[jira] [Commented] (HDFS-14610) HashMap is not thread safe. Field storageMap is typically synchronized by storageMap. However, in one place, field storageMap is not protected with synchronized.
[
https://issues.apache.org/jira/browse/HDFS-14610?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16876695#comment-16876695
]
Paul Ward commented on HDFS-14610:
--
Hi Anu,
Thank you for sending those links.
Indeed, I am not familiar with the CR flow for Hadoop, so those links
speed up a *lot* my figuring out what is where.
Thanks!!
> HashMap is not thread safe. Field storageMap is typically synchronized by
> storageMap. However, in one place, field storageMap is not protected with
> synchronized.
> -
>
> Key: HDFS-14610
> URL: https://issues.apache.org/jira/browse/HDFS-14610
> Project: Hadoop HDFS
> Issue Type: Bug
>Reporter: Paul Ward
>Assignee: Paul Ward
>Priority: Critical
> Labels: fix-provided, patch-available
> Fix For: 3.3.0
>
> Attachments: addingSynchronization.patch
>
>
> I submitted a CR for this issue at:
> [https://github.com/apache/hadoop/pull/1015]
> The field *storageMap* (a *HashMap*)
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L155]
> is typically protected by synchronization on *storageMap*, e.g.,
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L294]
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L443]
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L484]
> For a total of 9 locations.
> The reason is because *HashMap* is not thread safe.
> However, here:
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L455]
> {{DatanodeStorageInfo storage =}}
> {{ storageMap.get(report.getStorage().getStorageID());}}
> It is not synchronized.
> Note that in the same method:
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L484]
> *storageMap* is again protected by synchronization:
> {{synchronized (storageMap) {}}
> {{ storageMapSize = storageMap.size();}}
> {{}}}
>
> The CR I inlined above protected the above instance (line 455 ) with
> synchronization
> like in line 484 and in all other occurrences.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
-
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Commented] (HDFS-14610) HashMap is not thread safe. Field storageMap is typically synchronized by storageMap. However, in one place, field storageMap is not protected with synchronized.
[
https://issues.apache.org/jira/browse/HDFS-14610?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16876323#comment-16876323
]
Paul Ward commented on HDFS-14610:
--
Hi Anu,
Thank you for explaining this convention. I was not aware of it.
I will do the tagging from now on.
Thanks!
> HashMap is not thread safe. Field storageMap is typically synchronized by
> storageMap. However, in one place, field storageMap is not protected with
> synchronized.
> -
>
> Key: HDFS-14610
> URL: https://issues.apache.org/jira/browse/HDFS-14610
> Project: Hadoop HDFS
> Issue Type: Bug
>Reporter: Paul Ward
>Assignee: Paul Ward
>Priority: Critical
> Labels: fix-provided, patch-available
> Attachments: addingSynchronization.patch
>
>
> I submitted a CR for this issue at:
> [https://github.com/apache/hadoop/pull/1015]
> The field *storageMap* (a *HashMap*)
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L155]
> is typically protected by synchronization on *storageMap*, e.g.,
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L294]
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L443]
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L484]
> For a total of 9 locations.
> The reason is because *HashMap* is not thread safe.
> However, here:
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L455]
> {{DatanodeStorageInfo storage =}}
> {{ storageMap.get(report.getStorage().getStorageID());}}
> It is not synchronized.
> Note that in the same method:
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L484]
> *storageMap* is again protected by synchronization:
> {{synchronized (storageMap) {}}
> {{ storageMapSize = storageMap.size();}}
> {{}}}
>
> The CR I inlined above protected the above instance (line 455 ) with
> synchronization
> like in line 484 and in all other occurrences.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
-
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Commented] (HDFS-14610) HashMap is not thread safe. Field storageMap is typically synchronized by storageMap. However, in one place, field storageMap is not protected with synchronized.
[
https://issues.apache.org/jira/browse/HDFS-14610?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16875954#comment-16875954
]
Paul Ward commented on HDFS-14610:
--
Anu,
Can you please take a look at the patch, like you did for:
https://issues.apache.org/jira/browse/HDFS-14618
?
I.e., I do not think those tests fail due to the patch.
Thanks
> HashMap is not thread safe. Field storageMap is typically synchronized by
> storageMap. However, in one place, field storageMap is not protected with
> synchronized.
> -
>
> Key: HDFS-14610
> URL: https://issues.apache.org/jira/browse/HDFS-14610
> Project: Hadoop HDFS
> Issue Type: Bug
>Reporter: Paul Ward
>Priority: Critical
> Labels: fix-provided, patch-available
> Attachments: addingSynchronization.patch
>
>
> I submitted a CR for this issue at:
> [https://github.com/apache/hadoop/pull/1015]
> The field *storageMap* (a *HashMap*)
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L155]
> is typically protected by synchronization on *storageMap*, e.g.,
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L294]
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L443]
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L484]
> For a total of 9 locations.
> The reason is because *HashMap* is not thread safe.
> However, here:
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L455]
> {{DatanodeStorageInfo storage =}}
> {{ storageMap.get(report.getStorage().getStorageID());}}
> It is not synchronized.
> Note that in the same method:
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L484]
> *storageMap* is again protected by synchronization:
> {{synchronized (storageMap) {}}
> {{ storageMapSize = storageMap.size();}}
> {{}}}
>
> The CR I inlined above protected the above instance (line 455 ) with
> synchronization
> like in line 484 and in all other occurrences.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
-
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Assigned] (HDFS-14610) HashMap is not thread safe. Field storageMap is typically synchronized by storageMap. However, in one place, field storageMap is not protected with synchronized.
[
https://issues.apache.org/jira/browse/HDFS-14610?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Paul Ward reassigned HDFS-14610:
Assignee: Anu Engineer
> HashMap is not thread safe. Field storageMap is typically synchronized by
> storageMap. However, in one place, field storageMap is not protected with
> synchronized.
> -
>
> Key: HDFS-14610
> URL: https://issues.apache.org/jira/browse/HDFS-14610
> Project: Hadoop HDFS
> Issue Type: Bug
>Reporter: Paul Ward
>Assignee: Anu Engineer
>Priority: Critical
> Labels: fix-provided, patch-available
> Attachments: addingSynchronization.patch
>
>
> I submitted a CR for this issue at:
> [https://github.com/apache/hadoop/pull/1015]
> The field *storageMap* (a *HashMap*)
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L155]
> is typically protected by synchronization on *storageMap*, e.g.,
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L294]
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L443]
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L484]
> For a total of 9 locations.
> The reason is because *HashMap* is not thread safe.
> However, here:
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L455]
> {{DatanodeStorageInfo storage =}}
> {{ storageMap.get(report.getStorage().getStorageID());}}
> It is not synchronized.
> Note that in the same method:
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L484]
> *storageMap* is again protected by synchronization:
> {{synchronized (storageMap) {}}
> {{ storageMapSize = storageMap.size();}}
> {{}}}
>
> The CR I inlined above protected the above instance (line 455 ) with
> synchronization
> like in line 484 and in all other occurrences.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
-
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Assigned] (HDFS-14610) HashMap is not thread safe. Field storageMap is typically synchronized by storageMap. However, in one place, field storageMap is not protected with synchronized.
[
https://issues.apache.org/jira/browse/HDFS-14610?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Paul Ward reassigned HDFS-14610:
Assignee: (was: Paul Ward)
> HashMap is not thread safe. Field storageMap is typically synchronized by
> storageMap. However, in one place, field storageMap is not protected with
> synchronized.
> -
>
> Key: HDFS-14610
> URL: https://issues.apache.org/jira/browse/HDFS-14610
> Project: Hadoop HDFS
> Issue Type: Bug
>Reporter: Paul Ward
>Priority: Critical
> Labels: fix-provided, patch-available
> Attachments: addingSynchronization.patch
>
>
> I submitted a CR for this issue at:
> [https://github.com/apache/hadoop/pull/1015]
> The field *storageMap* (a *HashMap*)
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L155]
> is typically protected by synchronization on *storageMap*, e.g.,
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L294]
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L443]
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L484]
> For a total of 9 locations.
> The reason is because *HashMap* is not thread safe.
> However, here:
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L455]
> {{DatanodeStorageInfo storage =}}
> {{ storageMap.get(report.getStorage().getStorageID());}}
> It is not synchronized.
> Note that in the same method:
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L484]
> *storageMap* is again protected by synchronization:
> {{synchronized (storageMap) {}}
> {{ storageMapSize = storageMap.size();}}
> {{}}}
>
> The CR I inlined above protected the above instance (line 455 ) with
> synchronization
> like in line 484 and in all other occurrences.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
-
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Commented] (HDFS-14618) Incorrect synchronization of ArrayList field (ArrayList is thread-unsafe).
[
https://issues.apache.org/jira/browse/HDFS-14618?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16875302#comment-16875302
]
Paul Ward commented on HDFS-14618:
--
Hi Anu,
Btw, can you please also take a look at:
https://issues.apache.org/jira/browse/HDFS-14610
I suppose there may be the same problem with non-deterministic performance.
I also added there an explanation for why that patch does not introduce a
deadlock
Thanks.
> Incorrect synchronization of ArrayList field (ArrayList is thread-unsafe).
> --
>
> Key: HDFS-14618
> URL: https://issues.apache.org/jira/browse/HDFS-14618
> Project: Hadoop HDFS
> Issue Type: Bug
>Reporter: Paul Ward
>Assignee: Paul Ward
>Priority: Critical
> Labels: fix-provided, patch-available
> Fix For: 3.3.0
>
> Attachments: race.patch
>
>
> I submitted a CR for this issue at:
> https://github.com/apache/hadoop/pull/1030
> The field {{timedOutItems}} (an {{ArrayList}}, i.e., not thread safe):
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L70
> is protected by synchronization on itself ({{timedOutItems}}):
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L167-L168
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L267-L268
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L178
> However, in one place:
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L133-L135
> it is (trying to be) protected by synchronized using
> {{pendingReconstructions}} --- but this cannot protect {{timedOutItems}}.
> Synchronized on different objects does not ensure mutual exclusion with the
> other locations.
> I.e., 2 code locations, one synchronized by {{pendingReconstructions}} and
> the other by {{timedOutItems}} can still executed concurrently.
> This CR adds the synchronized on {{timedOutItems}}.
> Note that this CR keeps the synchronized on {{pendingReconstructions}}, which
> is needed for a different purpose (protect {{pendingReconstructions}})
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
-
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Commented] (HDFS-14610) HashMap is not thread safe. Field storageMap is typically synchronized by storageMap. However, in one place, field storageMap is not protected with synchronized.
[
https://issues.apache.org/jira/browse/HDFS-14610?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16875301#comment-16875301
]
Paul Ward commented on HDFS-14610:
--
I do not know why the tests fail and I am not familiar with Hadoop internals to
debug the tests.
However, note that the patch just adds a
{{synchronized (storageMap) { }}
at line 455
[https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L455]
just like the one at line 484
[https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L484]
Line 455 and 484 are in the same method body.
I.e., this patch does not introduce a deadlock or such.
Maybe the tests fail due to non-deterministic performance in the containers,
like here ?:
https://issues.apache.org/jira/browse/HDFS-14618
> HashMap is not thread safe. Field storageMap is typically synchronized by
> storageMap. However, in one place, field storageMap is not protected with
> synchronized.
> -
>
> Key: HDFS-14610
> URL: https://issues.apache.org/jira/browse/HDFS-14610
> Project: Hadoop HDFS
> Issue Type: Bug
>Reporter: Paul Ward
>Priority: Critical
> Labels: fix-provided, patch-available
> Attachments: addingSynchronization.patch
>
>
> I submitted a CR for this issue at:
> [https://github.com/apache/hadoop/pull/1015]
> The field *storageMap* (a *HashMap*)
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L155]
> is typically protected by synchronization on *storageMap*, e.g.,
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L294]
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L443]
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L484]
> For a total of 9 locations.
> The reason is because *HashMap* is not thread safe.
> However, here:
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L455]
> {{DatanodeStorageInfo storage =}}
> {{ storageMap.get(report.getStorage().getStorageID());}}
> It is not synchronized.
> Note that in the same method:
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L484]
> *storageMap* is again protected by synchronization:
> {{synchronized (storageMap) {}}
> {{ storageMapSize = storageMap.size();}}
> {{}}}
>
> The CR I inlined above protected the above instance (line 455 ) with
> synchronization
> like in line 484 and in all other occurrences.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
-
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Commented] (HDFS-14618) Incorrect synchronization of ArrayList field (ArrayList is thread-unsafe).
[
https://issues.apache.org/jira/browse/HDFS-14618?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16875294#comment-16875294
]
Paul Ward commented on HDFS-14618:
--
Hi Anu,
Ok, thank you for helping with those tests!
> Incorrect synchronization of ArrayList field (ArrayList is thread-unsafe).
> --
>
> Key: HDFS-14618
> URL: https://issues.apache.org/jira/browse/HDFS-14618
> Project: Hadoop HDFS
> Issue Type: Bug
>Reporter: Paul Ward
>Assignee: Paul Ward
>Priority: Critical
> Labels: fix-provided, patch-available
> Fix For: 3.3.0
>
> Attachments: race.patch
>
>
> I submitted a CR for this issue at:
> https://github.com/apache/hadoop/pull/1030
> The field {{timedOutItems}} (an {{ArrayList}}, i.e., not thread safe):
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L70
> is protected by synchronization on itself ({{timedOutItems}}):
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L167-L168
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L267-L268
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L178
> However, in one place:
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L133-L135
> it is (trying to be) protected by synchronized using
> {{pendingReconstructions}} --- but this cannot protect {{timedOutItems}}.
> Synchronized on different objects does not ensure mutual exclusion with the
> other locations.
> I.e., 2 code locations, one synchronized by {{pendingReconstructions}} and
> the other by {{timedOutItems}} can still executed concurrently.
> This CR adds the synchronized on {{timedOutItems}}.
> Note that this CR keeps the synchronized on {{pendingReconstructions}}, which
> is needed for a different purpose (protect {{pendingReconstructions}})
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
-
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Commented] (HDFS-14618) Incorrect synchronization of ArrayList field (ArrayList is thread-unsafe).
[
https://issues.apache.org/jira/browse/HDFS-14618?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16875251#comment-16875251
]
Paul Ward commented on HDFS-14618:
--
Hi Anu,
Thank you for assigning me this.
I am not familiar with Hadoop internals to debug those tests.
However, note that the patch grabs the locks in the same order as here:
[https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L257-L267]
I.e., this patch does *not* introduce a deadlock.
Beyond that, I don't know why this patch would cause anything to fail.
Can you please take a look?
Thanks
> Incorrect synchronization of ArrayList field (ArrayList is thread-unsafe).
> --
>
> Key: HDFS-14618
> URL: https://issues.apache.org/jira/browse/HDFS-14618
> Project: Hadoop HDFS
> Issue Type: Bug
>Reporter: Paul Ward
>Assignee: Paul Ward
>Priority: Critical
> Labels: fix-provided, patch-available
> Attachments: race.patch
>
>
> I submitted a CR for this issue at:
> https://github.com/apache/hadoop/pull/1030
> The field {{timedOutItems}} (an {{ArrayList}}, i.e., not thread safe):
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L70
> is protected by synchronization on itself ({{timedOutItems}}):
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L167-L168
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L267-L268
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L178
> However, in one place:
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L133-L135
> it is (trying to be) protected by synchronized using
> {{pendingReconstructions}} --- but this cannot protect {{timedOutItems}}.
> Synchronized on different objects does not ensure mutual exclusion with the
> other locations.
> I.e., 2 code locations, one synchronized by {{pendingReconstructions}} and
> the other by {{timedOutItems}} can still executed concurrently.
> This CR adds the synchronized on {{timedOutItems}}.
> Note that this CR keeps the synchronized on {{pendingReconstructions}}, which
> is needed for a different purpose (protect {{pendingReconstructions}})
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
-
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Updated] (HDFS-14618) Incorrect synchronization of ArrayList field (ArrayList is thread-unsafe).
[
https://issues.apache.org/jira/browse/HDFS-14618?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Paul Ward updated HDFS-14618:
-
Attachment: (was: race.patch)
> Incorrect synchronization of ArrayList field (ArrayList is thread-unsafe).
> --
>
> Key: HDFS-14618
> URL: https://issues.apache.org/jira/browse/HDFS-14618
> Project: Hadoop HDFS
> Issue Type: Bug
>Reporter: Paul Ward
>Assignee: Paul Ward
>Priority: Critical
> Labels: fix-provided, patch-available
> Attachments: race.patch
>
>
> I submitted a CR for this issue at:
> https://github.com/apache/hadoop/pull/1030
> The field {{timedOutItems}} (an {{ArrayList}}, i.e., not thread safe):
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L70
> is protected by synchronization on itself ({{timedOutItems}}):
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L167-L168
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L267-L268
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L178
> However, in one place:
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L133-L135
> it is (trying to be) protected by synchronized using
> {{pendingReconstructions}} --- but this cannot protect {{timedOutItems}}.
> Synchronized on different objects does not ensure mutual exclusion with the
> other locations.
> I.e., 2 code locations, one synchronized by {{pendingReconstructions}} and
> the other by {{timedOutItems}} can still executed concurrently.
> This CR adds the synchronized on {{timedOutItems}}.
> Note that this CR keeps the synchronized on {{pendingReconstructions}}, which
> is needed for a different purpose (protect {{pendingReconstructions}})
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
-
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Updated] (HDFS-14618) Incorrect synchronization of ArrayList field (ArrayList is thread-unsafe).
[
https://issues.apache.org/jira/browse/HDFS-14618?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Paul Ward updated HDFS-14618:
-
Attachment: race.patch
Status: Patch Available (was: In Progress)
> Incorrect synchronization of ArrayList field (ArrayList is thread-unsafe).
> --
>
> Key: HDFS-14618
> URL: https://issues.apache.org/jira/browse/HDFS-14618
> Project: Hadoop HDFS
> Issue Type: Bug
>Reporter: Paul Ward
>Assignee: Paul Ward
>Priority: Critical
> Labels: fix-provided, patch-available
> Attachments: race.patch, race.patch
>
>
> I submitted a CR for this issue at:
> https://github.com/apache/hadoop/pull/1030
> The field {{timedOutItems}} (an {{ArrayList}}, i.e., not thread safe):
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L70
> is protected by synchronization on itself ({{timedOutItems}}):
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L167-L168
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L267-L268
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L178
> However, in one place:
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L133-L135
> it is (trying to be) protected by synchronized using
> {{pendingReconstructions}} --- but this cannot protect {{timedOutItems}}.
> Synchronized on different objects does not ensure mutual exclusion with the
> other locations.
> I.e., 2 code locations, one synchronized by {{pendingReconstructions}} and
> the other by {{timedOutItems}} can still executed concurrently.
> This CR adds the synchronized on {{timedOutItems}}.
> Note that this CR keeps the synchronized on {{pendingReconstructions}}, which
> is needed for a different purpose (protect {{pendingReconstructions}})
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
-
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Updated] (HDFS-14618) Incorrect synchronization of ArrayList field (ArrayList is thread-unsafe).
[
https://issues.apache.org/jira/browse/HDFS-14618?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Paul Ward updated HDFS-14618:
-
Status: In Progress (was: Patch Available)
> Incorrect synchronization of ArrayList field (ArrayList is thread-unsafe).
> --
>
> Key: HDFS-14618
> URL: https://issues.apache.org/jira/browse/HDFS-14618
> Project: Hadoop HDFS
> Issue Type: Bug
>Reporter: Paul Ward
>Assignee: Paul Ward
>Priority: Critical
> Labels: fix-provided, patch-available
> Attachments: race.patch
>
>
> I submitted a CR for this issue at:
> https://github.com/apache/hadoop/pull/1030
> The field {{timedOutItems}} (an {{ArrayList}}, i.e., not thread safe):
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L70
> is protected by synchronization on itself ({{timedOutItems}}):
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L167-L168
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L267-L268
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L178
> However, in one place:
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L133-L135
> it is (trying to be) protected by synchronized using
> {{pendingReconstructions}} --- but this cannot protect {{timedOutItems}}.
> Synchronized on different objects does not ensure mutual exclusion with the
> other locations.
> I.e., 2 code locations, one synchronized by {{pendingReconstructions}} and
> the other by {{timedOutItems}} can still executed concurrently.
> This CR adds the synchronized on {{timedOutItems}}.
> Note that this CR keeps the synchronized on {{pendingReconstructions}}, which
> is needed for a different purpose (protect {{pendingReconstructions}})
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
-
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Updated] (HDFS-14610) HashMap is not thread safe. Field storageMap is typically synchronized by storageMap. However, in one place, field storageMap is not protected with synchronized.
[
https://issues.apache.org/jira/browse/HDFS-14610?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Paul Ward updated HDFS-14610:
-
Flags: Patch
> HashMap is not thread safe. Field storageMap is typically synchronized by
> storageMap. However, in one place, field storageMap is not protected with
> synchronized.
> -
>
> Key: HDFS-14610
> URL: https://issues.apache.org/jira/browse/HDFS-14610
> Project: Hadoop HDFS
> Issue Type: Bug
>Reporter: Paul Ward
>Priority: Critical
> Labels: fix-provided, patch-available
> Attachments: addingSynchronization.patch
>
>
> I submitted a CR for this issue at:
> [https://github.com/apache/hadoop/pull/1015]
> The field *storageMap* (a *HashMap*)
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L155]
> is typically protected by synchronization on *storageMap*, e.g.,
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L294]
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L443]
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L484]
> For a total of 9 locations.
> The reason is because *HashMap* is not thread safe.
> However, here:
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L455]
> {{DatanodeStorageInfo storage =}}
> {{ storageMap.get(report.getStorage().getStorageID());}}
> It is not synchronized.
> Note that in the same method:
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L484]
> *storageMap* is again protected by synchronization:
> {{synchronized (storageMap) {}}
> {{ storageMapSize = storageMap.size();}}
> {{}}}
>
> The CR I inlined above protected the above instance (line 455 ) with
> synchronization
> like in line 484 and in all other occurrences.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
-
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Updated] (HDFS-14618) Incorrect synchronization of ArrayList field (ArrayList is thread-unsafe).
[
https://issues.apache.org/jira/browse/HDFS-14618?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Paul Ward updated HDFS-14618:
-
Labels: fix-provided patch-available (was: )
> Incorrect synchronization of ArrayList field (ArrayList is thread-unsafe).
> --
>
> Key: HDFS-14618
> URL: https://issues.apache.org/jira/browse/HDFS-14618
> Project: Hadoop HDFS
> Issue Type: Bug
>Reporter: Paul Ward
>Priority: Critical
> Labels: fix-provided, patch-available
> Attachments: race.patch
>
>
> I submitted a CR for this issue at:
> https://github.com/apache/hadoop/pull/1030
> The field {{timedOutItems}} (an {{ArrayList}}, i.e., not thread safe):
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L70
> is protected by synchronization on itself ({{timedOutItems}}):
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L167-L168
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L267-L268
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L178
> However, in one place:
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L133-L135
> it is (trying to be) protected by synchronized using
> {{pendingReconstructions}} --- but this cannot protect {{timedOutItems}}.
> Synchronized on different objects does not ensure mutual exclusion with the
> other locations.
> I.e., 2 code locations, one synchronized by {{pendingReconstructions}} and
> the other by {{timedOutItems}} can still executed concurrently.
> This CR adds the synchronized on {{timedOutItems}}.
> Note that this CR keeps the synchronized on {{pendingReconstructions}}, which
> is needed for a different purpose (protect {{pendingReconstructions}})
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
-
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Updated] (HDFS-14610) HashMap is not thread safe. Field storageMap is typically synchronized by storageMap. However, in one place, field storageMap is not protected with synchronized.
[
https://issues.apache.org/jira/browse/HDFS-14610?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Paul Ward updated HDFS-14610:
-
Labels: fix-provided patch-available (was: fix-provided)
> HashMap is not thread safe. Field storageMap is typically synchronized by
> storageMap. However, in one place, field storageMap is not protected with
> synchronized.
> -
>
> Key: HDFS-14610
> URL: https://issues.apache.org/jira/browse/HDFS-14610
> Project: Hadoop HDFS
> Issue Type: Bug
>Reporter: Paul Ward
>Priority: Critical
> Labels: fix-provided, patch-available
> Attachments: addingSynchronization.patch
>
>
> I submitted a CR for this issue at:
> [https://github.com/apache/hadoop/pull/1015]
> The field *storageMap* (a *HashMap*)
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L155]
> is typically protected by synchronization on *storageMap*, e.g.,
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L294]
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L443]
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L484]
> For a total of 9 locations.
> The reason is because *HashMap* is not thread safe.
> However, here:
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L455]
> {{DatanodeStorageInfo storage =}}
> {{ storageMap.get(report.getStorage().getStorageID());}}
> It is not synchronized.
> Note that in the same method:
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L484]
> *storageMap* is again protected by synchronization:
> {{synchronized (storageMap) {}}
> {{ storageMapSize = storageMap.size();}}
> {{}}}
>
> The CR I inlined above protected the above instance (line 455 ) with
> synchronization
> like in line 484 and in all other occurrences.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
-
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Updated] (HDFS-14618) Incorrect synchronization of ArrayList field (ArrayList is thread-unsafe).
[
https://issues.apache.org/jira/browse/HDFS-14618?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Paul Ward updated HDFS-14618:
-
Attachment: (was: race.diff)
> Incorrect synchronization of ArrayList field (ArrayList is thread-unsafe).
> --
>
> Key: HDFS-14618
> URL: https://issues.apache.org/jira/browse/HDFS-14618
> Project: Hadoop HDFS
> Issue Type: Bug
>Reporter: Paul Ward
>Priority: Critical
> Attachments: race.patch
>
>
> I submitted a CR for this issue at:
> https://github.com/apache/hadoop/pull/1030
> The field {{timedOutItems}} (an {{ArrayList}}, i.e., not thread safe):
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L70
> is protected by synchronization on itself ({{timedOutItems}}):
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L167-L168
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L267-L268
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L178
> However, in one place:
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L133-L135
> it is (trying to be) protected by synchronized using
> {{pendingReconstructions}} --- but this cannot protect {{timedOutItems}}.
> Synchronized on different objects does not ensure mutual exclusion with the
> other locations.
> I.e., 2 code locations, one synchronized by {{pendingReconstructions}} and
> the other by {{timedOutItems}} can still executed concurrently.
> This CR adds the synchronized on {{timedOutItems}}.
> Note that this CR keeps the synchronized on {{pendingReconstructions}}, which
> is needed for a different purpose (protect {{pendingReconstructions}})
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
-
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Updated] (HDFS-14618) Incorrect synchronization of ArrayList field (ArrayList is thread-unsafe).
[
https://issues.apache.org/jira/browse/HDFS-14618?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Paul Ward updated HDFS-14618:
-
Attachment: race.patch
> Incorrect synchronization of ArrayList field (ArrayList is thread-unsafe).
> --
>
> Key: HDFS-14618
> URL: https://issues.apache.org/jira/browse/HDFS-14618
> Project: Hadoop HDFS
> Issue Type: Bug
>Reporter: Paul Ward
>Priority: Critical
> Attachments: race.patch
>
>
> I submitted a CR for this issue at:
> https://github.com/apache/hadoop/pull/1030
> The field {{timedOutItems}} (an {{ArrayList}}, i.e., not thread safe):
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L70
> is protected by synchronization on itself ({{timedOutItems}}):
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L167-L168
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L267-L268
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L178
> However, in one place:
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L133-L135
> it is (trying to be) protected by synchronized using
> {{pendingReconstructions}} --- but this cannot protect {{timedOutItems}}.
> Synchronized on different objects does not ensure mutual exclusion with the
> other locations.
> I.e., 2 code locations, one synchronized by {{pendingReconstructions}} and
> the other by {{timedOutItems}} can still executed concurrently.
> This CR adds the synchronized on {{timedOutItems}}.
> Note that this CR keeps the synchronized on {{pendingReconstructions}}, which
> is needed for a different purpose (protect {{pendingReconstructions}})
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
-
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Updated] (HDFS-14610) HashMap is not thread safe. Field storageMap is typically synchronized by storageMap. However, in one place, field storageMap is not protected with synchronized.
[
https://issues.apache.org/jira/browse/HDFS-14610?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Paul Ward updated HDFS-14610:
-
Description:
I submitted a CR for this issue at:
[https://github.com/apache/hadoop/pull/1015]
The field *storageMap* (a *HashMap*)
[https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L155]
is typically protected by synchronization on *storageMap*, e.g.,
[https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L294]
[https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L443]
[https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L484]
For a total of 9 locations.
The reason is because *HashMap* is not thread safe.
However, here:
[https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L455]
{{DatanodeStorageInfo storage =}}
{{ storageMap.get(report.getStorage().getStorageID());}}
It is not synchronized.
Note that in the same method:
[https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L484]
*storageMap* is again protected by synchronization:
{{synchronized (storageMap) {}}
{{ storageMapSize = storageMap.size();}}
{{}}}
The CR I inlined above protected the above instance (line 455 ) with
synchronization
like in line 484 and in all other occurrences.
was:
The field *storageMap* (a *HashMap*)
[https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L155]
is typically protected by synchronization on *storageMap*, e.g.,
[https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L294]
[https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L443]
[https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L484]
For a total of 9 locations.
The reason is because *HashMap* is not thread safe.
However, here:
[https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L455]
{{DatanodeStorageInfo storage =}}
{{ storageMap.get(report.getStorage().getStorageID());}}
It is not synchronized.
Note that in the same method:
[https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L484]
*storageMap* is again protected by synchronization:
{{synchronized (storageMap) {}}
{{ storageMapSize = storageMap.size();}}
{{}}}
I submitted a CR:
[https://github.com/apache/hadoop/pull/1015]
This CR protected the above instance (line 455 ) with synchronization
like in line 484 and in all other occurrences.
> HashMap is not thread safe. Field storageMap is typically synchronized by
> storageMap. However, in one place, field storageMap is not protected with
> synchronized.
> -
>
> Key: HDFS-14610
> URL: https://issues.apache.org/jira/browse/HDFS-14610
> Project: Hadoop HDFS
> Issue Type: Bug
>Reporter: Paul Ward
>Priority: Critical
> Labels: fix-provided
> Attachments: addingSynchronization.patch
>
>
> I submitted a CR for this issue at:
> [https://github.com/apache/hadoop/pull/1015]
> The field *storageMap* (a *HashMap*)
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L155]
> is typically protected by synchronization on *storageMap*, e.g.,
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L294]
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L443]
>
[jira] [Updated] (HDFS-14610) HashMap is not thread safe. Field storageMap is typically synchronized by storageMap. However, in one place, field storageMap is not protected with synchronized.
[
https://issues.apache.org/jira/browse/HDFS-14610?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Paul Ward updated HDFS-14610:
-
Attachment: addingSynchronization.patch
Status: Patch Available (was: Open)
This is the same patch I submitted in the CR I inlined in the text
> HashMap is not thread safe. Field storageMap is typically synchronized by
> storageMap. However, in one place, field storageMap is not protected with
> synchronized.
> -
>
> Key: HDFS-14610
> URL: https://issues.apache.org/jira/browse/HDFS-14610
> Project: Hadoop HDFS
> Issue Type: Bug
>Reporter: Paul Ward
>Priority: Critical
> Labels: fix-provided
> Attachments: addingSynchronization.patch
>
>
> The field *storageMap* (a *HashMap*)
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L155]
> is typically protected by synchronization on *storageMap*, e.g.,
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L294]
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L443]
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L484]
> For a total of 9 locations.
> The reason is because *HashMap* is not thread safe.
> However, here:
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L455]
> {{DatanodeStorageInfo storage =}}
> {{ storageMap.get(report.getStorage().getStorageID());}}
> It is not synchronized.
> Note that in the same method:
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L484]
> *storageMap* is again protected by synchronization:
> {{synchronized (storageMap) {}}
> {{ storageMapSize = storageMap.size();}}
> {{}}}
> I submitted a CR:
>
> [https://github.com/apache/hadoop/pull/1015]
>
> This CR protected the above instance (line 455 ) with synchronization
> like in line 484 and in all other occurrences.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
-
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Updated] (HDFS-14618) Incorrect synchronization of ArrayList field (ArrayList is thread-unsafe).
[
https://issues.apache.org/jira/browse/HDFS-14618?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Paul Ward updated HDFS-14618:
-
Attachment: race.diff
Status: Patch Available (was: Open)
This diff is the same from the Pull request I inlined
> Incorrect synchronization of ArrayList field (ArrayList is thread-unsafe).
> --
>
> Key: HDFS-14618
> URL: https://issues.apache.org/jira/browse/HDFS-14618
> Project: Hadoop HDFS
> Issue Type: Bug
>Reporter: Paul Ward
>Priority: Critical
> Attachments: race.diff
>
>
> I submitted a CR for this issue at:
> https://github.com/apache/hadoop/pull/1030
> The field {{timedOutItems}} (an {{ArrayList}}, i.e., not thread safe):
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L70
> is protected by synchronization on itself ({{timedOutItems}}):
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L167-L168
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L267-L268
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L178
> However, in one place:
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L133-L135
> it is (trying to be) protected by synchronized using
> {{pendingReconstructions}} --- but this cannot protect {{timedOutItems}}.
> Synchronized on different objects does not ensure mutual exclusion with the
> other locations.
> I.e., 2 code locations, one synchronized by {{pendingReconstructions}} and
> the other by {{timedOutItems}} can still executed concurrently.
> This CR adds the synchronized on {{timedOutItems}}.
> Note that this CR keeps the synchronized on {{pendingReconstructions}}, which
> is needed for a different purpose (protect {{pendingReconstructions}})
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
-
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Updated] (HDFS-14618) Incorrect synchronization of ArrayList field (ArrayList is thread-unsafe).
[
https://issues.apache.org/jira/browse/HDFS-14618?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Paul Ward updated HDFS-14618:
-
Description:
I submitted a CR for this issue at:
https://github.com/apache/hadoop/pull/1030
The field {{timedOutItems}} (an {{ArrayList}}, i.e., not thread safe):
https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L70
is protected by synchronization on itself ({{timedOutItems}}):
https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L167-L168
https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L267-L268
https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L178
However, in one place:
https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L133-L135
it is (trying to be) protected by synchronized using {{pendingReconstructions}}
--- but this cannot protect {{timedOutItems}}.
Synchronized on different objects does not ensure mutual exclusion with the
other locations.
I.e., 2 code locations, one synchronized by {{pendingReconstructions}} and the
other by {{timedOutItems}} can still executed concurrently.
This CR adds the synchronized on {{timedOutItems}}.
Note that this CR keeps the synchronized on {{pendingReconstructions}}, which
is needed for a different purpose (protect {{pendingReconstructions}})
was:
I submitted a CR for this issue at:
https://github.com/apache/hadoop/pull/1030
The field ```timedOutItems``` (an ```ArrayList```, i.e., not thread safe):
https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L70
is protected by synchronization on itself (```timedOutItems```):
https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L167-L168
https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L267-L268
https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L178
However, in one place:
https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L133-L135
it is (trying to be) protected by synchronized using
```pendingReconstructions``` --- but this cannot protect ```timedOutItems```.
Synchronized on different objects does not ensure mutual exclusion
with the other locations.
I.e., 2 code locations, one synchronized by
```pendingReconstructions``` and the other by ```timedOutItems``` can
still executed concurrently.
This CR adds the synchronized on ```timedOutItems```.
Note that this CR keeps the synchronized on
```pendingReconstructions```, which is needed for a different purpose
(protect ```pendingReconstructions```)
> Incorrect synchronization of ArrayList field (ArrayList is thread-unsafe).
> --
>
> Key: HDFS-14618
> URL: https://issues.apache.org/jira/browse/HDFS-14618
> Project: Hadoop HDFS
> Issue Type: Bug
>Reporter: Paul Ward
>Priority: Critical
>
> I submitted a CR for this issue at:
> https://github.com/apache/hadoop/pull/1030
> The field {{timedOutItems}} (an {{ArrayList}}, i.e., not thread safe):
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L70
> is protected by synchronization on itself ({{timedOutItems}}):
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L167-L168
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L267-L268
> https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L178
> However, in one place:
>
[jira] [Created] (HDFS-14618) Incorrect synchronization of ArrayList field (ArrayList is thread-unsafe).
Paul Ward created HDFS-14618: Summary: Incorrect synchronization of ArrayList field (ArrayList is thread-unsafe). Key: HDFS-14618 URL: https://issues.apache.org/jira/browse/HDFS-14618 Project: Hadoop HDFS Issue Type: Bug Reporter: Paul Ward I submitted a CR for this issue at: https://github.com/apache/hadoop/pull/1030 The field ```timedOutItems``` (an ```ArrayList```, i.e., not thread safe): https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L70 is protected by synchronization on itself (```timedOutItems```): https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L167-L168 https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L267-L268 https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L178 However, in one place: https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/PendingReconstructionBlocks.java#L133-L135 it is (trying to be) protected by synchronized using ```pendingReconstructions``` --- but this cannot protect ```timedOutItems```. Synchronized on different objects does not ensure mutual exclusion with the other locations. I.e., 2 code locations, one synchronized by ```pendingReconstructions``` and the other by ```timedOutItems``` can still executed concurrently. This CR adds the synchronized on ```timedOutItems```. Note that this CR keeps the synchronized on ```pendingReconstructions```, which is needed for a different purpose (protect ```pendingReconstructions```) -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Updated] (HDFS-14610) HashMap is not thread safe. Field storageMap is typically synchronized by storageMap. However, in one place, field storageMap is not protected with synchronized.
[
https://issues.apache.org/jira/browse/HDFS-14610?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Paul Ward updated HDFS-14610:
-
Description:
The field *storageMap* (a *HashMap*)
[https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L155]
is typically protected by synchronization on *storageMap*, e.g.,
[https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L294]
[https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L443]
[https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L484]
For a total of 9 locations.
The reason is because *HashMap* is not thread safe.
However, here:
[https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L455]
{{DatanodeStorageInfo storage =}}
{{ storageMap.get(report.getStorage().getStorageID());}}
It is not synchronized.
Note that in the same method:
[https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L484]
*storageMap* is again protected by synchronization:
{{synchronized (storageMap) {}}
{{ storageMapSize = storageMap.size();}}
{{}}}
I submitted a CR:
[https://github.com/apache/hadoop/pull/1015]
This CR protected the above instance (line 455 ) with synchronization
like in line 484 and in all other occurrences.
was:
The field *storageMap* (a *HashMap*)
[https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L155]
is typically protected by synchronization on *storageMap*, e.g.,
[https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L294]
[https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L443]
[https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L484]
For a total of 9 locations.
The reason is because *HashMap* is not thread safe.
However, here:
[https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L455]
{{ DatanodeStorageInfo storage =}}
{{ storageMap.get(report.getStorage().getStorageID());}}
It is not synchronized.
Note that in the same method:
[https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L484]
*storageMap* is again protected by synchronization:
{{ synchronized (storageMap) {}}
{{ storageMapSize = storageMap.size();}}
{{ }}}
I submitted a CR:
[https://github.com/apache/hadoop/pull/1015]
This CR protected the above instance (line 455 ) with synchronization
like in line 484 and in all other occurrences.
> HashMap is not thread safe. Field storageMap is typically synchronized by
> storageMap. However, in one place, field storageMap is not protected with
> synchronized.
> -
>
> Key: HDFS-14610
> URL: https://issues.apache.org/jira/browse/HDFS-14610
> Project: Hadoop HDFS
> Issue Type: Bug
>Reporter: Paul Ward
>Priority: Critical
>
> The field *storageMap* (a *HashMap*)
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L155]
> is typically protected by synchronization on *storageMap*, e.g.,
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L294]
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L443]
> [https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L484]
> For a total of 9 locations.
> The reason is because
[jira] [Created] (HDFS-14610) HashMap is not thread safe. Field storageMap is typically synchronized by storageMap. However, in one place, field storageMap is not protected with synchronized.
Paul Ward created HDFS-14610:
Summary: HashMap is not thread safe. Field storageMap is typically
synchronized by storageMap. However, in one place, field storageMap is not
protected with synchronized.
Key: HDFS-14610
URL: https://issues.apache.org/jira/browse/HDFS-14610
Project: Hadoop HDFS
Issue Type: Bug
Reporter: Paul Ward
The field *storageMap* (a *HashMap*)
[https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L155]
is typically protected by synchronization on *storageMap*, e.g.,
[https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L294]
[https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L443]
[https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L484]
For a total of 9 locations.
The reason is because *HashMap* is not thread safe.
However, here:
[https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L455]
{{ DatanodeStorageInfo storage =}}
{{ storageMap.get(report.getStorage().getStorageID());}}
It is not synchronized.
Note that in the same method:
[https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeDescriptor.java#L484]
*storageMap* is again protected by synchronization:
{{ synchronized (storageMap) {}}
{{ storageMapSize = storageMap.size();}}
{{ }}}
I submitted a CR:
[https://github.com/apache/hadoop/pull/1015]
This CR protected the above instance (line 455 ) with synchronization
like in line 484 and in all other occurrences.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
-
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
