[JBoss JIRA] (JGRP-2288) S3_PING: Under certain conditions, subclusters fail to merge after network partition
by Nick Sawadsky (Jira)
[ https://issues.jboss.org/browse/JGRP-2288?page=com.atlassian.jira.plugin.... ]
Nick Sawadsky edited comment on JGRP-2288 at 12/13/18 3:24 PM:
---------------------------------------------------------------
[~belaban] I've tested this out with the latest code from the 3.6 branch using S3_PING -- it's working well. Thanks very much!
I'm thinking that in the typical case, the {{writeAll()}} call will not be made, because the coordinator will find its own address in the responses: {{data == null || !data.isCoord()}} will be false. So my hope is that there shouldn't be a noticeable increase in writes to the cloud store with this change.
To ensure we don't incur extra writes to the cloud store, would it make sense to also check that {{members}} is null before doing the write? If {{members}} is non-null, then the responses will be filtered and may not include this coordinator.
{code}
if(write_data_on_find && (remove_all_files_on_view_change || remove_old_coords_on_view_change)
&& !initial_discovery && members == null && is_coord && (data == null || !data.isCoord()) && !isInfoWriterRunning()) {
// a coordinator in a separate partition may have deleted this coordinator's file
writeAll();
}
{code}
Otherwise, whenever {{TP.fetchResponsesFromDiscoveryProtocol()}} is called, we may end up writing to the cloud store.
Similarly, maybe the {{if(responses.isEmpty())}} check should be {{if ((responses.isEmpty() && members == null)}}
was (Author: nsawadsky):
[~belaban] I've tested this out with the latest code from the 3.6 branch using S3_PING -- it's working well. Thanks very much!
I'm thinking that in the typical case, the {{writeAll()}} call will not be made, because the coordinator will find its own address in the responses: {{data == null || !data.isCoord()}} will be false. So my hope is that there shouldn't be a noticeable increase in writes to the cloud store with this change.
To ensure we don't incur extra writes to the cloud store, would it make sense to also check that {{members}} is null before doing the write? If {{members}} is non-null, then the responses will be filtered and may not include this coordinator.
{code}
if(write_data_on_find && (remove_all_files_on_view_change || remove_old_coords_on_view_change)
&& !initial_discovery && members == null && is_coord && (data == null || !data.isCoord()) && !isInfoWriterRunning()) {
// a coordinator in a separate partition may have deleted this coordinator's file
writeAll();
}
{code}
> S3_PING: Under certain conditions, subclusters fail to merge after network partition
> ------------------------------------------------------------------------------------
>
> Key: JGRP-2288
> URL: https://issues.jboss.org/browse/JGRP-2288
> Project: JGroups
> Issue Type: Bug
> Affects Versions: 3.6.15
> Reporter: Nick Sawadsky
> Assignee: Bela Ban
> Priority: Major
> Fix For: 4.0.17, 3.6.17
>
>
> Repro steps:
> 1. Set up a cluster of four nodes, two on one machine (Host 1) and two on another (Host 2). Let's call the nodes A, B, C, and D.
> 2. Configure all 4 nodes with S3_PING as the discovery mechanism. Set remove_all_files_on_view_change to true.
> 3. Start up nodes in the order A, B, C, D.
> 4. In the S3 bucket, there should be a single file with all four nodes listed. Node A should be flagged as the coordinator. Ensure that the UUID for node B is larger than the UUID for node C, when compared as two's complement integers. If this is not the case, shut down all nodes and restart in order. Repeat until the desired relationship is achieved. Note that with two's complement, a UUID having a first hex digit of 8 or higher is treated as negative for comparison purposes. So, for example, a UUID starting with 'a' is less than a UUID starting with 'b' which is less than a UUID starting with '1'.
> 5. On Host 1, use iptables to block all traffic going to and coming from Host 2.
> sudo iptables -A INPUT -s <Host 2 IP addr> -j DROP
> sudo iptables -A OUTPUT -d <Host 2 IP addr> -j DROP
> 6. Allow a few minutes for the nodes to detect the network partition. Eventually you should see two files in the S3 bucket.
> 7. Using Ctrl-C, stop node A.
> 8. You should soon find only a single file in the bucket, containing a single entry for node B. This is a result of the remove_all_files_on_view_change setting on S3_PING, which we set to true to avoid accumulation of old files in the bucket.
> 9. Resolve the network partition:
> sudo iptables -F OUTPUT
> sudo iptables -F INPUT
> 10. You will find that, even after many minutes, the subclusters are not merged.
> I believe the reason why the subclusters are never merged is as follows:
> - MERGE3 on nodes B, C and D uses S3_PING to find members to send INFO messages to. Each one finds only node B in the discovery file. As a result, only node B's view consistency checker has anything to work with.
> - On node B, the consistency checker can see that there are two coordinators, B and C. However, node C has a lower UUID, so node B defers to it to perform the merge. Node C never performs the merge because, as mentioned above, it is not receiving any INFO messages.
> I this this problem would affect FILE_PING as well, and other protocols derived from FILE_PING. Looking at the latest 4.x code, it appears the problem still exists there.
> I think the crux of the issue is that the coordinator on Host 2 (node C) does not re-create its discovery file after it is deleted by node B. Would it be reasonable for FILE_PING.findMembers() to create the discovery file if the node is a coordinator and the file doesn't exist?
--
This message was sent by Atlassian Jira
(v7.12.1#712002)
7 years, 7 months
[JBoss JIRA] (JGRP-2288) S3_PING: Under certain conditions, subclusters fail to merge after network partition
by Nick Sawadsky (Jira)
[ https://issues.jboss.org/browse/JGRP-2288?page=com.atlassian.jira.plugin.... ]
Nick Sawadsky edited comment on JGRP-2288 at 12/13/18 3:21 PM:
---------------------------------------------------------------
[~belaban] I've tested this out with the latest code from the 3.6 branch using S3_PING -- it's working well. Thanks very much!
I'm thinking that in the typical case, the {{writeAll()}} call will not be made, because the coordinator will find its own address in the responses: {{data == null || !data.isCoord()}} will be false. So my hope is that there shouldn't be a noticeable increase in writes to the cloud store with this change.
To ensure we don't incur extra writes to the cloud store, would it make sense to also check that {{members}} is null before doing the write? If {{members}} is non-null, then the responses will be filtered and may not include this coordinator.
{code}
if(write_data_on_find && (remove_all_files_on_view_change || remove_old_coords_on_view_change)
&& !initial_discovery && members == null && is_coord && (data == null || !data.isCoord()) && !isInfoWriterRunning()) {
// a coordinator in a separate partition may have deleted this coordinator's file
writeAll();
}
{code}
was (Author: nsawadsky):
[~belaban] I've tested this out with the latest code from the 3.6 branch using S3_PING -- it's working well. Thanks very much!
I'm thinking that in the typical case, the {{writeAll()}} call will not be made, because the coordinator will find its own address in the responses: {{data == null || !data.isCoord()}} will be false. So my hope is that there shouldn't be a noticeable increase in writes to the cloud store with this change.
To ensure we don't incur extra writes to the cloud store, would it make sense to also check that members is null before doing the write? If members is non-null, then the responses will be filtered and may not include this coordinator.
{code}
if(write_data_on_find && (remove_all_files_on_view_change || remove_old_coords_on_view_change)
&& !initial_discovery && members == null && is_coord && (data == null || !data.isCoord()) && !isInfoWriterRunning()) {
// a coordinator in a separate partition may have deleted this coordinator's file
writeAll();
}
{code}
> S3_PING: Under certain conditions, subclusters fail to merge after network partition
> ------------------------------------------------------------------------------------
>
> Key: JGRP-2288
> URL: https://issues.jboss.org/browse/JGRP-2288
> Project: JGroups
> Issue Type: Bug
> Affects Versions: 3.6.15
> Reporter: Nick Sawadsky
> Assignee: Bela Ban
> Priority: Major
> Fix For: 4.0.17, 3.6.17
>
>
> Repro steps:
> 1. Set up a cluster of four nodes, two on one machine (Host 1) and two on another (Host 2). Let's call the nodes A, B, C, and D.
> 2. Configure all 4 nodes with S3_PING as the discovery mechanism. Set remove_all_files_on_view_change to true.
> 3. Start up nodes in the order A, B, C, D.
> 4. In the S3 bucket, there should be a single file with all four nodes listed. Node A should be flagged as the coordinator. Ensure that the UUID for node B is larger than the UUID for node C, when compared as two's complement integers. If this is not the case, shut down all nodes and restart in order. Repeat until the desired relationship is achieved. Note that with two's complement, a UUID having a first hex digit of 8 or higher is treated as negative for comparison purposes. So, for example, a UUID starting with 'a' is less than a UUID starting with 'b' which is less than a UUID starting with '1'.
> 5. On Host 1, use iptables to block all traffic going to and coming from Host 2.
> sudo iptables -A INPUT -s <Host 2 IP addr> -j DROP
> sudo iptables -A OUTPUT -d <Host 2 IP addr> -j DROP
> 6. Allow a few minutes for the nodes to detect the network partition. Eventually you should see two files in the S3 bucket.
> 7. Using Ctrl-C, stop node A.
> 8. You should soon find only a single file in the bucket, containing a single entry for node B. This is a result of the remove_all_files_on_view_change setting on S3_PING, which we set to true to avoid accumulation of old files in the bucket.
> 9. Resolve the network partition:
> sudo iptables -F OUTPUT
> sudo iptables -F INPUT
> 10. You will find that, even after many minutes, the subclusters are not merged.
> I believe the reason why the subclusters are never merged is as follows:
> - MERGE3 on nodes B, C and D uses S3_PING to find members to send INFO messages to. Each one finds only node B in the discovery file. As a result, only node B's view consistency checker has anything to work with.
> - On node B, the consistency checker can see that there are two coordinators, B and C. However, node C has a lower UUID, so node B defers to it to perform the merge. Node C never performs the merge because, as mentioned above, it is not receiving any INFO messages.
> I this this problem would affect FILE_PING as well, and other protocols derived from FILE_PING. Looking at the latest 4.x code, it appears the problem still exists there.
> I think the crux of the issue is that the coordinator on Host 2 (node C) does not re-create its discovery file after it is deleted by node B. Would it be reasonable for FILE_PING.findMembers() to create the discovery file if the node is a coordinator and the file doesn't exist?
--
This message was sent by Atlassian Jira
(v7.12.1#712002)
7 years, 7 months
[JBoss JIRA] (JGRP-2288) S3_PING: Under certain conditions, subclusters fail to merge after network partition
by Nick Sawadsky (Jira)
[ https://issues.jboss.org/browse/JGRP-2288?page=com.atlassian.jira.plugin.... ]
Nick Sawadsky edited comment on JGRP-2288 at 12/13/18 3:01 PM:
---------------------------------------------------------------
[~belaban] I've tested this out with the latest code from the 3.6 branch using S3_PING -- it's working well. Thanks very much!
I'm thinking that in the typical case, the {{writeAll()}} call will not be made, because the coordinator will find its own address in the responses: {{data == null || !data.isCoord()}} will be false. So my hope is that there shouldn't be a noticeable increase in writes to the cloud store with this change.
To ensure we don't incur extra writes to the cloud store, would it make sense to also check that members is null before doing the write? If members is non-null, then the responses will be filtered and may not include this coordinator.
{code}
if(write_data_on_find && (remove_all_files_on_view_change || remove_old_coords_on_view_change)
&& !initial_discovery && members == null && is_coord && (data == null || !data.isCoord()) && !isInfoWriterRunning()) {
// a coordinator in a separate partition may have deleted this coordinator's file
writeAll();
}
{code}
was (Author: nsawadsky):
[~belaban] I've tested this out with the latest code from the 3.6 branch using S3_PING -- it's working well. Thanks very much!
I'm thinking that in the typical case, the {{writeAll()}} call will not be made, because the coordinator will find its own address in the responses: {{data == null || !data.isCoord()}} will be false. So my hope is that there shouldn't be a noticeable increase in writes to the cloud store with this change.
> S3_PING: Under certain conditions, subclusters fail to merge after network partition
> ------------------------------------------------------------------------------------
>
> Key: JGRP-2288
> URL: https://issues.jboss.org/browse/JGRP-2288
> Project: JGroups
> Issue Type: Bug
> Affects Versions: 3.6.15
> Reporter: Nick Sawadsky
> Assignee: Bela Ban
> Priority: Major
> Fix For: 4.0.17, 3.6.17
>
>
> Repro steps:
> 1. Set up a cluster of four nodes, two on one machine (Host 1) and two on another (Host 2). Let's call the nodes A, B, C, and D.
> 2. Configure all 4 nodes with S3_PING as the discovery mechanism. Set remove_all_files_on_view_change to true.
> 3. Start up nodes in the order A, B, C, D.
> 4. In the S3 bucket, there should be a single file with all four nodes listed. Node A should be flagged as the coordinator. Ensure that the UUID for node B is larger than the UUID for node C, when compared as two's complement integers. If this is not the case, shut down all nodes and restart in order. Repeat until the desired relationship is achieved. Note that with two's complement, a UUID having a first hex digit of 8 or higher is treated as negative for comparison purposes. So, for example, a UUID starting with 'a' is less than a UUID starting with 'b' which is less than a UUID starting with '1'.
> 5. On Host 1, use iptables to block all traffic going to and coming from Host 2.
> sudo iptables -A INPUT -s <Host 2 IP addr> -j DROP
> sudo iptables -A OUTPUT -d <Host 2 IP addr> -j DROP
> 6. Allow a few minutes for the nodes to detect the network partition. Eventually you should see two files in the S3 bucket.
> 7. Using Ctrl-C, stop node A.
> 8. You should soon find only a single file in the bucket, containing a single entry for node B. This is a result of the remove_all_files_on_view_change setting on S3_PING, which we set to true to avoid accumulation of old files in the bucket.
> 9. Resolve the network partition:
> sudo iptables -F OUTPUT
> sudo iptables -F INPUT
> 10. You will find that, even after many minutes, the subclusters are not merged.
> I believe the reason why the subclusters are never merged is as follows:
> - MERGE3 on nodes B, C and D uses S3_PING to find members to send INFO messages to. Each one finds only node B in the discovery file. As a result, only node B's view consistency checker has anything to work with.
> - On node B, the consistency checker can see that there are two coordinators, B and C. However, node C has a lower UUID, so node B defers to it to perform the merge. Node C never performs the merge because, as mentioned above, it is not receiving any INFO messages.
> I this this problem would affect FILE_PING as well, and other protocols derived from FILE_PING. Looking at the latest 4.x code, it appears the problem still exists there.
> I think the crux of the issue is that the coordinator on Host 2 (node C) does not re-create its discovery file after it is deleted by node B. Would it be reasonable for FILE_PING.findMembers() to create the discovery file if the node is a coordinator and the file doesn't exist?
--
This message was sent by Atlassian Jira
(v7.12.1#712002)
7 years, 7 months
[JBoss JIRA] (JGRP-2288) S3_PING: Under certain conditions, subclusters fail to merge after network partition
by Nick Sawadsky (Jira)
[ https://issues.jboss.org/browse/JGRP-2288?page=com.atlassian.jira.plugin.... ]
Nick Sawadsky edited comment on JGRP-2288 at 12/13/18 2:45 PM:
---------------------------------------------------------------
[~belaban] I've tested this out with the latest code from the 3.6 branch using S3_PING -- it's working well. Thanks very much!
I'm thinking that in the typical case, the {{writeAll()}} call will not be made, because the coordinator will find its own address in the responses: {{data == null || !data.isCoord()}} will be false. So my hope is that there shouldn't be a noticeable increase in writes to the cloud store with this change.
was (Author: nsawadsky):
[~belaban] I've tested this out with the latest code from the 3.6 branch using S3_PING -- it's working well. Thanks very much!
I'm thinking that in the typical case, the writeAll() call will not be made, because the coordinator will find its own address in the responses: {{data == null || !data.isCoord()}} will be false. So my hope is that there shouldn't be a noticeable increase in writes to the cloud store with this change.
> S3_PING: Under certain conditions, subclusters fail to merge after network partition
> ------------------------------------------------------------------------------------
>
> Key: JGRP-2288
> URL: https://issues.jboss.org/browse/JGRP-2288
> Project: JGroups
> Issue Type: Bug
> Affects Versions: 3.6.15
> Reporter: Nick Sawadsky
> Assignee: Bela Ban
> Priority: Major
> Fix For: 4.0.17, 3.6.17
>
>
> Repro steps:
> 1. Set up a cluster of four nodes, two on one machine (Host 1) and two on another (Host 2). Let's call the nodes A, B, C, and D.
> 2. Configure all 4 nodes with S3_PING as the discovery mechanism. Set remove_all_files_on_view_change to true.
> 3. Start up nodes in the order A, B, C, D.
> 4. In the S3 bucket, there should be a single file with all four nodes listed. Node A should be flagged as the coordinator. Ensure that the UUID for node B is larger than the UUID for node C, when compared as two's complement integers. If this is not the case, shut down all nodes and restart in order. Repeat until the desired relationship is achieved. Note that with two's complement, a UUID having a first hex digit of 8 or higher is treated as negative for comparison purposes. So, for example, a UUID starting with 'a' is less than a UUID starting with 'b' which is less than a UUID starting with '1'.
> 5. On Host 1, use iptables to block all traffic going to and coming from Host 2.
> sudo iptables -A INPUT -s <Host 2 IP addr> -j DROP
> sudo iptables -A OUTPUT -d <Host 2 IP addr> -j DROP
> 6. Allow a few minutes for the nodes to detect the network partition. Eventually you should see two files in the S3 bucket.
> 7. Using Ctrl-C, stop node A.
> 8. You should soon find only a single file in the bucket, containing a single entry for node B. This is a result of the remove_all_files_on_view_change setting on S3_PING, which we set to true to avoid accumulation of old files in the bucket.
> 9. Resolve the network partition:
> sudo iptables -F OUTPUT
> sudo iptables -F INPUT
> 10. You will find that, even after many minutes, the subclusters are not merged.
> I believe the reason why the subclusters are never merged is as follows:
> - MERGE3 on nodes B, C and D uses S3_PING to find members to send INFO messages to. Each one finds only node B in the discovery file. As a result, only node B's view consistency checker has anything to work with.
> - On node B, the consistency checker can see that there are two coordinators, B and C. However, node C has a lower UUID, so node B defers to it to perform the merge. Node C never performs the merge because, as mentioned above, it is not receiving any INFO messages.
> I this this problem would affect FILE_PING as well, and other protocols derived from FILE_PING. Looking at the latest 4.x code, it appears the problem still exists there.
> I think the crux of the issue is that the coordinator on Host 2 (node C) does not re-create its discovery file after it is deleted by node B. Would it be reasonable for FILE_PING.findMembers() to create the discovery file if the node is a coordinator and the file doesn't exist?
--
This message was sent by Atlassian Jira
(v7.12.1#712002)
7 years, 7 months
[JBoss JIRA] (JGRP-2288) S3_PING: Under certain conditions, subclusters fail to merge after network partition
by Nick Sawadsky (Jira)
[ https://issues.jboss.org/browse/JGRP-2288?page=com.atlassian.jira.plugin.... ]
Nick Sawadsky edited comment on JGRP-2288 at 12/13/18 2:43 PM:
---------------------------------------------------------------
[~belaban] I've tested this out with the latest code from the 3.6 branch using S3_PING -- it's working well. Thanks very much!
I'm thinking that in the typical case, the writeAll() call will not be made, because the coordinator will find its own address in the responses: {{data == null || !data.isCoord()}} will be false. So my hope is that there shouldn't be a noticeable increase in writes to the cloud store with this change.
was (Author: nsawadsky):
[~belaban] I've tested this out with the latest code from the 3.6 branch using S3_PING -- it's working well. Thanks very much!
I'm thinking that in the typical case, the writeAll() call will not be made, because the coordinator will find its own address in the responses: {{ (data == null || !data.isCoord()) }} will be false. So my hope is that there shouldn't be a noticeable increase in writes to the cloud store with this change.
> S3_PING: Under certain conditions, subclusters fail to merge after network partition
> ------------------------------------------------------------------------------------
>
> Key: JGRP-2288
> URL: https://issues.jboss.org/browse/JGRP-2288
> Project: JGroups
> Issue Type: Bug
> Affects Versions: 3.6.15
> Reporter: Nick Sawadsky
> Assignee: Bela Ban
> Priority: Major
> Fix For: 4.0.17, 3.6.17
>
>
> Repro steps:
> 1. Set up a cluster of four nodes, two on one machine (Host 1) and two on another (Host 2). Let's call the nodes A, B, C, and D.
> 2. Configure all 4 nodes with S3_PING as the discovery mechanism. Set remove_all_files_on_view_change to true.
> 3. Start up nodes in the order A, B, C, D.
> 4. In the S3 bucket, there should be a single file with all four nodes listed. Node A should be flagged as the coordinator. Ensure that the UUID for node B is larger than the UUID for node C, when compared as two's complement integers. If this is not the case, shut down all nodes and restart in order. Repeat until the desired relationship is achieved. Note that with two's complement, a UUID having a first hex digit of 8 or higher is treated as negative for comparison purposes. So, for example, a UUID starting with 'a' is less than a UUID starting with 'b' which is less than a UUID starting with '1'.
> 5. On Host 1, use iptables to block all traffic going to and coming from Host 2.
> sudo iptables -A INPUT -s <Host 2 IP addr> -j DROP
> sudo iptables -A OUTPUT -d <Host 2 IP addr> -j DROP
> 6. Allow a few minutes for the nodes to detect the network partition. Eventually you should see two files in the S3 bucket.
> 7. Using Ctrl-C, stop node A.
> 8. You should soon find only a single file in the bucket, containing a single entry for node B. This is a result of the remove_all_files_on_view_change setting on S3_PING, which we set to true to avoid accumulation of old files in the bucket.
> 9. Resolve the network partition:
> sudo iptables -F OUTPUT
> sudo iptables -F INPUT
> 10. You will find that, even after many minutes, the subclusters are not merged.
> I believe the reason why the subclusters are never merged is as follows:
> - MERGE3 on nodes B, C and D uses S3_PING to find members to send INFO messages to. Each one finds only node B in the discovery file. As a result, only node B's view consistency checker has anything to work with.
> - On node B, the consistency checker can see that there are two coordinators, B and C. However, node C has a lower UUID, so node B defers to it to perform the merge. Node C never performs the merge because, as mentioned above, it is not receiving any INFO messages.
> I this this problem would affect FILE_PING as well, and other protocols derived from FILE_PING. Looking at the latest 4.x code, it appears the problem still exists there.
> I think the crux of the issue is that the coordinator on Host 2 (node C) does not re-create its discovery file after it is deleted by node B. Would it be reasonable for FILE_PING.findMembers() to create the discovery file if the node is a coordinator and the file doesn't exist?
--
This message was sent by Atlassian Jira
(v7.12.1#712002)
7 years, 7 months
[JBoss JIRA] (JGRP-2288) S3_PING: Under certain conditions, subclusters fail to merge after network partition
by Nick Sawadsky (Jira)
[ https://issues.jboss.org/browse/JGRP-2288?page=com.atlassian.jira.plugin.... ]
Nick Sawadsky edited comment on JGRP-2288 at 12/13/18 2:43 PM:
---------------------------------------------------------------
[~belaban] I've tested this out with the latest code from the 3.6 branch using S3_PING -- it's working well. Thanks very much!
I'm thinking that in the typical case, the writeAll() call will not be made, because the coordinator will find its own address in the responses: {{ (data == null || !data.isCoord()) }} will be false. So my hope is that there shouldn't be a noticeable increase in writes to the cloud store with this change.
was (Author: nsawadsky):
[~belaban] I've tested this out with the latest code from the 3.6 branch using S3_PING -- it's working well. Thanks very much!
I'm thinking that in the typical case, the writeAll() call will not be made, because the coordinator will find its own address in the responses -- {{data == null || !data.isCoord()}} will be false. So my hope is that there shouldn't be a noticeable increase in writes to the cloud store with this change.
> S3_PING: Under certain conditions, subclusters fail to merge after network partition
> ------------------------------------------------------------------------------------
>
> Key: JGRP-2288
> URL: https://issues.jboss.org/browse/JGRP-2288
> Project: JGroups
> Issue Type: Bug
> Affects Versions: 3.6.15
> Reporter: Nick Sawadsky
> Assignee: Bela Ban
> Priority: Major
> Fix For: 4.0.17, 3.6.17
>
>
> Repro steps:
> 1. Set up a cluster of four nodes, two on one machine (Host 1) and two on another (Host 2). Let's call the nodes A, B, C, and D.
> 2. Configure all 4 nodes with S3_PING as the discovery mechanism. Set remove_all_files_on_view_change to true.
> 3. Start up nodes in the order A, B, C, D.
> 4. In the S3 bucket, there should be a single file with all four nodes listed. Node A should be flagged as the coordinator. Ensure that the UUID for node B is larger than the UUID for node C, when compared as two's complement integers. If this is not the case, shut down all nodes and restart in order. Repeat until the desired relationship is achieved. Note that with two's complement, a UUID having a first hex digit of 8 or higher is treated as negative for comparison purposes. So, for example, a UUID starting with 'a' is less than a UUID starting with 'b' which is less than a UUID starting with '1'.
> 5. On Host 1, use iptables to block all traffic going to and coming from Host 2.
> sudo iptables -A INPUT -s <Host 2 IP addr> -j DROP
> sudo iptables -A OUTPUT -d <Host 2 IP addr> -j DROP
> 6. Allow a few minutes for the nodes to detect the network partition. Eventually you should see two files in the S3 bucket.
> 7. Using Ctrl-C, stop node A.
> 8. You should soon find only a single file in the bucket, containing a single entry for node B. This is a result of the remove_all_files_on_view_change setting on S3_PING, which we set to true to avoid accumulation of old files in the bucket.
> 9. Resolve the network partition:
> sudo iptables -F OUTPUT
> sudo iptables -F INPUT
> 10. You will find that, even after many minutes, the subclusters are not merged.
> I believe the reason why the subclusters are never merged is as follows:
> - MERGE3 on nodes B, C and D uses S3_PING to find members to send INFO messages to. Each one finds only node B in the discovery file. As a result, only node B's view consistency checker has anything to work with.
> - On node B, the consistency checker can see that there are two coordinators, B and C. However, node C has a lower UUID, so node B defers to it to perform the merge. Node C never performs the merge because, as mentioned above, it is not receiving any INFO messages.
> I this this problem would affect FILE_PING as well, and other protocols derived from FILE_PING. Looking at the latest 4.x code, it appears the problem still exists there.
> I think the crux of the issue is that the coordinator on Host 2 (node C) does not re-create its discovery file after it is deleted by node B. Would it be reasonable for FILE_PING.findMembers() to create the discovery file if the node is a coordinator and the file doesn't exist?
--
This message was sent by Atlassian Jira
(v7.12.1#712002)
7 years, 7 months
[JBoss JIRA] (JGRP-2288) S3_PING: Under certain conditions, subclusters fail to merge after network partition
by Nick Sawadsky (Jira)
[ https://issues.jboss.org/browse/JGRP-2288?page=com.atlassian.jira.plugin.... ]
Nick Sawadsky edited comment on JGRP-2288 at 12/13/18 2:42 PM:
---------------------------------------------------------------
[~belaban] I've tested this out with the latest code from the 3.6 branch using S3_PING -- it's working well. Thanks very much!
I'm thinking that in the typical case, the writeAll() call will not be made, because the coordinator will find its own address in the responses -- {{data == null || !data.isCoord()}} will be false. So my hope is that there shouldn't be a noticeable increase in writes to the cloud store with this change.
was (Author: nsawadsky):
[~belaban] I've tested this out with the latest code from the 3.6 branch using S3_PING -- it's working well. Thanks very much!
I'm thinking that in the typical case, the writeAll() call will not be made, because the coordinator will find its own address in the responses -- {{(data == null || !data.isCoord()})} will be false. So my hope is that there shouldn't be a noticeable increase in writes to the cloud store with this change.
> S3_PING: Under certain conditions, subclusters fail to merge after network partition
> ------------------------------------------------------------------------------------
>
> Key: JGRP-2288
> URL: https://issues.jboss.org/browse/JGRP-2288
> Project: JGroups
> Issue Type: Bug
> Affects Versions: 3.6.15
> Reporter: Nick Sawadsky
> Assignee: Bela Ban
> Priority: Major
> Fix For: 4.0.17, 3.6.17
>
>
> Repro steps:
> 1. Set up a cluster of four nodes, two on one machine (Host 1) and two on another (Host 2). Let's call the nodes A, B, C, and D.
> 2. Configure all 4 nodes with S3_PING as the discovery mechanism. Set remove_all_files_on_view_change to true.
> 3. Start up nodes in the order A, B, C, D.
> 4. In the S3 bucket, there should be a single file with all four nodes listed. Node A should be flagged as the coordinator. Ensure that the UUID for node B is larger than the UUID for node C, when compared as two's complement integers. If this is not the case, shut down all nodes and restart in order. Repeat until the desired relationship is achieved. Note that with two's complement, a UUID having a first hex digit of 8 or higher is treated as negative for comparison purposes. So, for example, a UUID starting with 'a' is less than a UUID starting with 'b' which is less than a UUID starting with '1'.
> 5. On Host 1, use iptables to block all traffic going to and coming from Host 2.
> sudo iptables -A INPUT -s <Host 2 IP addr> -j DROP
> sudo iptables -A OUTPUT -d <Host 2 IP addr> -j DROP
> 6. Allow a few minutes for the nodes to detect the network partition. Eventually you should see two files in the S3 bucket.
> 7. Using Ctrl-C, stop node A.
> 8. You should soon find only a single file in the bucket, containing a single entry for node B. This is a result of the remove_all_files_on_view_change setting on S3_PING, which we set to true to avoid accumulation of old files in the bucket.
> 9. Resolve the network partition:
> sudo iptables -F OUTPUT
> sudo iptables -F INPUT
> 10. You will find that, even after many minutes, the subclusters are not merged.
> I believe the reason why the subclusters are never merged is as follows:
> - MERGE3 on nodes B, C and D uses S3_PING to find members to send INFO messages to. Each one finds only node B in the discovery file. As a result, only node B's view consistency checker has anything to work with.
> - On node B, the consistency checker can see that there are two coordinators, B and C. However, node C has a lower UUID, so node B defers to it to perform the merge. Node C never performs the merge because, as mentioned above, it is not receiving any INFO messages.
> I this this problem would affect FILE_PING as well, and other protocols derived from FILE_PING. Looking at the latest 4.x code, it appears the problem still exists there.
> I think the crux of the issue is that the coordinator on Host 2 (node C) does not re-create its discovery file after it is deleted by node B. Would it be reasonable for FILE_PING.findMembers() to create the discovery file if the node is a coordinator and the file doesn't exist?
--
This message was sent by Atlassian Jira
(v7.12.1#712002)
7 years, 7 months
[JBoss JIRA] (JGRP-2288) S3_PING: Under certain conditions, subclusters fail to merge after network partition
by Nick Sawadsky (Jira)
[ https://issues.jboss.org/browse/JGRP-2288?page=com.atlassian.jira.plugin.... ]
Nick Sawadsky commented on JGRP-2288:
-------------------------------------
[~belaban] I've tested this out with the latest code from the 3.6 branch using S3_PING -- it's working well. Thanks very much!
I'm thinking that in the typical case, the writeAll() call will not be made, because the coordinator will find its own address in the responses -- {{(data == null || !data.isCoord()})} will be false. So my hope is that there shouldn't be a noticeable increase in writes to the cloud store with this change.
> S3_PING: Under certain conditions, subclusters fail to merge after network partition
> ------------------------------------------------------------------------------------
>
> Key: JGRP-2288
> URL: https://issues.jboss.org/browse/JGRP-2288
> Project: JGroups
> Issue Type: Bug
> Affects Versions: 3.6.15
> Reporter: Nick Sawadsky
> Assignee: Bela Ban
> Priority: Major
> Fix For: 4.0.17, 3.6.17
>
>
> Repro steps:
> 1. Set up a cluster of four nodes, two on one machine (Host 1) and two on another (Host 2). Let's call the nodes A, B, C, and D.
> 2. Configure all 4 nodes with S3_PING as the discovery mechanism. Set remove_all_files_on_view_change to true.
> 3. Start up nodes in the order A, B, C, D.
> 4. In the S3 bucket, there should be a single file with all four nodes listed. Node A should be flagged as the coordinator. Ensure that the UUID for node B is larger than the UUID for node C, when compared as two's complement integers. If this is not the case, shut down all nodes and restart in order. Repeat until the desired relationship is achieved. Note that with two's complement, a UUID having a first hex digit of 8 or higher is treated as negative for comparison purposes. So, for example, a UUID starting with 'a' is less than a UUID starting with 'b' which is less than a UUID starting with '1'.
> 5. On Host 1, use iptables to block all traffic going to and coming from Host 2.
> sudo iptables -A INPUT -s <Host 2 IP addr> -j DROP
> sudo iptables -A OUTPUT -d <Host 2 IP addr> -j DROP
> 6. Allow a few minutes for the nodes to detect the network partition. Eventually you should see two files in the S3 bucket.
> 7. Using Ctrl-C, stop node A.
> 8. You should soon find only a single file in the bucket, containing a single entry for node B. This is a result of the remove_all_files_on_view_change setting on S3_PING, which we set to true to avoid accumulation of old files in the bucket.
> 9. Resolve the network partition:
> sudo iptables -F OUTPUT
> sudo iptables -F INPUT
> 10. You will find that, even after many minutes, the subclusters are not merged.
> I believe the reason why the subclusters are never merged is as follows:
> - MERGE3 on nodes B, C and D uses S3_PING to find members to send INFO messages to. Each one finds only node B in the discovery file. As a result, only node B's view consistency checker has anything to work with.
> - On node B, the consistency checker can see that there are two coordinators, B and C. However, node C has a lower UUID, so node B defers to it to perform the merge. Node C never performs the merge because, as mentioned above, it is not receiving any INFO messages.
> I this this problem would affect FILE_PING as well, and other protocols derived from FILE_PING. Looking at the latest 4.x code, it appears the problem still exists there.
> I think the crux of the issue is that the coordinator on Host 2 (node C) does not re-create its discovery file after it is deleted by node B. Would it be reasonable for FILE_PING.findMembers() to create the discovery file if the node is a coordinator and the file doesn't exist?
--
This message was sent by Atlassian Jira
(v7.12.1#712002)
7 years, 7 months
[JBoss JIRA] (WFLY-9529) Using injected JMS in a background task/thread leads to NameNotFoundException: java:comp/TransactionSynchronizationRegistry
by Cody Lerum (Jira)
[ https://issues.jboss.org/browse/WFLY-9529?page=com.atlassian.jira.plugin.... ]
Cody Lerum commented on WFLY-9529:
----------------------------------
Temp workaround for me is as follows
*Original*
{code:java}
import java.io.Serializable;
import javax.annotation.Resource;
import javax.inject.Inject;
import javax.jms.JMSContext;
import javax.jms.Topic;
import javax.transaction.Transactional;
import javax.transaction.Transactional.TxType;
public class EntityUpdatedMessageSender implements Serializable {
private static final long serialVersionUID = 1L;
@Inject
private JMSContext jmsContext;
@Resource(lookup = JMSResources.ENTITY_UPDATED_TOPIC)
private Topic topic;
@Transactional(value = TxType.REQUIRES_NEW)
public void send(VersionedEntity entity) {
jmsContext.createProducer().send(topic, new EntityUpdate(entity));
}
}
{code}
*Workaround*
{code:java}
import java.io.Serializable;
import javax.annotation.Resource;
import javax.jms.ConnectionFactory;
import javax.jms.JMSContext;
import javax.jms.Topic;
import javax.transaction.Transactional;
import javax.transaction.Transactional.TxType;
public class EntityUpdatedMessageSender implements Serializable {
private static final long serialVersionUID = 1L;
@Resource
private ConnectionFactory connectionFactory;
@Resource(lookup = JMSResources.ENTITY_UPDATED_TOPIC)
private Topic topic;
@Transactional(value = TxType.REQUIRES_NEW)
public void send(VersionedEntity entity) {
try (JMSContext context = connectionFactory.createContext()) {
context.createProducer().send(topic, new EntityUpdate(entity));
}
}
}
{code}
> Using injected JMS in a background task/thread leads to NameNotFoundException: java:comp/TransactionSynchronizationRegistry
> ---------------------------------------------------------------------------------------------------------------------------
>
> Key: WFLY-9529
> URL: https://issues.jboss.org/browse/WFLY-9529
> Project: WildFly
> Issue Type: Bug
> Components: CDI / Weld, JMS, Naming
> Affects Versions: 14.0.1.Final, 10.1.0.Final, 11.0.0.Final, 12.0.0.Final, 13.0.0.Final
> Environment: Running on Windows 10, Java 64-bit 1.8.0_131
> Reporter: Scott Van Wart
> Assignee: Eduardo Martins
> Priority: Major
> Labels: ActiveMQ, jms, transaction
> Attachments: injected-jms.zip, injected-jms2.zip, wildfly-11-injected-jms.txt
>
>
> If I try to use an @Injected JMSContext while executing within a background task (ManagedExecutorService) or thread (ManagedThreadFactory), I get the attached stacktrace. I've experienced this a number of times with Wildfly 10.1.0, including messages sent in Infinispan's expiry task thread.
> My original workaround was to submit an additional task on a separate thread to send the message, then wait for it to complete. That seemed unreliable (sometimes it would still produce NameNotFoundException). I've resorted to creating my own JMSContext by using @Resource( lookup="java:/ConnectionFactory" ) and sending messages that way. Both workarounds prevent the message sending logic from participating in any ongoing transactions.
> I'm also attaching a sample EAR project. It can be built with Maven. It creates a background task that waits 3 seconds and then tries to send a JMS message in a new transaction using an injected JMS context. I use the standalone-full.xml profile to run it.
--
This message was sent by Atlassian Jira
(v7.12.1#712002)
7 years, 7 months
[JBoss JIRA] (WFLY-11349) Tests in org.jboss.as.test.clustering.cluster fails with security manager
by Radoslav Husar (Jira)
[ https://issues.jboss.org/browse/WFLY-11349?page=com.atlassian.jira.plugin... ]
Radoslav Husar commented on WFLY-11349:
---------------------------------------
An Infinispan upgrade will be needed to address some of these, see WFLY-11511.
> Tests in org.jboss.as.test.clustering.cluster fails with security manager
> -------------------------------------------------------------------------
>
> Key: WFLY-11349
> URL: https://issues.jboss.org/browse/WFLY-11349
> Project: WildFly
> Issue Type: Bug
> Components: Clustering, Test Suite
> Affects Versions: 14.0.0.Final
> Reporter: Martin Choma
> Assignee: Radoslav Husar
> Priority: Major
> Labels: security-manager
>
> {noformat}
> org.jboss.as.test.clustering.cluster.cdi (1)
> CdiFailoverTestCase.test
> org.jboss.as.test.clustering.cluster.jsf (2)
> JSFFailoverTestCase.testGracefulSimpleFailover
> JSFFailoverTestCase.testGracefulUndeployFailover
> org.jboss.as.test.clustering.cluster.provider (1)
> ServiceProviderRegistrationTestCase.test
> org.jboss.as.test.clustering.cluster.singleton (4)
> SingletonBackupServiceTestCase.testSingletonService
> SingletonDeploymentDescriptorTestCase.test
> SingletonDeploymentJBossAllTestCase.test
> SingletonPartitionTestCase.testSingletonService
> org.jboss.as.test.clustering.cluster.xsite (1)
> XSiteSimpleTestCase.test
> {noformat}
> Seems to me doPrivileged block is missing in server code somewhere. There are errors like
> {noformat}
> Caused by: java.security.AccessControlException: WFSM000001: Permission check failed (permission "("java.lang.RuntimePermission" "accessClassInPackage.sun.reflect")" in code source "(vfs:/content/CdiFailoverTestCase.war/WEB-INF/classes <no signer certificates>)" of "ModuleClassLoader for Module "deployment.CdiFailoverTestCase.war" from Service Module Loader")
> at org.wildfly.security.manager.WildFlySecurityManager.checkPermission(WildFlySecurityManager.java:294)
> at org.wildfly.security.manager.WildFlySecurityManager.checkPermission(WildFlySecurityManager.java:191)
> at java.lang.SecurityManager.checkPackageAccess(SecurityManager.java:1564)
> at org.wildfly.security.manager.WildFlySecurityManager.checkPackageAccess(WildFlySecurityManager.java:490)
> at sun.misc.Launcher$AppClassLoader.loadClass(Launcher.java:318)
> at java.lang.ClassLoader.loadClass(ClassLoader.java:357)
> at org.jboss.modules.JDKSpecific.getSystemClass(JDKSpecific.java:180)
> at org.jboss.modules.ConcurrentClassLoader.performLoadClass(ConcurrentClassLoader.java:395)
> at org.jboss.modules.ConcurrentClassLoader.loadClass(ConcurrentClassLoader.java:116)
> at org.jboss.marshalling.river.Protocol.<clinit>(Protocol.java:310)
> Caused by: java.security.AccessControlException: WFSM000001: Permission check failed (permission \"(\"java.lang.RuntimePermission\" \"accessDeclaredMembers\")\" in code source \"(vfs:/content/XSiteSimpleTestCase.war/WEB-INF/classes <no signer certificates>)\" of \"ModuleClassLoader for Module \"deployment.XSiteSimpleTestCase.war\" from Service Module Loader\")"}}}}
> Caused by: java.lang.ExceptionInInitializerError
> at org.jboss.marshalling.river.RiverMarshaller.<clinit>(RiverMarshaller.java:1274)
> {noformat}
> [1] https://ci.wildfly.org/viewLog.html?buildId=128138&buildTypeId=WF_MasterS...
--
This message was sent by Atlassian Jira
(v7.12.1#712002)
7 years, 7 months