="replication/replica.lua"')
  test_run:cmd('start server replica with wait=True, wait_load=True')

  test_run:switch('default')
  _ = box.schema.space.create('sync', {is_sync=true, engine=engine})
  _ = box.space.sync:create_index('pk')
  box.space.sync:insert{1}
  box.snapshot()
  box.space.sync:drop()

  test_run:cmd('stop server replica')
  test_run:cmd('delete server replica')
  test_run:cleanup_cluster()
  box.schema.user.revoke('guest', 'replication')

the later running tests with commands, like:

  test_run:cmd("restart server default")

on the same test-run worker failed with output, like:

  main/103/master C> Tarantool 2.6.0-52-g71a24b9f2
  main/103/master C> log level 5
  main/103/master I> mapping 117440512 bytes for memtx tuple arena...
  main/103/master I> mapping 134217728 bytes for vinyl tuple arena...
  main/103/master I> instance uuid b086288b-5e2d-4696-b690-fd26222a6274
  main/103/master I> instance vclock {1: 13}
  iproto/101/main I> binary: bound to unix/:/Users/tntmac02.tarantool.i/tnt/test/var/001_replication/master.socket-iproto
  main/103/master I> recovery start
  main/103/master I> recovering from `/Users/tntmac02.tarantool.i/tnt/test/var/001_replication/master/00000000000000000007.snap'
  main/103/master I> cluster uuid 889d55d8-2cf1-4b62-966e-228fc7c86f58
  main/103/master I> assigned id 1 to replica b086288b-5e2d-4696-b690-fd26222a6274
  main/103/master I> assigned id 2 to replica 9c712376-d95d-4ed1-9eee-695353997c3c
  main/103/master I> recover from `/Users/tntmac02.tarantool.i/tnt/test/var/001_replication/master/00000000000000000007.xlog'
  main/103/master I> removed replica 9c712376-d95d-4ed1-9eee-695353997c3c
  main/103/master I> done `/Users/tntmac02.tarantool.i/tnt/test/var/001_replication/master/00000000000000000007.xlog'
  main/103/master I> recover from `/Users/tntmac02.tarantool.i/tnt/test/var/001_replication/master/00000000000000000013.xlog'
  main/103/master I> done `/Users/tntmac02.tarantool.i/tnt/test/var/001_replication/master/00000000000000000013.xlog'
  main/103/master cbus.c:442 !> SystemError timed out: Operation timed out
  main/103/master F> can't initialize storage: timed out
  main/103/master F> can't initialize storage: timed out

Found that this command used in many tests:

  find test/replication -name "*.test.lua" -exec grep "restart server default" {} \; | wc -l
  16

To save the testing stability the test qsync_snapshots.test.lua must
be disabled either run after all of the tests that use this command.
Decided to add this test to 'fragile' test-run list to set it run
at the very end of the replication suite tests run list.

Part of #5288
---

Github: https://github.com/tarantool/tarantool/tree/avtikhon/gh-5288-qsync-snapshots-block
Issue: https://github.com/tarantool/tarantool/issues/5288

 test/replication/suite.ini | 1 +
 1 file changed, 1 insertion(+)

diff --git a/test/replication/suite.ini b/test/replication/suite.ini
index a6d653d3b..3baa3af47 100644
--- a/test/replication/suite.ini
+++ b/test/replication/suite.ini
@@ -24,3 +24,4 @@ fragile = errinj.test.lua            ; gh-3870
           gh-4605-empty-password.test.lua         ; gh-5030
           anon.test.lua              ; gh-5058
           status.test.lua            ; gh-5110
+          qsync_snapshots.test.lua   ; gh-5288
-- 
2.17.1

Re: [Tarantool-patches] [PATCH v1] test: block qsync_snapshots.test.lua

[Vladislav Shpilevoy]

I think this patch can be dropped. The problem is
properly fixed in another pending patch I sent
recently.