[
https://issues.apache.org/jira/browse/IGNITE-23321?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Igor updated IGNITE-23321:
--------------------------
Description:
h3. Steps to reproduce:
Run the next queries in cycle with 50 repeats in single connection:
{code:java}
drop table if exists dropNoMoreIndexedColumn
create zone if not exists "rocksdb" engine rocksdb
create table dropNoMoreIndexedColumn(k1 TIMESTAMP not null, k2 INTEGER not
null, v0 TINYINT not null, v1 SMALLINT not null, v2 INTEGER not null, v3 BIGINT
not null, v4 VARCHAR not null, v5 TIMESTAMP not null, primary key (k1, k2))
with PRIMARY_ZONE='rocksdb'
create index dropNoMoreIndexedColumn_v1idx on dropNoMoreIndexedColumn using
TREE (v1)
drop index dropNoMoreIndexedColumn_v1idx
alter table dropNoMoreIndexedColumn drop column v1 {code}
h3. Expected:
All queries are executed.
h3. Actual:
On repeat 31 the client freeze for infinite amount of time.
h3. Analysis:
The servers contain repeated error in logs:
{code:java}
2024-09-30 02:58:36:290 +0000
[WARNING][%DropColumnsTest_cluster_2%JRaft-FSMCaller-Disruptormetastorage_stripe_0-0][LeaseUpdater]
Lease update invocation failed because of concurrent update.
2024-09-30 02:58:36:809 +0000
[WARNING][%DropColumnsTest_cluster_2%JRaft-FSMCaller-Disruptormetastorage_stripe_0-0][LeaseUpdater]
Lease update invocation failed because of concurrent update.
2024-09-30 02:58:37:045 +0000 [INFO][Thread-259][LoggingRocksDbFlushListener]
Starting rocksdb flush process [name='meta-storage log',
reason=WRITE_BUFFER_FULL]
2024-09-30 02:58:37:332 +0000
[WARNING][%DropColumnsTest_cluster_2%JRaft-FSMCaller-Disruptormetastorage_stripe_0-0][LeaseUpdater]
Lease update invocation failed because of concurrent update.
2024-09-30 02:58:38:178 +0000
[WARNING][%DropColumnsTest_cluster_2%JRaft-FSMCaller-Disruptormetastorage_stripe_0-0][LeaseUpdater]
Lease update invocation failed because of concurrent update.
2024-09-30 02:58:38:743 +0000
[WARNING][%DropColumnsTest_cluster_2%JRaft-FSMCaller-Disruptormetastorage_stripe_0-0][LeaseUpdater]
Lease update invocation failed because of concurrent update.
2024-09-30 02:58:39:047 +0000
[WARNING][%DropColumnsTest_cluster_2%JRaft-FSMCaller-Disruptormetastorage_stripe_0-0][LeaseUpdater]
Lease update invocation failed because of concurrent update.{code}
[^server_logs.zip]
was:
h3. Steps to reproduce:
Run the next queries in cycle with 50 repeats in single connection:
{code:java}
drop table if exists dropNoMoreIndexedColumn
create zone if not exists "rocksdb" engine rocksdb
create table dropNoMoreIndexedColumn(k1 TIMESTAMP not null, k2 INTEGER not
null, v0 TINYINT not null, v1 SMALLINT not null, v2 INTEGER not null, v3 BIGINT
not null, v4 VARCHAR not null, v5 TIMESTAMP not null, primary key (k1, k2))
with PRIMARY_ZONE='rocksdb'
create index dropNoMoreIndexedColumn_v1idx on dropNoMoreIndexedColumn using
TREE (v1)
drop index dropNoMoreIndexedColumn_v1idx
alter table dropNoMoreIndexedColumn drop column v1 {code}
h3. Expected:
All queries are executed.
h3. Actual:
On repeat 31 the client freeze for infinite amount of time.
h3. Analysis:
With storages aimem and aipersist the issue is not happen.
The servers contain repeated error in logs:
{code:java}
2024-09-30 02:58:36:290 +0000
[WARNING][%DropColumnsTest_cluster_2%JRaft-FSMCaller-Disruptormetastorage_stripe_0-0][LeaseUpdater]
Lease update invocation failed because of concurrent update.
2024-09-30 02:58:36:809 +0000
[WARNING][%DropColumnsTest_cluster_2%JRaft-FSMCaller-Disruptormetastorage_stripe_0-0][LeaseUpdater]
Lease update invocation failed because of concurrent update.
2024-09-30 02:58:37:045 +0000 [INFO][Thread-259][LoggingRocksDbFlushListener]
Starting rocksdb flush process [name='meta-storage log',
reason=WRITE_BUFFER_FULL]
2024-09-30 02:58:37:332 +0000
[WARNING][%DropColumnsTest_cluster_2%JRaft-FSMCaller-Disruptormetastorage_stripe_0-0][LeaseUpdater]
Lease update invocation failed because of concurrent update.
2024-09-30 02:58:38:178 +0000
[WARNING][%DropColumnsTest_cluster_2%JRaft-FSMCaller-Disruptormetastorage_stripe_0-0][LeaseUpdater]
Lease update invocation failed because of concurrent update.
2024-09-30 02:58:38:743 +0000
[WARNING][%DropColumnsTest_cluster_2%JRaft-FSMCaller-Disruptormetastorage_stripe_0-0][LeaseUpdater]
Lease update invocation failed because of concurrent update.
2024-09-30 02:58:39:047 +0000
[WARNING][%DropColumnsTest_cluster_2%JRaft-FSMCaller-Disruptormetastorage_stripe_0-0][LeaseUpdater]
Lease update invocation failed because of concurrent update.{code}
[^server_logs.zip]
> RocksDB: repeat of create table and drop column leads to freeze of client
> -------------------------------------------------------------------------
>
> Key: IGNITE-23321
> URL: https://issues.apache.org/jira/browse/IGNITE-23321
> Project: Ignite
> Issue Type: Bug
> Components: persistence
> Affects Versions: 3.0.0-beta1
> Environment: 3 nodes cluster running locally.
> Reporter: Igor
> Priority: Major
> Labels: ignite-3
> Attachments: server_logs.zip
>
>
> h3. Steps to reproduce:
> Run the next queries in cycle with 50 repeats in single connection:
> {code:java}
> drop table if exists dropNoMoreIndexedColumn
> create zone if not exists "rocksdb" engine rocksdb
> create table dropNoMoreIndexedColumn(k1 TIMESTAMP not null, k2 INTEGER not
> null, v0 TINYINT not null, v1 SMALLINT not null, v2 INTEGER not null, v3
> BIGINT not null, v4 VARCHAR not null, v5 TIMESTAMP not null, primary key (k1,
> k2)) with PRIMARY_ZONE='rocksdb'
> create index dropNoMoreIndexedColumn_v1idx on dropNoMoreIndexedColumn using
> TREE (v1)
> drop index dropNoMoreIndexedColumn_v1idx
> alter table dropNoMoreIndexedColumn drop column v1 {code}
> h3. Expected:
> All queries are executed.
> h3. Actual:
> On repeat 31 the client freeze for infinite amount of time.
> h3. Analysis:
> The servers contain repeated error in logs:
> {code:java}
> 2024-09-30 02:58:36:290 +0000
> [WARNING][%DropColumnsTest_cluster_2%JRaft-FSMCaller-Disruptormetastorage_stripe_0-0][LeaseUpdater]
> Lease update invocation failed because of concurrent update.
> 2024-09-30 02:58:36:809 +0000
> [WARNING][%DropColumnsTest_cluster_2%JRaft-FSMCaller-Disruptormetastorage_stripe_0-0][LeaseUpdater]
> Lease update invocation failed because of concurrent update.
> 2024-09-30 02:58:37:045 +0000 [INFO][Thread-259][LoggingRocksDbFlushListener]
> Starting rocksdb flush process [name='meta-storage log',
> reason=WRITE_BUFFER_FULL]
> 2024-09-30 02:58:37:332 +0000
> [WARNING][%DropColumnsTest_cluster_2%JRaft-FSMCaller-Disruptormetastorage_stripe_0-0][LeaseUpdater]
> Lease update invocation failed because of concurrent update.
> 2024-09-30 02:58:38:178 +0000
> [WARNING][%DropColumnsTest_cluster_2%JRaft-FSMCaller-Disruptormetastorage_stripe_0-0][LeaseUpdater]
> Lease update invocation failed because of concurrent update.
> 2024-09-30 02:58:38:743 +0000
> [WARNING][%DropColumnsTest_cluster_2%JRaft-FSMCaller-Disruptormetastorage_stripe_0-0][LeaseUpdater]
> Lease update invocation failed because of concurrent update.
> 2024-09-30 02:58:39:047 +0000
> [WARNING][%DropColumnsTest_cluster_2%JRaft-FSMCaller-Disruptormetastorage_stripe_0-0][LeaseUpdater]
> Lease update invocation failed because of concurrent update.{code}
> [^server_logs.zip]
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
