[jira] [Commented] (IGNITE-6804) Print a warning if HashMap is passed into bulk update operations
[ https://issues.apache.org/jira/browse/IGNITE-6804?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17034404#comment-17034404 ] Andrey N. Gura commented on IGNITE-6804: [~ilyak] It us unimportant after removal of deadlock detection related logic. One cosmetic comment: {{Map}} and {{Collection}} parameters in {{warnIfUnordered}} should be declared as {{Map}} and {{Collection}} respectively. LGTM. Please proceed with merge. > Print a warning if HashMap is passed into bulk update operations > > > Key: IGNITE-6804 > URL: https://issues.apache.org/jira/browse/IGNITE-6804 > Project: Ignite > Issue Type: Improvement > Components: cache >Reporter: Denis A. Magda >Assignee: Ilya Kasnacheev >Priority: Critical > Labels: usability > Time Spent: 5h 10m > Remaining Estimate: 0h > > Ignite newcomers tend to stumble on deadlocks simply because the keys are > passed in an unordered HashMap. Propose to do the following: > * update bulk operations Java docs. > * print out a warning if not SortedMap (e.g. HashMap, > Weak/Identity/Concurrent/Linked HashMap etc) is passed into > a bulk method (instead of SortedMap) and contains more than 1 element. > However, we should make sure that we only print that warning once and not > every time the API is called. > * do not produce warning for explicit optimistic transactions > More details are here: > http://apache-ignite-developers.2346864.n4.nabble.com/Re-Ignite-2-0-0-GridUnsafe-unmonitor-td23706.html -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (IGNITE-6804) Print a warning if HashMap is passed into bulk update operations
[ https://issues.apache.org/jira/browse/IGNITE-6804?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17033794#comment-17033794 ] Andrey N. Gura commented on IGNITE-6804: [~ilyak] It seems you missed the following comment: {quote} @alamar It seems we call checkKeysOrdered before syncOp method. But implicit transactions always start in syncOp method. So curTxDeadlockDetecting will always return false for any implicit transaction. {quote} > Print a warning if HashMap is passed into bulk update operations > > > Key: IGNITE-6804 > URL: https://issues.apache.org/jira/browse/IGNITE-6804 > Project: Ignite > Issue Type: Improvement > Components: cache >Reporter: Denis A. Magda >Assignee: Ilya Kasnacheev >Priority: Critical > Labels: usability > Time Spent: 5h 10m > Remaining Estimate: 0h > > Ignite newcomers tend to stumble on deadlocks simply because the keys are > passed in an unordered HashMap. Propose to do the following: > * update bulk operations Java docs. > * print out a warning if not SortedMap (e.g. HashMap, > Weak/Identity/Concurrent/Linked HashMap etc) is passed into > a bulk method (instead of SortedMap) and contains more than 1 element. > However, we should make sure that we only print that warning once and not > every time the API is called. > * do not produce warning for explicit optimistic transactions > More details are here: > http://apache-ignite-developers.2346864.n4.nabble.com/Re-Ignite-2-0-0-GridUnsafe-unmonitor-td23706.html -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (IGNITE-6804) Print a warning if HashMap is passed into bulk update operations
[ https://issues.apache.org/jira/browse/IGNITE-6804?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17032271#comment-17032271 ] Ignite TC Bot commented on IGNITE-6804: --- {panel:title=Branch: [pull/6976/head] Base: [master] : No blockers found!|borderStyle=dashed|borderColor=#ccc|titleBGColor=#D6F7C1}{panel} [TeamCity *-- Run :: All* Results|https://ci.ignite.apache.org/viewLog.html?buildId=4990981buildTypeId=IgniteTests24Java8_RunAll] > Print a warning if HashMap is passed into bulk update operations > > > Key: IGNITE-6804 > URL: https://issues.apache.org/jira/browse/IGNITE-6804 > Project: Ignite > Issue Type: Improvement > Components: cache >Reporter: Denis A. Magda >Assignee: Ilya Kasnacheev >Priority: Critical > Labels: usability > Time Spent: 5h 10m > Remaining Estimate: 0h > > Ignite newcomers tend to stumble on deadlocks simply because the keys are > passed in an unordered HashMap. Propose to do the following: > * update bulk operations Java docs. > * print out a warning if not SortedMap (e.g. HashMap, > Weak/Identity/Concurrent/Linked HashMap etc) is passed into > a bulk method (instead of SortedMap) and contains more than 1 element. > However, we should make sure that we only print that warning once and not > every time the API is called. > * do not produce warning for explicit optimistic transactions > More details are here: > http://apache-ignite-developers.2346864.n4.nabble.com/Re-Ignite-2-0-0-GridUnsafe-unmonitor-td23706.html -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (IGNITE-6804) Print a warning if HashMap is passed into bulk update operations
[ https://issues.apache.org/jira/browse/IGNITE-6804?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17031707#comment-17031707 ] Ilya Kasnacheev commented on IGNITE-6804: - [~agura] please check amended fix (you can just look at the last commit) - logic is simplified, extra method removed, it should now be very obvious. > Print a warning if HashMap is passed into bulk update operations > > > Key: IGNITE-6804 > URL: https://issues.apache.org/jira/browse/IGNITE-6804 > Project: Ignite > Issue Type: Improvement > Components: cache >Reporter: Denis A. Magda >Assignee: Ilya Kasnacheev >Priority: Critical > Labels: usability > Time Spent: 5h 10m > Remaining Estimate: 0h > > Ignite newcomers tend to stumble on deadlocks simply because the keys are > passed in an unordered HashMap. Propose to do the following: > * update bulk operations Java docs. > * print out a warning if not SortedMap (e.g. HashMap, > Weak/Identity/Concurrent/Linked HashMap etc) is passed into > a bulk method (instead of SortedMap) and contains more than 1 element. > However, we should make sure that we only print that warning once and not > every time the API is called. > * do not produce warning for explicit optimistic transactions > More details are here: > http://apache-ignite-developers.2346864.n4.nabble.com/Re-Ignite-2-0-0-GridUnsafe-unmonitor-td23706.html -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (IGNITE-6804) Print a warning if HashMap is passed into bulk update operations
[ https://issues.apache.org/jira/browse/IGNITE-6804?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17031613#comment-17031613 ] Andrey N. Gura commented on IGNITE-6804: [~ilyak] Ok. I think the logic is too overwhelmed. Moreover, it doesn't matter whether deadlock detection enabled or not. Unsafe API usage is still could lead to deadlocks and performance degradation. So we can simplify conditions under which warning about deadlocks possibility tested. Let's just remove this check. > Print a warning if HashMap is passed into bulk update operations > > > Key: IGNITE-6804 > URL: https://issues.apache.org/jira/browse/IGNITE-6804 > Project: Ignite > Issue Type: Improvement > Components: cache >Reporter: Denis A. Magda >Assignee: Ilya Kasnacheev >Priority: Critical > Labels: usability > Time Spent: 5h 10m > Remaining Estimate: 0h > > Ignite newcomers tend to stumble on deadlocks simply because the keys are > passed in an unordered HashMap. Propose to do the following: > * update bulk operations Java docs. > * print out a warning if not SortedMap (e.g. HashMap, > Weak/Identity/Concurrent/Linked HashMap etc) is passed into > a bulk method (instead of SortedMap) and contains more than 1 element. > However, we should make sure that we only print that warning once and not > every time the API is called. > * do not produce warning for explicit optimistic transactions > More details are here: > http://apache-ignite-developers.2346864.n4.nabble.com/Re-Ignite-2-0-0-GridUnsafe-unmonitor-td23706.html -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (IGNITE-6804) Print a warning if HashMap is passed into bulk update operations
[ https://issues.apache.org/jira/browse/IGNITE-6804?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17031503#comment-17031503 ] Ilya Kasnacheev commented on IGNITE-6804: - [~agura] I have just checked this use case, warning was issued. Are we good now? > Print a warning if HashMap is passed into bulk update operations > > > Key: IGNITE-6804 > URL: https://issues.apache.org/jira/browse/IGNITE-6804 > Project: Ignite > Issue Type: Improvement > Components: cache >Reporter: Denis A. Magda >Assignee: Ilya Kasnacheev >Priority: Critical > Labels: usability > Time Spent: 5h 10m > Remaining Estimate: 0h > > Ignite newcomers tend to stumble on deadlocks simply because the keys are > passed in an unordered HashMap. Propose to do the following: > * update bulk operations Java docs. > * print out a warning if not SortedMap (e.g. HashMap, > Weak/Identity/Concurrent/Linked HashMap etc) is passed into > a bulk method (instead of SortedMap) and contains more than 1 element. > However, we should make sure that we only print that warning once and not > every time the API is called. > * do not produce warning for explicit optimistic transactions > More details are here: > http://apache-ignite-developers.2346864.n4.nabble.com/Re-Ignite-2-0-0-GridUnsafe-unmonitor-td23706.html -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (IGNITE-6804) Print a warning if HashMap is passed into bulk update operations
[ https://issues.apache.org/jira/browse/IGNITE-6804?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17030778#comment-17030778 ] Andrey N. Gura commented on IGNITE-6804: [~ilyak] Two concurrent {{invokeAll()}} on transactional cache will start optimistic repeatable read transactions implicitly and they could be deadly locked. > Print a warning if HashMap is passed into bulk update operations > > > Key: IGNITE-6804 > URL: https://issues.apache.org/jira/browse/IGNITE-6804 > Project: Ignite > Issue Type: Improvement > Components: cache >Reporter: Denis A. Magda >Assignee: Ilya Kasnacheev >Priority: Critical > Labels: usability > Time Spent: 5h 10m > Remaining Estimate: 0h > > Ignite newcomers tend to stumble on deadlocks simply because the keys are > passed in an unordered HashMap. Propose to do the following: > * update bulk operations Java docs. > * print out a warning if not SortedMap (e.g. HashMap, > Weak/Identity/Concurrent/Linked HashMap etc) is passed into > a bulk method (instead of SortedMap) and contains more than 1 element. > However, we should make sure that we only print that warning once and not > every time the API is called. > * do not produce warning for explicit optimistic transactions > More details are here: > http://apache-ignite-developers.2346864.n4.nabble.com/Re-Ignite-2-0-0-GridUnsafe-unmonitor-td23706.html -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (IGNITE-6804) Print a warning if HashMap is passed into bulk update operations
[ https://issues.apache.org/jira/browse/IGNITE-6804?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17029852#comment-17029852 ] Ilya Kasnacheev commented on IGNITE-6804: - [~agura] I have fixed on most of your comments. I don't think that `implicit' is a problem, this just means we will consider any implicit transaction non-detecting, will issue a warning. Can you show examples where this is not correct? > Print a warning if HashMap is passed into bulk update operations > > > Key: IGNITE-6804 > URL: https://issues.apache.org/jira/browse/IGNITE-6804 > Project: Ignite > Issue Type: Improvement > Components: cache >Reporter: Denis A. Magda >Assignee: Ilya Kasnacheev >Priority: Critical > Labels: usability > Time Spent: 5h 10m > Remaining Estimate: 0h > > Ignite newcomers tend to stumble on deadlocks simply because the keys are > passed in an unordered HashMap. Propose to do the following: > * update bulk operations Java docs. > * print out a warning if not SortedMap (e.g. HashMap, > Weak/Identity/Concurrent/Linked HashMap etc) is passed into > a bulk method (instead of SortedMap) and contains more than 1 element. > However, we should make sure that we only print that warning once and not > every time the API is called. > * do not produce warning for explicit optimistic transactions > More details are here: > http://apache-ignite-developers.2346864.n4.nabble.com/Re-Ignite-2-0-0-GridUnsafe-unmonitor-td23706.html -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (IGNITE-6804) Print a warning if HashMap is passed into bulk update operations
[ https://issues.apache.org/jira/browse/IGNITE-6804?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17028907#comment-17028907 ] Ilya Kasnacheev commented on IGNITE-6804: - [~agura] I have fixed the remaining javadocs, please review. > Print a warning if HashMap is passed into bulk update operations > > > Key: IGNITE-6804 > URL: https://issues.apache.org/jira/browse/IGNITE-6804 > Project: Ignite > Issue Type: Improvement > Components: cache >Reporter: Denis A. Magda >Assignee: Ilya Kasnacheev >Priority: Critical > Labels: usability > Time Spent: 3h 20m > Remaining Estimate: 0h > > Ignite newcomers tend to stumble on deadlocks simply because the keys are > passed in an unordered HashMap. Propose to do the following: > * update bulk operations Java docs. > * print out a warning if not SortedMap (e.g. HashMap, > Weak/Identity/Concurrent/Linked HashMap etc) is passed into > a bulk method (instead of SortedMap) and contains more than 1 element. > However, we should make sure that we only print that warning once and not > every time the API is called. > * do not produce warning for explicit optimistic transactions > More details are here: > http://apache-ignite-developers.2346864.n4.nabble.com/Re-Ignite-2-0-0-GridUnsafe-unmonitor-td23706.html -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (IGNITE-6804) Print a warning if HashMap is passed into bulk update operations
[ https://issues.apache.org/jira/browse/IGNITE-6804?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17027412#comment-17027412 ] Ignite TC Bot commented on IGNITE-6804: --- {panel:title=Branch: [pull/6976/head] Base: [master] : No blockers found!|borderStyle=dashed|borderColor=#ccc|titleBGColor=#D6F7C1}{panel} [TeamCity *-- Run :: All* Results|https://ci.ignite.apache.org/viewLog.html?buildId=4973988buildTypeId=IgniteTests24Java8_RunAll] > Print a warning if HashMap is passed into bulk update operations > > > Key: IGNITE-6804 > URL: https://issues.apache.org/jira/browse/IGNITE-6804 > Project: Ignite > Issue Type: Improvement > Components: cache >Reporter: Denis A. Magda >Assignee: Ilya Kasnacheev >Priority: Critical > Labels: usability > Time Spent: 3h 20m > Remaining Estimate: 0h > > Ignite newcomers tend to stumble on deadlocks simply because the keys are > passed in an unordered HashMap. Propose to do the following: > * update bulk operations Java docs. > * print out a warning if not SortedMap (e.g. HashMap, > Weak/Identity/Concurrent/Linked HashMap etc) is passed into > a bulk method (instead of SortedMap) and contains more than 1 element. > However, we should make sure that we only print that warning once and not > every time the API is called. > * do not produce warning for explicit optimistic transactions > More details are here: > http://apache-ignite-developers.2346864.n4.nabble.com/Re-Ignite-2-0-0-GridUnsafe-unmonitor-td23706.html -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (IGNITE-6804) Print a warning if HashMap is passed into bulk update operations
[ https://issues.apache.org/jira/browse/IGNITE-6804?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17026779#comment-17026779 ] Ignite TC Bot commented on IGNITE-6804: --- {panel:title=Branch: [pull/6976/head] Base: [master] : Possible Blockers (3)|borderStyle=dashed|borderColor=#ccc|titleBGColor=#F7D6C1} {color:#d04437}Queries 1{color} [[tests 1|https://ci.ignite.apache.org/viewLog.html?buildId=4973964]] * IgniteBinaryCacheQueryTestSuite: IgniteSqlSplitterSelfTest.testReplicatedTablesUsingPartitionedCacheClient - Test has low fail rate in base branch 0,0% and is not flaky {color:#d04437}Platform .NET (Core Linux){color} [[tests 1|https://ci.ignite.apache.org/viewLog.html?buildId=4973959]] * dll: CacheTest.TestCacheWithExpiryPolicyOnUpdate - Test has low fail rate in base branch 0,0% and is not flaky {color:#d04437}Platform .NET (Inspections)*{color} [[tests 0 TIMEOUT , TC_BUILD_FAILURE |https://ci.ignite.apache.org/viewLog.html?buildId=4973960]] {panel} [TeamCity *-- Run :: All* Results|https://ci.ignite.apache.org/viewLog.html?buildId=4973988buildTypeId=IgniteTests24Java8_RunAll] > Print a warning if HashMap is passed into bulk update operations > > > Key: IGNITE-6804 > URL: https://issues.apache.org/jira/browse/IGNITE-6804 > Project: Ignite > Issue Type: Improvement > Components: cache >Reporter: Denis A. Magda >Assignee: Ilya Kasnacheev >Priority: Critical > Labels: usability > Time Spent: 3h 20m > Remaining Estimate: 0h > > Ignite newcomers tend to stumble on deadlocks simply because the keys are > passed in an unordered HashMap. Propose to do the following: > * update bulk operations Java docs. > * print out a warning if not SortedMap (e.g. HashMap, > Weak/Identity/Concurrent/Linked HashMap etc) is passed into > a bulk method (instead of SortedMap) and contains more than 1 element. > However, we should make sure that we only print that warning once and not > every time the API is called. > * do not produce warning for explicit optimistic transactions > More details are here: > http://apache-ignite-developers.2346864.n4.nabble.com/Re-Ignite-2-0-0-GridUnsafe-unmonitor-td23706.html -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (IGNITE-6804) Print a warning if HashMap is passed into bulk update operations
[ https://issues.apache.org/jira/browse/IGNITE-6804?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17026662#comment-17026662 ] Ilya Kasnacheev commented on IGNITE-6804: - [~agura] I have also added detection of transactional getAll(). Please review amended fix. Tests are running, will be checked before merge. > Print a warning if HashMap is passed into bulk update operations > > > Key: IGNITE-6804 > URL: https://issues.apache.org/jira/browse/IGNITE-6804 > Project: Ignite > Issue Type: Improvement > Components: cache >Reporter: Denis A. Magda >Assignee: Ilya Kasnacheev >Priority: Critical > Labels: usability > Time Spent: 3h 20m > Remaining Estimate: 0h > > Ignite newcomers tend to stumble on deadlocks simply because the keys are > passed in an unordered HashMap. Propose to do the following: > * update bulk operations Java docs. > * print out a warning if not SortedMap (e.g. HashMap, > Weak/Identity/Concurrent/Linked HashMap etc) is passed into > a bulk method (instead of SortedMap) and contains more than 1 element. > However, we should make sure that we only print that warning once and not > every time the API is called. > * do not produce warning for explicit optimistic transactions > More details are here: > http://apache-ignite-developers.2346864.n4.nabble.com/Re-Ignite-2-0-0-GridUnsafe-unmonitor-td23706.html -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (IGNITE-6804) Print a warning if HashMap is passed into bulk update operations
[ https://issues.apache.org/jira/browse/IGNITE-6804?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17025915#comment-17025915 ] Andrey N. Gura commented on IGNITE-6804: [~ilyak] Please take a look on my comments in provided PR. > Print a warning if HashMap is passed into bulk update operations > > > Key: IGNITE-6804 > URL: https://issues.apache.org/jira/browse/IGNITE-6804 > Project: Ignite > Issue Type: Improvement > Components: cache >Reporter: Denis A. Magda >Assignee: Ilya Kasnacheev >Priority: Critical > Labels: usability > Time Spent: 2h 10m > Remaining Estimate: 0h > > Ignite newcomers tend to stumble on deadlocks simply because the keys are > passed in an unordered HashMap. Propose to do the following: > * update bulk operations Java docs. > * print out a warning if not SortedMap (e.g. HashMap, > Weak/Identity/Concurrent/Linked HashMap etc) is passed into > a bulk method (instead of SortedMap) and contains more than 1 element. > However, we should make sure that we only print that warning once and not > every time the API is called. > * do not produce warning for explicit optimistic transactions > More details are here: > http://apache-ignite-developers.2346864.n4.nabble.com/Re-Ignite-2-0-0-GridUnsafe-unmonitor-td23706.html -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (IGNITE-6804) Print a warning if HashMap is passed into bulk update operations
[ https://issues.apache.org/jira/browse/IGNITE-6804?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17021237#comment-17021237 ] Ignite TC Bot commented on IGNITE-6804: --- {panel:title=Branch: [pull/6976/head] Base: [master] : No blockers found!|borderStyle=dashed|borderColor=#ccc|titleBGColor=#D6F7C1}{panel} [TeamCity *-- Run :: All* Results|https://ci.ignite.apache.org/viewLog.html?buildId=4947337buildTypeId=IgniteTests24Java8_RunAll] > Print a warning if HashMap is passed into bulk update operations > > > Key: IGNITE-6804 > URL: https://issues.apache.org/jira/browse/IGNITE-6804 > Project: Ignite > Issue Type: Improvement > Components: cache >Reporter: Denis A. Magda >Assignee: Ilya Kasnacheev >Priority: Critical > Labels: usability > Time Spent: 50m > Remaining Estimate: 0h > > Ignite newcomers tend to stumble on deadlocks simply because the keys are > passed in an unordered HashMap. Propose to do the following: > * update bulk operations Java docs. > * print out a warning if not SortedMap (e.g. HashMap, > Weak/Identity/Concurrent/Linked HashMap etc) is passed into > a bulk method (instead of SortedMap) and contains more than 1 element. > However, we should make sure that we only print that warning once and not > every time the API is called. > * do not produce warning for explicit optimistic transactions > More details are here: > http://apache-ignite-developers.2346864.n4.nabble.com/Re-Ignite-2-0-0-GridUnsafe-unmonitor-td23706.html -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (IGNITE-6804) Print a warning if HashMap is passed into bulk update operations
[ https://issues.apache.org/jira/browse/IGNITE-6804?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17020467#comment-17020467 ] Ignite TC Bot commented on IGNITE-6804: --- {panel:title=Branch: [pull/6976/head] Base: [master] : Possible Blockers (109)|borderStyle=dashed|borderColor=#ccc|titleBGColor=#F7D6C1} {color:#d04437}Cache 6{color} [[tests 0 TIMEOUT , Exit Code |https://ci.ignite.apache.org/viewLog.html?buildId=4947291]] {color:#d04437}ZooKeeper (Discovery) 2{color} [[tests 105|https://ci.ignite.apache.org/viewLog.html?buildId=4947261]] * ZookeeperDiscoverySpiTestSuite2: GridCacheReplicatedMultiNodeFullApiSelfTest.testPutxIfAbsentAsync - Test has low fail rate in base branch 0,0% and is not flaky * ZookeeperDiscoverySpiTestSuite2: GridCacheReplicatedMultiNodeFullApiSelfTest.testRemoveAllSkipStore - Test has low fail rate in base branch 0,0% and is not flaky * ZookeeperDiscoverySpiTestSuite2: GridCacheReplicatedMultiNodeFullApiSelfTest.testNoReadThroughTx - Test has low fail rate in base branch 0,0% and is not flaky * ZookeeperDiscoverySpiTestSuite2: GridCacheReplicatedMultiNodeFullApiSelfTest.testPutIfAbsent - Test has low fail rate in base branch 0,0% and is not flaky * ZookeeperDiscoverySpiTestSuite2: GridCacheReplicatedMultiNodeFullApiSelfTest.testRemoveAllDuplicatesTx - Test has low fail rate in base branch 0,0% and is not flaky * ZookeeperDiscoverySpiTestSuite2: GridCacheReplicatedMultiNodeFullApiSelfTest.testGetAndPutObject - Test has low fail rate in base branch 0,0% and is not flaky * ZookeeperDiscoverySpiTestSuite2: GridCacheReplicatedMultiNodeFullApiSelfTest.testContainsKeyTx - Test has low fail rate in base branch 0,0% and is not flaky * ZookeeperDiscoverySpiTestSuite2: GridCacheReplicatedMultiNodeFullApiSelfTest.testLoadAll - Test has low fail rate in base branch 0,0% and is not flaky * ZookeeperDiscoverySpiTestSuite2: GridCacheReplicatedMultiNodeFullApiSelfTest.testGetAndPutIfAbsent - Test has low fail rate in base branch 0,0% and is not flaky * ZookeeperDiscoverySpiTestSuite2: GridCacheReplicatedMultiNodeFullApiSelfTest.testTransformAllWithNulls - Test has low fail rate in base branch 0,0% and is not flaky * ZookeeperDiscoverySpiTestSuite2: GridCommandHandlerTest.testBaselineAutoAdjustmentAutoRemoveNode - Test has low fail rate in base branch 0,0% and is not flaky ... and 94 tests blockers {color:#d04437}Cache 2{color} [[tests 2|https://ci.ignite.apache.org/viewLog.html?buildId=4947287]] * IgniteCacheTestSuite2: GridCachePartitionedTxSingleThreadedSelfTest.testOptimisticRepeatableReadRollback - Test has low fail rate in base branch 0,0% and is not flaky * IgniteCacheTestSuite2: GridCachePartitionedTxSingleThreadedSelfTest.testOptimisticReadCommittedCommit - Test has low fail rate in base branch 0,0% and is not flaky {color:#d04437}PDS 4{color} [[tests 0 TIMEOUT , Exit Code |https://ci.ignite.apache.org/viewLog.html?buildId=4947306]] {panel} [TeamCity *-- Run :: All* Results|https://ci.ignite.apache.org/viewLog.html?buildId=4947337buildTypeId=IgniteTests24Java8_RunAll] > Print a warning if HashMap is passed into bulk update operations > > > Key: IGNITE-6804 > URL: https://issues.apache.org/jira/browse/IGNITE-6804 > Project: Ignite > Issue Type: Improvement > Components: cache >Reporter: Denis A. Magda >Assignee: Ilya Kasnacheev >Priority: Critical > Labels: usability > Time Spent: 50m > Remaining Estimate: 0h > > Ignite newcomers tend to stumble on deadlocks simply because the keys are > passed in an unordered HashMap. Propose to do the following: > * update bulk operations Java docs. > * print out a warning if not SortedMap (e.g. HashMap, > Weak/Identity/Concurrent/Linked HashMap etc) is passed into > a bulk method (instead of SortedMap) and contains more than 1 element. > However, we should make sure that we only print that warning once and not > every time the API is called. > * do not produce warning for explicit optimistic transactions > More details are here: > http://apache-ignite-developers.2346864.n4.nabble.com/Re-Ignite-2-0-0-GridUnsafe-unmonitor-td23706.html -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (IGNITE-6804) Print a warning if HashMap is passed into bulk update operations
[ https://issues.apache.org/jira/browse/IGNITE-6804?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17020293#comment-17020293 ] Ilya Kasnacheev commented on IGNITE-6804: - [~agura] thank you for review! Please review amended changes in PR. Either way I'm going to re-run for visa. > Print a warning if HashMap is passed into bulk update operations > > > Key: IGNITE-6804 > URL: https://issues.apache.org/jira/browse/IGNITE-6804 > Project: Ignite > Issue Type: Improvement > Components: cache >Reporter: Denis A. Magda >Assignee: Ilya Kasnacheev >Priority: Critical > Labels: usability > Time Spent: 50m > Remaining Estimate: 0h > > Ignite newcomers tend to stumble on deadlocks simply because the keys are > passed in an unordered HashMap. Propose to do the following: > * update bulk operations Java docs. > * print out a warning if not SortedMap (e.g. HashMap, > Weak/Identity/Concurrent/Linked HashMap etc) is passed into > a bulk method (instead of SortedMap) and contains more than 1 element. > However, we should make sure that we only print that warning once and not > every time the API is called. > * do not produce warning for explicit optimistic transactions > More details are here: > http://apache-ignite-developers.2346864.n4.nabble.com/Re-Ignite-2-0-0-GridUnsafe-unmonitor-td23706.html -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (IGNITE-6804) Print a warning if HashMap is passed into bulk update operations
[ https://issues.apache.org/jira/browse/IGNITE-6804?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17018091#comment-17018091 ] Ilya Kasnacheev commented on IGNITE-6804: - [~nizhikov] please review proposed fix. I have also fixed a few tests to not produce warnings, but far more exist. > Print a warning if HashMap is passed into bulk update operations > > > Key: IGNITE-6804 > URL: https://issues.apache.org/jira/browse/IGNITE-6804 > Project: Ignite > Issue Type: Improvement > Components: cache >Reporter: Denis A. Magda >Assignee: Ilya Kasnacheev >Priority: Critical > Labels: usability > Time Spent: 50m > Remaining Estimate: 0h > > Ignite newcomers tend to stumble on deadlocks simply because the keys are > passed in an unordered HashMap. Propose to do the following: > * update bulk operations Java docs. > * print out a warning if not SortedMap (e.g. HashMap, > Weak/Identity/Concurrent/Linked HashMap etc) is passed into > a bulk method (instead of SortedMap) and contains more than 1 element. > However, we should make sure that we only print that warning once and not > every time the API is called. > * do not produce warning for explicit optimistic transactions > More details are here: > http://apache-ignite-developers.2346864.n4.nabble.com/Re-Ignite-2-0-0-GridUnsafe-unmonitor-td23706.html -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (IGNITE-6804) Print a warning if HashMap is passed into bulk update operations
[ https://issues.apache.org/jira/browse/IGNITE-6804?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17018089#comment-17018089 ] Ignite TC Bot commented on IGNITE-6804: --- {panel:title=Branch: [pull/6976/head] Base: [master] : No blockers found!|borderStyle=dashed|borderColor=#ccc|titleBGColor=#D6F7C1}{panel} [TeamCity *-- Run :: All* Results|https://ci.ignite.apache.org/viewLog.html?buildId=4936570buildTypeId=IgniteTests24Java8_RunAll] > Print a warning if HashMap is passed into bulk update operations > > > Key: IGNITE-6804 > URL: https://issues.apache.org/jira/browse/IGNITE-6804 > Project: Ignite > Issue Type: Improvement > Components: cache >Reporter: Denis A. Magda >Assignee: Ilya Kasnacheev >Priority: Critical > Labels: usability > Time Spent: 50m > Remaining Estimate: 0h > > Ignite newcomers tend to stumble on deadlocks simply because the keys are > passed in an unordered HashMap. Propose to do the following: > * update bulk operations Java docs. > * print out a warning if not SortedMap (e.g. HashMap, > Weak/Identity/Concurrent/Linked HashMap etc) is passed into > a bulk method (instead of SortedMap) and contains more than 1 element. > However, we should make sure that we only print that warning once and not > every time the API is called. > * do not produce warning for explicit optimistic transactions > More details are here: > http://apache-ignite-developers.2346864.n4.nabble.com/Re-Ignite-2-0-0-GridUnsafe-unmonitor-td23706.html -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (IGNITE-6804) Print a warning if HashMap is passed into bulk update operations
[ https://issues.apache.org/jira/browse/IGNITE-6804?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17017461#comment-17017461 ] Ignite TC Bot commented on IGNITE-6804: --- {panel:title=Branch: [pull/6976/head] Base: [master] : Possible Blockers (49)|borderStyle=dashed|borderColor=#ccc|titleBGColor=#F7D6C1} {color:#d04437}Start Nodes{color} [[tests 1|https://ci.ignite.apache.org/viewLog.html?buildId=4936485]] * IgniteStartStopRestartTestSuite: IgniteProjectionStartStopRestartSelfTest.testCustomScript - Test has low fail rate in base branch 0,0% and is not flaky {color:#d04437}Queries 1{color} [[tests 42|https://ci.ignite.apache.org/viewLog.html?buildId=4936546]] * IgniteBinaryCacheQueryTestSuite: IgniteCachePartitionedTransactionalSnapshotColumnConstraintTest.testPutLongStringKeyField - Test has low fail rate in base branch 0,0% and is not flaky * IgniteBinaryCacheQueryTestSuite: IgniteCachePartitionedTransactionalSnapshotColumnConstraintTest.testPutValidDecimalKeyAndValueField - Test has low fail rate in base branch 0,0% and is not flaky * IgniteBinaryCacheQueryTestSuite: IgniteCachePartitionedTransactionalSnapshotColumnConstraintTest.testPutLongStringValueField - Test has low fail rate in base branch 0,0% and is not flaky * IgniteBinaryCacheQueryTestSuite: IgniteCacheReplicatedTransactionalSnapshotColumnConstraintTest.testPutValidDecimalKeyAndValueField - Test has low fail rate in base branch 0,0% and is not flaky * IgniteBinaryCacheQueryTestSuite: IgniteCacheReplicatedTransactionalSnapshotColumnConstraintTest.testPutLongStringKeyField - Test has low fail rate in base branch 0,0% and is not flaky * IgniteBinaryCacheQueryTestSuite: IgniteCacheReplicatedTransactionalSnapshotColumnConstraintTest.testPutLongStringValueField - Test has low fail rate in base branch 0,0% and is not flaky * IgniteBinaryCacheQueryTestSuite: IgniteCachePartitionedAtomicColumnConstraintsTest.testPutValidDecimalKeyAndValueField - Test has low fail rate in base branch 0,0% and is not flaky * IgniteBinaryCacheQueryTestSuite: IgniteCachePartitionedAtomicColumnConstraintsTest.testPutTooLongStringKeyFieldFail - Test has low fail rate in base branch 0,0% and is not flaky * IgniteBinaryCacheQueryTestSuite: IgniteCachePartitionedAtomicColumnConstraintsTest.testPutLongStringKeyField - Test has low fail rate in base branch 0,0% and is not flaky * IgniteBinaryCacheQueryTestSuite: IgniteCachePartitionedAtomicColumnConstraintsTest.testPutTooLongStringValueFieldFail - Test has low fail rate in base branch 0,0% and is not flaky * IgniteBinaryCacheQueryTestSuite: IgniteCachePartitionedAtomicColumnConstraintsTest.testPutTooLongDecimalValueFieldScaleFail - Test has low fail rate in base branch 0,0% and is not flaky ... and 31 tests blockers {color:#d04437}Data Structures{color} [[tests 4|https://ci.ignite.apache.org/viewLog.html?buildId=4936531]] * IgniteCacheDataStructuresSelfTestSuite: SingleBackupImplicitTransactionalReadRepairTest.test[getEntry=false, async=false] - Test has low fail rate in base branch 0,0% and is not flaky * IgniteCacheDataStructuresSelfTestSuite: SingleBackupImplicitTransactionalReadRepairTest.test[getEntry=false, async=true] - Test has low fail rate in base branch 0,0% and is not flaky * IgniteCacheDataStructuresSelfTestSuite: SingleBackupImplicitTransactionalReadRepairTest.test[getEntry=true, async=false] - Test has low fail rate in base branch 0,0% and is not flaky * IgniteCacheDataStructuresSelfTestSuite: SingleBackupImplicitTransactionalReadRepairTest.test[getEntry=true, async=true] - Test has low fail rate in base branch 0,0% and is not flaky {color:#d04437}Cache 9{color} [[tests 1|https://ci.ignite.apache.org/viewLog.html?buildId=4936527]] * IgniteCacheTestSuite9: TxPartitionCounterStateOnePrimaryTwoBackupsTest.testPartialPrepare_3TX_6_1 - Test has low fail rate in base branch 0,0% and is not flaky {color:#d04437}MVCC Cache 5{color} [[tests 1|https://ci.ignite.apache.org/viewLog.html?buildId=4936558]] * IgniteCacheMvccTestSuite5: GridCacheHashMapPutAllWarningsTest.testHashMapPutAllExplicitOptimistic - History for base branch is absent. {panel} [TeamCity *-- Run :: All* Results|https://ci.ignite.apache.org/viewLog.html?buildId=4936570buildTypeId=IgniteTests24Java8_RunAll] > Print a warning if HashMap is passed into bulk update operations > > > Key: IGNITE-6804 > URL: https://issues.apache.org/jira/browse/IGNITE-6804 > Project: Ignite > Issue Type: Improvement > Components: cache >Reporter: Denis A. Magda >Assignee: Ilya Kasnacheev >Priority: Critical > Labels: usability > Time Spent: 50m > Remaining Estimate: 0h > > Ignite newcomers tend to stumble on deadlocks simply because the keys are > passed in an unordered HashMap.
[jira] [Commented] (IGNITE-6804) Print a warning if HashMap is passed into bulk update operations
[ https://issues.apache.org/jira/browse/IGNITE-6804?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16989749#comment-16989749 ] Vyacheslav Koptilin commented on IGNITE-6804: - Hello [~nizhikov], [~avinogradov], > Maybe we should deprecate putAll(Map) and introduce putAll(SortedMap) instead? I don't think that {{putAll(Map)}} can be deprecated just because it is a part of JSR-109. > We should just perform instanceOf check inside the putAll() method if > pessimistic. The user can use any kind of implementation of Map interface. It does not seem a reliable solution. > This will affect atomics and optimistic transactions, which are resistant to > this issue. It must be done for the atomic cache as well. Please take a look at {{GridDhtAtomicCache.updateAllAsyncInternal0}} {code:java} // If batch store update is enabled, we need to lock all entries. // First, need to acquire locks on cache entries, then check filter. List locked = lockEntries(req, req.topologyVersion());; {code} > Print a warning if HashMap is passed into bulk update operations > > > Key: IGNITE-6804 > URL: https://issues.apache.org/jira/browse/IGNITE-6804 > Project: Ignite > Issue Type: Improvement > Components: cache >Reporter: Denis A. Magda >Assignee: Ilya Kasnacheev >Priority: Critical > Labels: usability > Time Spent: 50m > Remaining Estimate: 0h > > Ignite newcomers tend to stumble on deadlocks simply because the keys are > passed in an unordered HashMap. Propose to do the following: > * update bulk operations Java docs. > * print out a warning if not SortedMap (e.g. HashMap, > Weak/Identity/Concurrent/Linked HashMap etc) is passed into > a bulk method (instead of SortedMap) and contains more than 1 element. > However, we should make sure that we only print that warning once and not > every time the API is called. > * do not produce warning for explicit optimistic transactions > More details are here: > http://apache-ignite-developers.2346864.n4.nabble.com/Re-Ignite-2-0-0-GridUnsafe-unmonitor-td23706.html -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (IGNITE-6804) Print a warning if HashMap is passed into bulk update operations
[ https://issues.apache.org/jira/browse/IGNITE-6804?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16989716#comment-16989716 ] Anton Vinogradov commented on IGNITE-6804: -- [~nizhikov] This will affect atomics and optimistic transactions, which are resistant to this issue. > Print a warning if HashMap is passed into bulk update operations > > > Key: IGNITE-6804 > URL: https://issues.apache.org/jira/browse/IGNITE-6804 > Project: Ignite > Issue Type: Improvement > Components: cache >Reporter: Denis A. Magda >Assignee: Ilya Kasnacheev >Priority: Critical > Labels: usability > Time Spent: 50m > Remaining Estimate: 0h > > Ignite newcomers tend to stumble on deadlocks simply because the keys are > passed in an unordered HashMap. Propose to do the following: > * update bulk operations Java docs. > * print out a warning if not SortedMap (e.g. HashMap, > Weak/Identity/Concurrent/Linked HashMap etc) is passed into > a bulk method (instead of SortedMap) and contains more than 1 element. > However, we should make sure that we only print that warning once and not > every time the API is called. > * do not produce warning for explicit optimistic transactions > More details are here: > http://apache-ignite-developers.2346864.n4.nabble.com/Re-Ignite-2-0-0-GridUnsafe-unmonitor-td23706.html -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (IGNITE-6804) Print a warning if HashMap is passed into bulk update operations
[ https://issues.apache.org/jira/browse/IGNITE-6804?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16989707#comment-16989707 ] Nikolay Izhikov commented on IGNITE-6804: - [~avinogradov] Maybe we should deprecate putAll(Map) and introduce putAll(SortedMap) instead? > Print a warning if HashMap is passed into bulk update operations > > > Key: IGNITE-6804 > URL: https://issues.apache.org/jira/browse/IGNITE-6804 > Project: Ignite > Issue Type: Improvement > Components: cache >Reporter: Denis A. Magda >Assignee: Ilya Kasnacheev >Priority: Critical > Labels: usability > Time Spent: 50m > Remaining Estimate: 0h > > Ignite newcomers tend to stumble on deadlocks simply because the keys are > passed in an unordered HashMap. Propose to do the following: > * update bulk operations Java docs. > * print out a warning if not SortedMap (e.g. HashMap, > Weak/Identity/Concurrent/Linked HashMap etc) is passed into > a bulk method (instead of SortedMap) and contains more than 1 element. > However, we should make sure that we only print that warning once and not > every time the API is called. > * do not produce warning for explicit optimistic transactions > More details are here: > http://apache-ignite-developers.2346864.n4.nabble.com/Re-Ignite-2-0-0-GridUnsafe-unmonitor-td23706.html -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (IGNITE-6804) Print a warning if HashMap is passed into bulk update operations
[ https://issues.apache.org/jira/browse/IGNITE-6804?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16989706#comment-16989706 ] Anton Vinogradov commented on IGNITE-6804: -- Folks, A warning will be ignored. JavadoŃ will be ignored too. We should just perform instanceOf check inside the putAll() method if pessimistic. Such check will not drop overall performance since it will be evaluated on the client node. BTW, Can we perform same check by some annotation contract using static analysis? > Print a warning if HashMap is passed into bulk update operations > > > Key: IGNITE-6804 > URL: https://issues.apache.org/jira/browse/IGNITE-6804 > Project: Ignite > Issue Type: Improvement > Components: cache >Reporter: Denis A. Magda >Assignee: Ilya Kasnacheev >Priority: Critical > Labels: usability > Time Spent: 50m > Remaining Estimate: 0h > > Ignite newcomers tend to stumble on deadlocks simply because the keys are > passed in an unordered HashMap. Propose to do the following: > * update bulk operations Java docs. > * print out a warning if not SortedMap (e.g. HashMap, > Weak/Identity/Concurrent/Linked HashMap etc) is passed into > a bulk method (instead of SortedMap) and contains more than 1 element. > However, we should make sure that we only print that warning once and not > every time the API is called. > * do not produce warning for explicit optimistic transactions > More details are here: > http://apache-ignite-developers.2346864.n4.nabble.com/Re-Ignite-2-0-0-GridUnsafe-unmonitor-td23706.html -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (IGNITE-6804) Print a warning if HashMap is passed into bulk update operations
[ https://issues.apache.org/jira/browse/IGNITE-6804?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16989688#comment-16989688 ] Nikolay Izhikov commented on IGNITE-6804: - Hello, [~ilyak]. Any progress on that issue? As far as I know, users still stumble with this issue. It a very confusing Ignite behaviour, let's fix it! > Print a warning if HashMap is passed into bulk update operations > > > Key: IGNITE-6804 > URL: https://issues.apache.org/jira/browse/IGNITE-6804 > Project: Ignite > Issue Type: Improvement > Components: cache >Reporter: Denis A. Magda >Assignee: Ilya Kasnacheev >Priority: Critical > Labels: usability > Time Spent: 50m > Remaining Estimate: 0h > > Ignite newcomers tend to stumble on deadlocks simply because the keys are > passed in an unordered HashMap. Propose to do the following: > * update bulk operations Java docs. > * print out a warning if not SortedMap (e.g. HashMap, > Weak/Identity/Concurrent/Linked HashMap etc) is passed into > a bulk method (instead of SortedMap) and contains more than 1 element. > However, we should make sure that we only print that warning once and not > every time the API is called. > * do not produce warning for explicit optimistic transactions > More details are here: > http://apache-ignite-developers.2346864.n4.nabble.com/Re-Ignite-2-0-0-GridUnsafe-unmonitor-td23706.html -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (IGNITE-6804) Print a warning if HashMap is passed into bulk update operations
[ https://issues.apache.org/jira/browse/IGNITE-6804?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16975712#comment-16975712 ] Ignite TC Bot commented on IGNITE-6804: --- {panel:title=Branch: [pull/6976/head] Base: [master] : Possible Blockers (47)|borderStyle=dashed|borderColor=#ccc|titleBGColor=#F7D6C1} {color:#d04437}Cache 6{color} [[tests 0 Exit Code |https://ci.ignite.apache.org/viewLog.html?buildId=4755357]] {color:#d04437}Cache 5{color} [[tests 1|https://ci.ignite.apache.org/viewLog.html?buildId=4755178]] * IgniteCacheTestSuite5: ClientDelayedJoinTest.testClientJoinAndCacheStop - Test has low fail rate in base branch 0,0% and is not flaky {color:#d04437}Queries 1{color} [[tests 42|https://ci.ignite.apache.org/viewLog.html?buildId=4755201]] * IgniteBinaryCacheQueryTestSuite: IgniteCachePartitionedTransactionalSnapshotColumnConstraintTest.testPutLongStringKeyField - Test has low fail rate in base branch 0,0% and is not flaky * IgniteBinaryCacheQueryTestSuite: IgniteCachePartitionedTransactionalSnapshotColumnConstraintTest.testPutValidDecimalKeyAndValueField - Test has low fail rate in base branch 0,0% and is not flaky * IgniteBinaryCacheQueryTestSuite: IgniteCachePartitionedTransactionalSnapshotColumnConstraintTest.testPutLongStringValueField - Test has low fail rate in base branch 0,0% and is not flaky * IgniteBinaryCacheQueryTestSuite: IgniteCacheReplicatedTransactionalSnapshotColumnConstraintTest.testPutValidDecimalKeyAndValueField - Test has low fail rate in base branch 0,0% and is not flaky * IgniteBinaryCacheQueryTestSuite: IgniteCacheReplicatedTransactionalSnapshotColumnConstraintTest.testPutLongStringKeyField - Test has low fail rate in base branch 0,0% and is not flaky * IgniteBinaryCacheQueryTestSuite: IgniteCacheReplicatedTransactionalSnapshotColumnConstraintTest.testPutLongStringValueField - Test has low fail rate in base branch 0,0% and is not flaky * IgniteBinaryCacheQueryTestSuite: IgniteCachePartitionedAtomicColumnConstraintsTest.testPutValidDecimalKeyAndValueField - Test has low fail rate in base branch 0,0% and is not flaky * IgniteBinaryCacheQueryTestSuite: IgniteCachePartitionedAtomicColumnConstraintsTest.testPutTooLongStringKeyFieldFail - Test has low fail rate in base branch 0,0% and is not flaky * IgniteBinaryCacheQueryTestSuite: IgniteCachePartitionedAtomicColumnConstraintsTest.testPutLongStringKeyField - Test has low fail rate in base branch 0,0% and is not flaky * IgniteBinaryCacheQueryTestSuite: IgniteCachePartitionedAtomicColumnConstraintsTest.testPutTooLongStringValueFieldFail - Test has low fail rate in base branch 0,0% and is not flaky * IgniteBinaryCacheQueryTestSuite: IgniteCachePartitionedAtomicColumnConstraintsTest.testPutTooLongDecimalValueFieldScaleFail - Test has low fail rate in base branch 0,0% and is not flaky ... and 31 tests blockers {color:#d04437}MVCC Cache 5{color} [[tests 1|https://ci.ignite.apache.org/viewLog.html?buildId=4755213]] * IgniteCacheMvccTestSuite5: GridCacheHashMapPutAllWarningsTest.testHashMapPutAllExplicitOptimistic - History for base branch is absent. {color:#d04437}PDS 3{color} [[tests 1|https://ci.ignite.apache.org/viewLog.html?buildId=4755193]] * IgnitePdsTestSuite3: IgnitePdsContinuousRestartTest.testRebalancingDuringLoad_8000_8000_8_1 - Test has low fail rate in base branch 0,0% and is not flaky {color:#d04437}Platform .NET (Inspections)*{color} [[tests 0 Failure on metric |https://ci.ignite.apache.org/viewLog.html?buildId=4755197]] {panel} [TeamCity *-- Run :: All* Results|https://ci.ignite.apache.org/viewLog.html?buildId=4755225buildTypeId=IgniteTests24Java8_RunAll] > Print a warning if HashMap is passed into bulk update operations > > > Key: IGNITE-6804 > URL: https://issues.apache.org/jira/browse/IGNITE-6804 > Project: Ignite > Issue Type: Improvement > Components: cache >Reporter: Denis A. Magda >Assignee: Ilya Kasnacheev >Priority: Critical > Labels: usability > Time Spent: 50m > Remaining Estimate: 0h > > Ignite newcomers tend to stumble on deadlocks simply because the keys are > passed in an unordered HashMap. Propose to do the following: > * update bulk operations Java docs. > * print out a warning if not SortedMap (e.g. HashMap, > Weak/Identity/Concurrent/Linked HashMap etc) is passed into > a bulk method (instead of SortedMap) and contains more than 1 element. > However, we should make sure that we only print that warning once and not > every time the API is called. > * do not produce warning for explicit optimistic transactions > More details are here: > http://apache-ignite-developers.2346864.n4.nabble.com/Re-Ignite-2-0-0-GridUnsafe-unmonitor-td23706.html -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (IGNITE-6804) Print a warning if HashMap is passed into bulk update operations
[ https://issues.apache.org/jira/browse/IGNITE-6804?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16953880#comment-16953880 ] Ignite TC Bot commented on IGNITE-6804: --- {panel:title=Branch: [pull/6976/head] Base: [master] : Possible Blockers (3)|borderStyle=dashed|borderColor=#ccc|titleBGColor=#F7D6C1} {color:#d04437}Start Nodes{color} [[tests 1|https://ci.ignite.apache.org/viewLog.html?buildId=4697296]] * IgniteStartStopRestartTestSuite: IgniteProjectionStartStopRestartSelfTest.testStartOneNode - Test has low fail rate in base branch 0,0% and is not flaky {color:#d04437}Platform .NET (Inspections)*{color} [[tests 0 Failure on metric |https://ci.ignite.apache.org/viewLog.html?buildId=4698855]] {color:#d04437}~Build Apache Ignite~{color} [[tests 0 Exit Code |https://ci.ignite.apache.org/viewLog.html?buildId=4705081]] {panel} [TeamCity *-- Run :: All* Results|https://ci.ignite.apache.org/viewLog.html?buildId=4705196buildTypeId=IgniteTests24Java8_RunAll] > Print a warning if HashMap is passed into bulk update operations > > > Key: IGNITE-6804 > URL: https://issues.apache.org/jira/browse/IGNITE-6804 > Project: Ignite > Issue Type: Improvement > Components: cache >Reporter: Denis A. Magda >Assignee: Ilya Kasnacheev >Priority: Critical > Labels: usability > Time Spent: 50m > Remaining Estimate: 0h > > Ignite newcomers tend to stumble on deadlocks simply because the keys are > passed in an unordered HashMap. Propose to do the following: > * update bulk operations Java docs. > * print out a warning if not SortedMap (e.g. HashMap, > Weak/Identity/Concurrent/Linked HashMap etc) is passed into > a bulk method (instead of SortedMap) and contains more than 1 element. > However, we should make sure that we only print that warning once and not > every time the API is called. > * do not produce warning for explicit optimistic transactions > More details are here: > http://apache-ignite-developers.2346864.n4.nabble.com/Re-Ignite-2-0-0-GridUnsafe-unmonitor-td23706.html -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (IGNITE-6804) Print a warning if HashMap is passed into bulk update operations
[ https://issues.apache.org/jira/browse/IGNITE-6804?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16953852#comment-16953852 ] Ignite TC Bot commented on IGNITE-6804: --- {panel:title=Branch: [pull/6976/head] Base: [master] : Possible Blockers (3)|borderStyle=dashed|borderColor=#ccc|titleBGColor=#F7D6C1} {color:#d04437}Start Nodes{color} [[tests 1|https://ci.ignite.apache.org/viewLog.html?buildId=4697296]] * IgniteStartStopRestartTestSuite: IgniteProjectionStartStopRestartSelfTest.testStartOneNode - Test has low fail rate in base branch 0,0% and is not flaky {color:#d04437}Platform .NET (Inspections)*{color} [[tests 0 Failure on metric |https://ci.ignite.apache.org/viewLog.html?buildId=4698855]] {color:#d04437}~Build Apache Ignite~{color} [[tests 0 Exit Code |https://ci.ignite.apache.org/viewLog.html?buildId=4704733]] {panel} [TeamCity *-- Run :: All* Results|https://ci.ignite.apache.org/viewLog.html?buildId=4704848buildTypeId=IgniteTests24Java8_RunAll] > Print a warning if HashMap is passed into bulk update operations > > > Key: IGNITE-6804 > URL: https://issues.apache.org/jira/browse/IGNITE-6804 > Project: Ignite > Issue Type: Improvement > Components: cache >Reporter: Denis A. Magda >Assignee: Ilya Kasnacheev >Priority: Critical > Labels: usability > Time Spent: 50m > Remaining Estimate: 0h > > Ignite newcomers tend to stumble on deadlocks simply because the keys are > passed in an unordered HashMap. Propose to do the following: > * update bulk operations Java docs. > * print out a warning if not SortedMap (e.g. HashMap, > Weak/Identity/Concurrent/Linked HashMap etc) is passed into > a bulk method (instead of SortedMap) and contains more than 1 element. > However, we should make sure that we only print that warning once and not > every time the API is called. > * do not produce warning for explicit optimistic transactions > More details are here: > http://apache-ignite-developers.2346864.n4.nabble.com/Re-Ignite-2-0-0-GridUnsafe-unmonitor-td23706.html -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (IGNITE-6804) Print a warning if HashMap is passed into bulk update operations
[ https://issues.apache.org/jira/browse/IGNITE-6804?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16953276#comment-16953276 ] Denis A. Magda commented on IGNITE-6804: [~ilyak], thanks for taking over this task. I suggested some edits in PR. > Print a warning if HashMap is passed into bulk update operations > > > Key: IGNITE-6804 > URL: https://issues.apache.org/jira/browse/IGNITE-6804 > Project: Ignite > Issue Type: Improvement > Components: cache >Reporter: Denis A. Magda >Assignee: Ilya Kasnacheev >Priority: Critical > Labels: usability > Time Spent: 0.5h > Remaining Estimate: 0h > > Ignite newcomers tend to stumble on deadlocks simply because the keys are > passed in an unordered HashMap. Propose to do the following: > * update bulk operations Java docs. > * print out a warning if not SortedMap (e.g. HashMap, > Weak/Identity/Concurrent/Linked HashMap etc) is passed into > a bulk method (instead of SortedMap) and contains more than 1 element. > However, we should make sure that we only print that warning once and not > every time the API is called. > * do not produce warning for explicit optimistic transactions > More details are here: > http://apache-ignite-developers.2346864.n4.nabble.com/Re-Ignite-2-0-0-GridUnsafe-unmonitor-td23706.html -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (IGNITE-6804) Print a warning if HashMap is passed into bulk update operations
[ https://issues.apache.org/jira/browse/IGNITE-6804?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16952637#comment-16952637 ] Ilya Kasnacheev commented on IGNITE-6804: - [~ptupitsyn]If you have exactly the same number of same keys every time, it will work. If you have different sets of keys, deadlock. > Print a warning if HashMap is passed into bulk update operations > > > Key: IGNITE-6804 > URL: https://issues.apache.org/jira/browse/IGNITE-6804 > Project: Ignite > Issue Type: Improvement > Components: cache >Reporter: Denis A. Magda >Assignee: Ilya Kasnacheev >Priority: Critical > Labels: usability > Time Spent: 10m > Remaining Estimate: 0h > > Ignite newcomers tend to stumble on deadlocks simply because the keys are > passed in an unordered HashMap. Propose to do the following: > * update bulk operations Java docs. > * print out a warning if not SortedMap (e.g. HashMap, > Weak/Identity/Concurrent/Linked HashMap etc) is passed into > a bulk method (instead of SortedMap) and contains more than 1 element. > However, we should make sure that we only print that warning once and not > every time the API is called. > * do not produce warning for explicit optimistic transactions > More details are here: > http://apache-ignite-developers.2346864.n4.nabble.com/Re-Ignite-2-0-0-GridUnsafe-unmonitor-td23706.html -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (IGNITE-6804) Print a warning if HashMap is passed into bulk update operations
[ https://issues.apache.org/jira/browse/IGNITE-6804?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16952260#comment-16952260 ] Pavel Tupitsyn commented on IGNITE-6804: [~ilyak] if you add items to a new empty hashtable, the resulting order is going to be the same every time. And, I guess, this is the most common scenario. This also indicates a problem with the API. We waste CPU time populating HashMap or SortedMap, when Ignite only iterates over the result, throwing away all the hash-related stuff. Anyway, you can keep "will" if you like, but it is better to be technically correct (the best kind of correct). > Print a warning if HashMap is passed into bulk update operations > > > Key: IGNITE-6804 > URL: https://issues.apache.org/jira/browse/IGNITE-6804 > Project: Ignite > Issue Type: Improvement > Components: cache >Reporter: Denis A. Magda >Assignee: Ilya Kasnacheev >Priority: Critical > Labels: usability > Time Spent: 10m > Remaining Estimate: 0h > > Ignite newcomers tend to stumble on deadlocks simply because the keys are > passed in an unordered HashMap. Propose to do the following: > * update bulk operations Java docs. > * print out a warning if not SortedMap (e.g. HashMap, > Weak/Identity/Concurrent/Linked HashMap etc) is passed into > a bulk method (instead of SortedMap) and contains more than 1 element. > However, we should make sure that we only print that warning once and not > every time the API is called. > * do not produce warning for explicit optimistic transactions > More details are here: > http://apache-ignite-developers.2346864.n4.nabble.com/Re-Ignite-2-0-0-GridUnsafe-unmonitor-td23706.html -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (IGNITE-6804) Print a warning if HashMap is passed into bulk update operations
[ https://issues.apache.org/jira/browse/IGNITE-6804?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16952025#comment-16952025 ] Igor Sapego commented on IGNITE-6804: - [~ilyak] C++ changes looks good to me. > Print a warning if HashMap is passed into bulk update operations > > > Key: IGNITE-6804 > URL: https://issues.apache.org/jira/browse/IGNITE-6804 > Project: Ignite > Issue Type: Improvement > Components: cache >Reporter: Denis A. Magda >Assignee: Ilya Kasnacheev >Priority: Critical > Labels: usability > Time Spent: 10m > Remaining Estimate: 0h > > Ignite newcomers tend to stumble on deadlocks simply because the keys are > passed in an unordered HashMap. Propose to do the following: > * update bulk operations Java docs. > * print out a warning if not SortedMap (e.g. HashMap, > Weak/Identity/Concurrent/Linked HashMap etc) is passed into > a bulk method (instead of SortedMap) and contains more than 1 element. > However, we should make sure that we only print that warning once and not > every time the API is called. > * do not produce warning for explicit optimistic transactions > More details are here: > http://apache-ignite-developers.2346864.n4.nabble.com/Re-Ignite-2-0-0-GridUnsafe-unmonitor-td23706.html -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (IGNITE-6804) Print a warning if HashMap is passed into bulk update operations
[ https://issues.apache.org/jira/browse/IGNITE-6804?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16951967#comment-16951967 ] Ilya Kasnacheev commented on IGNITE-6804: - [~ptupitsyn] There's practically no way to not cause deadlock with HashMap so it makes sense to write "will". It's like with gets() function whose documentation starts with "+Never use this function.+" > Print a warning if HashMap is passed into bulk update operations > > > Key: IGNITE-6804 > URL: https://issues.apache.org/jira/browse/IGNITE-6804 > Project: Ignite > Issue Type: Improvement > Components: cache >Reporter: Denis A. Magda >Assignee: Ilya Kasnacheev >Priority: Critical > Labels: usability > Time Spent: 10m > Remaining Estimate: 0h > > Ignite newcomers tend to stumble on deadlocks simply because the keys are > passed in an unordered HashMap. Propose to do the following: > * update bulk operations Java docs. > * print out a warning if not SortedMap (e.g. HashMap, > Weak/Identity/Concurrent/Linked HashMap etc) is passed into > a bulk method (instead of SortedMap) and contains more than 1 element. > However, we should make sure that we only print that warning once and not > every time the API is called. > * do not produce warning for explicit optimistic transactions > More details are here: > http://apache-ignite-developers.2346864.n4.nabble.com/Re-Ignite-2-0-0-GridUnsafe-unmonitor-td23706.html -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (IGNITE-6804) Print a warning if HashMap is passed into bulk update operations
[ https://issues.apache.org/jira/browse/IGNITE-6804?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16951963#comment-16951963 ] Ilya Kasnacheev commented on IGNITE-6804: - [~dmagda] please review proposed documentation changes and developer warnings. > Print a warning if HashMap is passed into bulk update operations > > > Key: IGNITE-6804 > URL: https://issues.apache.org/jira/browse/IGNITE-6804 > Project: Ignite > Issue Type: Improvement > Components: cache >Reporter: Denis A. Magda >Assignee: Ilya Kasnacheev >Priority: Critical > Labels: usability > Time Spent: 10m > Remaining Estimate: 0h > > Ignite newcomers tend to stumble on deadlocks simply because the keys are > passed in an unordered HashMap. Propose to do the following: > * update bulk operations Java docs. > * print out a warning if not SortedMap (e.g. HashMap, > Weak/Identity/Concurrent/Linked HashMap etc) is passed into > a bulk method (instead of SortedMap) and contains more than 1 element. > However, we should make sure that we only print that warning once and not > every time the API is called. > * do not produce warning for explicit optimistic transactions > More details are here: > http://apache-ignite-developers.2346864.n4.nabble.com/Re-Ignite-2-0-0-GridUnsafe-unmonitor-td23706.html -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (IGNITE-6804) Print a warning if HashMap is passed into bulk update operations
[ https://issues.apache.org/jira/browse/IGNITE-6804?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16951962#comment-16951962 ] Ilya Kasnacheev commented on IGNITE-6804: - Failure not related to my change. > Print a warning if HashMap is passed into bulk update operations > > > Key: IGNITE-6804 > URL: https://issues.apache.org/jira/browse/IGNITE-6804 > Project: Ignite > Issue Type: Improvement > Components: cache >Reporter: Denis A. Magda >Assignee: Ilya Kasnacheev >Priority: Critical > Labels: usability > Time Spent: 10m > Remaining Estimate: 0h > > Ignite newcomers tend to stumble on deadlocks simply because the keys are > passed in an unordered HashMap. Propose to do the following: > * update bulk operations Java docs. > * print out a warning if not SortedMap (e.g. HashMap, > Weak/Identity/Concurrent/Linked HashMap etc) is passed into > a bulk method (instead of SortedMap) and contains more than 1 element. > However, we should make sure that we only print that warning once and not > every time the API is called. > * do not produce warning for explicit optimistic transactions > More details are here: > http://apache-ignite-developers.2346864.n4.nabble.com/Re-Ignite-2-0-0-GridUnsafe-unmonitor-td23706.html -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (IGNITE-6804) Print a warning if HashMap is passed into bulk update operations
[ https://issues.apache.org/jira/browse/IGNITE-6804?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16951960#comment-16951960 ] Ignite TC Bot commented on IGNITE-6804: --- {panel:title=Branch: [pull/6976/head] Base: [master] : Possible Blockers (1)|borderStyle=dashed|borderColor=#ccc|titleBGColor=#F7D6C1} {color:#d04437}Platform .NET (Inspections)*{color} [[tests 0 Failure on metric |https://ci.ignite.apache.org/viewLog.html?buildId=4698855]] {panel} [TeamCity *-- Run :: All* Results|https://ci.ignite.apache.org/viewLog.html?buildId=4697381buildTypeId=IgniteTests24Java8_RunAll] > Print a warning if HashMap is passed into bulk update operations > > > Key: IGNITE-6804 > URL: https://issues.apache.org/jira/browse/IGNITE-6804 > Project: Ignite > Issue Type: Improvement > Components: cache >Reporter: Denis A. Magda >Assignee: Ilya Kasnacheev >Priority: Critical > Labels: usability > Time Spent: 10m > Remaining Estimate: 0h > > Ignite newcomers tend to stumble on deadlocks simply because the keys are > passed in an unordered HashMap. Propose to do the following: > * update bulk operations Java docs. > * print out a warning if not SortedMap (e.g. HashMap, > Weak/Identity/Concurrent/Linked HashMap etc) is passed into > a bulk method (instead of SortedMap) and contains more than 1 element. > However, we should make sure that we only print that warning once and not > every time the API is called. > * do not produce warning for explicit optimistic transactions > More details are here: > http://apache-ignite-developers.2346864.n4.nabble.com/Re-Ignite-2-0-0-GridUnsafe-unmonitor-td23706.html -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (IGNITE-6804) Print a warning if HashMap is passed into bulk update operations
[ https://issues.apache.org/jira/browse/IGNITE-6804?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16951734#comment-16951734 ] Pavel Tupitsyn commented on IGNITE-6804: [~ilyak] # I've fixed collection class references in the docs (generic specification is mandatory there), pushed the branch (hope it's ok for you). # "will lead to deadlock" is not technically correct, I would change it to "may lead to deadlock". # Yes, we can do it in C# the same way as in Java. But I don't see how the warning is "throttled" - looks like it will be displayed on every operation, am I missing something? > Print a warning if HashMap is passed into bulk update operations > > > Key: IGNITE-6804 > URL: https://issues.apache.org/jira/browse/IGNITE-6804 > Project: Ignite > Issue Type: Improvement > Components: cache >Reporter: Denis A. Magda >Assignee: Ilya Kasnacheev >Priority: Critical > Labels: usability > Time Spent: 10m > Remaining Estimate: 0h > > Ignite newcomers tend to stumble on deadlocks simply because the keys are > passed in an unordered HashMap. Propose to do the following: > * update bulk operations Java docs. > * print out a warning if not SortedMap (e.g. HashMap, > Weak/Identity/Concurrent/Linked HashMap etc) is passed into > a bulk method (instead of SortedMap) and contains more than 1 element. > However, we should make sure that we only print that warning once and not > every time the API is called. > * do not produce warning for explicit optimistic transactions > More details are here: > http://apache-ignite-developers.2346864.n4.nabble.com/Re-Ignite-2-0-0-GridUnsafe-unmonitor-td23706.html -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (IGNITE-6804) Print a warning if HashMap is passed into bulk update operations
[ https://issues.apache.org/jira/browse/IGNITE-6804?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16951166#comment-16951166 ] Ignite TC Bot commented on IGNITE-6804: --- {panel:title=Branch: [pull/6976/head] Base: [master] : Possible Blockers (113)|borderStyle=dashed|borderColor=#ccc|titleBGColor=#F7D6C1} {color:#d04437}Platform C++ (Linux)*{color} [[tests 0 CANCELLED|https://ci.ignite.apache.org/viewLog.html?buildId=4696766]] {color:#d04437}Cache 6{color} [[tests 0 CANCELLED|https://ci.ignite.apache.org/viewLog.html?buildId=4696795]] {color:#d04437}SPI{color} [[tests 0 CANCELLED|https://ci.ignite.apache.org/viewLog.html?buildId=4696757]] {color:#d04437}ZooKeeper (Discovery) 3{color} [[tests 0 CANCELLED|https://ci.ignite.apache.org/viewLog.html?buildId=4696819]] {color:#d04437}ZooKeeper (Discovery) 1{color} [[tests 0 CANCELLED|https://ci.ignite.apache.org/viewLog.html?buildId=4696764]] {color:#d04437}Scala (Examples){color} [[tests 0 CANCELLED|https://ci.ignite.apache.org/viewLog.html?buildId=4696762]] {color:#d04437}Platform .NET (Long Running){color} [[tests 0 CANCELLED|https://ci.ignite.apache.org/viewLog.html?buildId=4696815]] {color:#d04437}Examples{color} [[tests 0 CANCELLED|https://ci.ignite.apache.org/viewLog.html?buildId=4696738]] {color:#d04437}PDS (Compatibility){color} [[tests 0 CANCELLED|https://ci.ignite.apache.org/viewLog.html?buildId=4696804]] {color:#d04437}Scala (Visor Console){color} [[tests 0 CANCELLED|https://ci.ignite.apache.org/viewLog.html?buildId=4696763]] {color:#d04437}RDD{color} [[tests 0 CANCELLED|https://ci.ignite.apache.org/viewLog.html?buildId=4696751]] {color:#d04437}PDS (Indexing){color} [[tests 0 CANCELLED|https://ci.ignite.apache.org/viewLog.html?buildId=4696805]] {color:#d04437}Cache (Expiry Policy){color} [[tests 0 CANCELLED|https://ci.ignite.apache.org/viewLog.html?buildId=4696780]] {color:#d04437}ZooKeeper (Discovery) 2{color} [[tests 0 CANCELLED|https://ci.ignite.apache.org/viewLog.html?buildId=4696765]] {color:#d04437}Streamers{color} [[tests 0 CANCELLED|https://ci.ignite.apache.org/viewLog.html?buildId=4696755]] {color:#d04437}Cache 2{color} [[tests 0 CANCELLED|https://ci.ignite.apache.org/viewLog.html?buildId=4696791]] {color:#d04437}Start Nodes{color} [[tests 0 CANCELLED|https://ci.ignite.apache.org/viewLog.html?buildId=4696756]] {color:#d04437}Platform .NET{color} [[tests 0 CANCELLED|https://ci.ignite.apache.org/viewLog.html?buildId=4696811]] {color:#d04437}Platform .NET (Core Linux){color} [[tests 0 CANCELLED|https://ci.ignite.apache.org/viewLog.html?buildId=4696812]] {color:#d04437}MVCC Queries{color} [[tests 0 CANCELLED|https://ci.ignite.apache.org/viewLog.html?buildId=4696775]] {color:#d04437}Continuous Query 3{color} [[tests 0 CANCELLED|https://ci.ignite.apache.org/viewLog.html?buildId=4696800]] {color:#d04437}Cache 5{color} [[tests 0 CANCELLED|https://ci.ignite.apache.org/viewLog.html?buildId=4696794]] {color:#d04437}PDS 1{color} [[tests 0 CANCELLED|https://ci.ignite.apache.org/viewLog.html?buildId=4696807]] {color:#d04437}Queries 1{color} [[tests 0 CANCELLED|https://ci.ignite.apache.org/viewLog.html?buildId=4696817]] {color:#d04437}ZooKeeper (Discovery) 4{color} [[tests 0 CANCELLED|https://ci.ignite.apache.org/viewLog.html?buildId=4696820]] {color:#d04437}MVCC PDS 1{color} [[tests 0 CANCELLED|https://ci.ignite.apache.org/viewLog.html?buildId=4696834]] {color:#d04437}Cache 1{color} [[tests 0 CANCELLED|https://ci.ignite.apache.org/viewLog.html?buildId=4696790]] {color:#d04437}Thin Client: Java{color} [[tests 0 CANCELLED|https://ci.ignite.apache.org/viewLog.html?buildId=4696748]] {color:#d04437}Compute (Grid){color} [[tests 0 CANCELLED|https://ci.ignite.apache.org/viewLog.html?buildId=4696735]] {color:#d04437}Basic 3{color} [[tests 0 CANCELLED|https://ci.ignite.apache.org/viewLog.html?buildId=4696770]] {color:#d04437}Data Structures{color} [[tests 0 CANCELLED|https://ci.ignite.apache.org/viewLog.html?buildId=4696802]] {color:#d04437}MVCC PDS 2{color} [[tests 0 CANCELLED|https://ci.ignite.apache.org/viewLog.html?buildId=4696835]] {color:#d04437}Cache 7{color} [[tests 0 CANCELLED|https://ci.ignite.apache.org/viewLog.html?buildId=4696796]] {color:#d04437}Continuous Query 4{color} [[tests 0 CANCELLED|https://ci.ignite.apache.org/viewLog.html?buildId=4696801]] {color:#d04437}Basic 1{color} [[tests 0 CANCELLED|https://ci.ignite.apache.org/viewLog.html?buildId=4696769]] {color:#d04437}MVCC Cache 1{color} [[tests 0 CANCELLED|https://ci.ignite.apache.org/viewLog.html?buildId=4696825]] {color:#d04437}Cache 8{color} [[tests 0 CANCELLED|https://ci.ignite.apache.org/viewLog.html?buildId=4696797]] {color:#d04437}Cache 3{color} [[tests 0 CANCELLED|https://ci.ignite.apache.org/viewLog.html?buildId=4696792]] {color:#d04437}PDS 3{color} [[tests 0 CANCELLED|https://ci.ignite.apache.org/viewLog.html?buildId=4696809]]
[jira] [Commented] (IGNITE-6804) Print a warning if HashMap is passed into bulk update operations
[ https://issues.apache.org/jira/browse/IGNITE-6804?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16951124#comment-16951124 ] Ilya Kasnacheev commented on IGNITE-6804: - [~ptupitsyn] please check if my additions to C# doc are well formed and correct. Can we issue warning from C# too? We would like to throttle it, i.e., only display once. Is it doable in C#? Any examples? [~isapego] please check if my additions to C++ doc are well formed and correct. > Print a warning if HashMap is passed into bulk update operations > > > Key: IGNITE-6804 > URL: https://issues.apache.org/jira/browse/IGNITE-6804 > Project: Ignite > Issue Type: Improvement > Components: cache >Reporter: Denis A. Magda >Assignee: Ilya Kasnacheev >Priority: Critical > Labels: usability > Time Spent: 10m > Remaining Estimate: 0h > > Ignite newcomers tend to stumble on deadlocks simply because the keys are > passed in an unordered HashMap. Propose to do the following: > * update bulk operations Java docs. > * print out a warning if not SortedMap (e.g. HashMap, > Weak/Identity/Concurrent/Linked HashMap etc) is passed into > a bulk method (instead of SortedMap) and contains more than 1 element. > However, we should make sure that we only print that warning once and not > every time the API is called. > * do not produce warning for explicit optimistic transactions > More details are here: > http://apache-ignite-developers.2346864.n4.nabble.com/Re-Ignite-2-0-0-GridUnsafe-unmonitor-td23706.html -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (IGNITE-6804) Print a warning if HashMap is passed into bulk update operations
[ https://issues.apache.org/jira/browse/IGNITE-6804?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16950996#comment-16950996 ] Ilya Kasnacheev commented on IGNITE-6804: - Do we also need to check invokeAll? > Print a warning if HashMap is passed into bulk update operations > > > Key: IGNITE-6804 > URL: https://issues.apache.org/jira/browse/IGNITE-6804 > Project: Ignite > Issue Type: Improvement > Components: cache >Reporter: Denis A. Magda >Assignee: Ilya Kasnacheev >Priority: Critical > Labels: usability > > Ignite newcomers tend to stumble on deadlocks simply because the keys are > passed in an unordered HashMap. Propose to do the following: > * update bulk operations Java docs. > * print out a warning if not SortedMap (e.g. HashMap, > Weak/Identity/Concurrent/Linked HashMap etc) is passed into > a bulk method (instead of SortedMap) and contains more than 1 element. > However, we should make sure that we only print that warning once and not > every time the API is called. > * do not produce warning for explicit optimistic transactions > More details are here: > http://apache-ignite-developers.2346864.n4.nabble.com/Re-Ignite-2-0-0-GridUnsafe-unmonitor-td23706.html -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (IGNITE-6804) Print a warning if HashMap is passed into bulk update operations
[ https://issues.apache.org/jira/browse/IGNITE-6804?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16950992#comment-16950992 ] Ilya Kasnacheev commented on IGNITE-6804: - Do we need to check for RemoveAll with HashSet? :) > Print a warning if HashMap is passed into bulk update operations > > > Key: IGNITE-6804 > URL: https://issues.apache.org/jira/browse/IGNITE-6804 > Project: Ignite > Issue Type: Improvement > Components: cache >Reporter: Denis A. Magda >Assignee: Ilya Kasnacheev >Priority: Critical > Labels: usability > > Ignite newcomers tend to stumble on deadlocks simply because the keys are > passed in an unordered HashMap. Propose to do the following: > * update bulk operations Java docs. > * print out a warning if not SortedMap (e.g. HashMap, > Weak/Identity/Concurrent/Linked HashMap etc) is passed into > a bulk method (instead of SortedMap) and contains more than 1 element. > However, we should make sure that we only print that warning once and not > every time the API is called. > * do not produce warning for explicit optimistic transactions > More details are here: > http://apache-ignite-developers.2346864.n4.nabble.com/Re-Ignite-2-0-0-GridUnsafe-unmonitor-td23706.html -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (IGNITE-6804) Print a warning if HashMap is passed into bulk update operations
[ https://issues.apache.org/jira/browse/IGNITE-6804?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16227492#comment-16227492 ] Denis Magda commented on IGNITE-6804: - [~dpavlov], yes thanks for clarifying the description. > Print a warning if HashMap is passed into bulk update operations > > > Key: IGNITE-6804 > URL: https://issues.apache.org/jira/browse/IGNITE-6804 > Project: Ignite > Issue Type: Improvement > Security Level: Public(Viewable by anyone) > Components: cache >Reporter: Denis Magda >Priority: Critical > Labels: usability > Fix For: 2.4 > > > Ignite newcomers tend to stumble on deadlocks simply because the keys are > passed in an unordered HashMap. Propose to do the following: > * update bulk operations Java docs. > * print out a warning if not SortedMap (e.g. HashMap, > Weak/Identity/Concurrent/Linked HashMap etc) is passed into > a bulk method (instead of SortedMap) and contains more than 1 element. > However, we should make sure that we only print that warning once and not > every time the API is called. > * do not produce warning for explicit optimistic transactions > More details are here: > http://apache-ignite-developers.2346864.n4.nabble.com/Re-Ignite-2-0-0-GridUnsafe-unmonitor-td23706.html -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Commented] (IGNITE-6804) Print a warning if HashMap is passed into bulk update operations
[ https://issues.apache.org/jira/browse/IGNITE-6804?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16227386#comment-16227386 ] Dmitriy Pavlov commented on IGNITE-6804: [~dmagda], I've changed description from HashMap to any not SortedMap, there can be a tons of maps provided into api. I suggest to check condition: not instanceof SortedMap. Are you agree? > Print a warning if HashMap is passed into bulk update operations > > > Key: IGNITE-6804 > URL: https://issues.apache.org/jira/browse/IGNITE-6804 > Project: Ignite > Issue Type: Improvement > Security Level: Public(Viewable by anyone) > Components: cache >Reporter: Denis Magda >Priority: Critical > Labels: usability > Fix For: 2.4 > > > Ignite newcomers tend to stumble on deadlocks simply because the keys are > passed in an unordered HashMap. Propose to do the following: > * update bulk operations Java docs. > * print out a warning if not SortedMap (e.g. HashMap, > Weak/Identity/Concurrent/Linked HashMap etc) is passed into > a bulk method (instead of SortedMap) and contains more than 1 element. > However, we should make sure that we only print that warning once and not > every time the API is called. > * do not produce warning for explicit optimistic transactions > More details are here: > http://apache-ignite-developers.2346864.n4.nabble.com/Re-Ignite-2-0-0-GridUnsafe-unmonitor-td23706.html -- This message was sent by Atlassian JIRA (v6.4.14#64029)