[jira] [Updated] (CASSANDRA-6747) MessagingService should handle failures on remote nodes.
[ https://issues.apache.org/jira/browse/CASSANDRA-6747?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Yuki Morishita updated CASSANDRA-6747: -- Attachment: 6747-v3.txt Thanks [~kohlisankalp], I updated your patch with following: * MessageOut object is immutable and MessageOut#withParameter returns new object, so we have to use that instead of original. * RTE throwed from ActiveRepairService#prepareForRepair has to be catched and notified to client so repair command not to hang. For remote snapshot fail, the patch certainly catches the error on coordinator side, but it still hangs(marked as TODO in RepairJob#sendTreeRequest). This is handled in CASSANDRA-6455. MessagingService should handle failures on remote nodes. Key: CASSANDRA-6747 URL: https://issues.apache.org/jira/browse/CASSANDRA-6747 Project: Cassandra Issue Type: Improvement Reporter: sankalp kohli Assignee: sankalp kohli Priority: Minor Labels: Core Fix For: 2.1 beta2 Attachments: 6747-v3.txt, CASSANDRA-6747-v2.diff, CASSANDRA-6747.diff While going through the code of MessagingService, I discovered that we don't handle callbacks on failure very well. If a Verb Handler on the remote machine throws an exception, it goes right through uncaught exception handler. The machine which triggered the message will keep waiting and will timeout. On timeout, it will so some stuff hard coded in the MS like hints and add to Latency. There is no way in IAsyncCallback to specify that to do on timeouts and also on failures. Here are some examples which I found will help if we enhance this system to also propagate failures back. So IAsyncCallback will have methods like onFailure. 1) From ActiveRepairService.prepareForRepair IAsyncCallback callback = new IAsyncCallback() { @Override public void response(MessageIn msg) { prepareLatch.countDown(); } @Override public boolean isLatencyForSnitch() { return false; } }; ListUUID cfIds = new ArrayList(columnFamilyStores.size()); for (ColumnFamilyStore cfs : columnFamilyStores) cfIds.add(cfs.metadata.cfId); for(InetAddress neighbour : endpoints) { PrepareMessage message = new PrepareMessage(parentRepairSession, cfIds, ranges); MessageOutRepairMessage msg = message.createMessage(); MessagingService.instance().sendRR(msg, neighbour, callback); } try { prepareLatch.await(1, TimeUnit.HOURS); } catch (InterruptedException e) { parentRepairSessions.remove(parentRepairSession); throw new RuntimeException(Did not get replies from all endpoints., e); } 2) During snapshot phase in repair, if SnapshotVerbHandler throws an exception, we will wait forever. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Updated] (CASSANDRA-6747) MessagingService should handle failures on remote nodes.
[ https://issues.apache.org/jira/browse/CASSANDRA-6747?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Yuki Morishita updated CASSANDRA-6747: -- Fix Version/s: 2.1 beta2 MessagingService should handle failures on remote nodes. Key: CASSANDRA-6747 URL: https://issues.apache.org/jira/browse/CASSANDRA-6747 Project: Cassandra Issue Type: Improvement Reporter: sankalp kohli Assignee: sankalp kohli Priority: Minor Labels: Core Fix For: 2.1 beta2 Attachments: CASSANDRA-6747.diff While going through the code of MessagingService, I discovered that we don't handle callbacks on failure very well. If a Verb Handler on the remote machine throws an exception, it goes right through uncaught exception handler. The machine which triggered the message will keep waiting and will timeout. On timeout, it will so some stuff hard coded in the MS like hints and add to Latency. There is no way in IAsyncCallback to specify that to do on timeouts and also on failures. Here are some examples which I found will help if we enhance this system to also propagate failures back. So IAsyncCallback will have methods like onFailure. 1) From ActiveRepairService.prepareForRepair IAsyncCallback callback = new IAsyncCallback() { @Override public void response(MessageIn msg) { prepareLatch.countDown(); } @Override public boolean isLatencyForSnitch() { return false; } }; ListUUID cfIds = new ArrayList(columnFamilyStores.size()); for (ColumnFamilyStore cfs : columnFamilyStores) cfIds.add(cfs.metadata.cfId); for(InetAddress neighbour : endpoints) { PrepareMessage message = new PrepareMessage(parentRepairSession, cfIds, ranges); MessageOutRepairMessage msg = message.createMessage(); MessagingService.instance().sendRR(msg, neighbour, callback); } try { prepareLatch.await(1, TimeUnit.HOURS); } catch (InterruptedException e) { parentRepairSessions.remove(parentRepairSession); throw new RuntimeException(Did not get replies from all endpoints., e); } 2) During snapshot phase in repair, if SnapshotVerbHandler throws an exception, we will wait forever. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Updated] (CASSANDRA-6747) MessagingService should handle failures on remote nodes.
[ https://issues.apache.org/jira/browse/CASSANDRA-6747?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Yuki Morishita updated CASSANDRA-6747: -- Reviewer: Yuki Morishita [~kohlisankalp] I like your approach. One thing you need to change is in SnapshotTask's callback#onFailure, you can't just throw RuntimeException, you have to call task.setException so repair knows there's exception during snapshotting. MessagingService should handle failures on remote nodes. Key: CASSANDRA-6747 URL: https://issues.apache.org/jira/browse/CASSANDRA-6747 Project: Cassandra Issue Type: Improvement Reporter: sankalp kohli Assignee: sankalp kohli Priority: Minor Labels: Core Fix For: 2.1 beta2 Attachments: CASSANDRA-6747.diff While going through the code of MessagingService, I discovered that we don't handle callbacks on failure very well. If a Verb Handler on the remote machine throws an exception, it goes right through uncaught exception handler. The machine which triggered the message will keep waiting and will timeout. On timeout, it will so some stuff hard coded in the MS like hints and add to Latency. There is no way in IAsyncCallback to specify that to do on timeouts and also on failures. Here are some examples which I found will help if we enhance this system to also propagate failures back. So IAsyncCallback will have methods like onFailure. 1) From ActiveRepairService.prepareForRepair IAsyncCallback callback = new IAsyncCallback() { @Override public void response(MessageIn msg) { prepareLatch.countDown(); } @Override public boolean isLatencyForSnitch() { return false; } }; ListUUID cfIds = new ArrayList(columnFamilyStores.size()); for (ColumnFamilyStore cfs : columnFamilyStores) cfIds.add(cfs.metadata.cfId); for(InetAddress neighbour : endpoints) { PrepareMessage message = new PrepareMessage(parentRepairSession, cfIds, ranges); MessageOutRepairMessage msg = message.createMessage(); MessagingService.instance().sendRR(msg, neighbour, callback); } try { prepareLatch.await(1, TimeUnit.HOURS); } catch (InterruptedException e) { parentRepairSessions.remove(parentRepairSession); throw new RuntimeException(Did not get replies from all endpoints., e); } 2) During snapshot phase in repair, if SnapshotVerbHandler throws an exception, we will wait forever. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Updated] (CASSANDRA-6747) MessagingService should handle failures on remote nodes.
[ https://issues.apache.org/jira/browse/CASSANDRA-6747?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] sankalp kohli updated CASSANDRA-6747: - Attachment: CASSANDRA-6747-v2.diff MessagingService should handle failures on remote nodes. Key: CASSANDRA-6747 URL: https://issues.apache.org/jira/browse/CASSANDRA-6747 Project: Cassandra Issue Type: Improvement Reporter: sankalp kohli Assignee: sankalp kohli Priority: Minor Labels: Core Fix For: 2.1 beta2 Attachments: CASSANDRA-6747-v2.diff, CASSANDRA-6747.diff While going through the code of MessagingService, I discovered that we don't handle callbacks on failure very well. If a Verb Handler on the remote machine throws an exception, it goes right through uncaught exception handler. The machine which triggered the message will keep waiting and will timeout. On timeout, it will so some stuff hard coded in the MS like hints and add to Latency. There is no way in IAsyncCallback to specify that to do on timeouts and also on failures. Here are some examples which I found will help if we enhance this system to also propagate failures back. So IAsyncCallback will have methods like onFailure. 1) From ActiveRepairService.prepareForRepair IAsyncCallback callback = new IAsyncCallback() { @Override public void response(MessageIn msg) { prepareLatch.countDown(); } @Override public boolean isLatencyForSnitch() { return false; } }; ListUUID cfIds = new ArrayList(columnFamilyStores.size()); for (ColumnFamilyStore cfs : columnFamilyStores) cfIds.add(cfs.metadata.cfId); for(InetAddress neighbour : endpoints) { PrepareMessage message = new PrepareMessage(parentRepairSession, cfIds, ranges); MessageOutRepairMessage msg = message.createMessage(); MessagingService.instance().sendRR(msg, neighbour, callback); } try { prepareLatch.await(1, TimeUnit.HOURS); } catch (InterruptedException e) { parentRepairSessions.remove(parentRepairSession); throw new RuntimeException(Did not get replies from all endpoints., e); } 2) During snapshot phase in repair, if SnapshotVerbHandler throws an exception, we will wait forever. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Updated] (CASSANDRA-6747) MessagingService should handle failures on remote nodes.
[ https://issues.apache.org/jira/browse/CASSANDRA-6747?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] sankalp kohli updated CASSANDRA-6747: - Attachment: CASSANDRA-6747.diff I am adding a new interface which could be used if we want negative ack back from the remote node or need a callback on timeout. This new interface is used in these two places. We can use this at other places but I have only updated two places.I can create new JIRA for these if that make more sense. SnapshotTask.java and ActiveRepairService.java This also fixes the problem in the previous comment for these use cases. MessagingService should handle failures on remote nodes. Key: CASSANDRA-6747 URL: https://issues.apache.org/jira/browse/CASSANDRA-6747 Project: Cassandra Issue Type: Improvement Reporter: sankalp kohli Priority: Minor Labels: Core Attachments: CASSANDRA-6747.diff While going through the code of MessagingService, I discovered that we don't handle callbacks on failure very well. If a Verb Handler on the remote machine throws an exception, it goes right through uncaught exception handler. The machine which triggered the message will keep waiting and will timeout. On timeout, it will so some stuff hard coded in the MS like hints and add to Latency. There is no way in IAsyncCallback to specify that to do on timeouts and also on failures. Here are some examples which I found will help if we enhance this system to also propagate failures back. So IAsyncCallback will have methods like onFailure. 1) From ActiveRepairService.prepareForRepair IAsyncCallback callback = new IAsyncCallback() { @Override public void response(MessageIn msg) { prepareLatch.countDown(); } @Override public boolean isLatencyForSnitch() { return false; } }; ListUUID cfIds = new ArrayList(columnFamilyStores.size()); for (ColumnFamilyStore cfs : columnFamilyStores) cfIds.add(cfs.metadata.cfId); for(InetAddress neighbour : endpoints) { PrepareMessage message = new PrepareMessage(parentRepairSession, cfIds, ranges); MessageOutRepairMessage msg = message.createMessage(); MessagingService.instance().sendRR(msg, neighbour, callback); } try { prepareLatch.await(1, TimeUnit.HOURS); } catch (InterruptedException e) { parentRepairSessions.remove(parentRepairSession); throw new RuntimeException(Did not get replies from all endpoints., e); } 2) During snapshot phase in repair, if SnapshotVerbHandler throws an exception, we will wait forever. -- This message was sent by Atlassian JIRA (v6.2#6252)