[jira] [Commented] (ARTEMIS-2678) Incomplete records for pages under high load
[ https://issues.apache.org/jira/browse/ARTEMIS-2678?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17065774#comment-17065774 ] Francesco Nigro commented on ARTEMIS-2678: -- {quote}1) The broker is running on all our stages with the same version. I guess that using the latest upstream is nothing I would want to run on any setup but a local one.{quote} We have had some changes on OpenWire and paging so it's worth to check IMO {quote}4) So what do you suggest for max-size-bytes and page-size-bytes? The default is using 50% of max heap and 10Mb for page.{quote} Depends on your requirements in term of number of addresses and your global-max-size (that's by default is half of your max heap size). > Incomplete records for pages under high load > > > Key: ARTEMIS-2678 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2678 > Project: ActiveMQ Artemis > Issue Type: Bug > Components: Broker > Environment: Linux >Reporter: Ansgar J. Sachs >Priority: Major > Attachments: Bildschirmfoto 2020-03-23 um 17.35.41.png > > > {quote}As developer, I expect paging to be resource saving and resilient to > high load{quote} > h3. Current Situation > During a performance test with a throughput of ~25.000 messages per second > that run mulitple hours, some consumers were too slow to consume and messages > piled up on the broker. For this reason, the broker started to page the > messages of growing queues. When we reduced the load from the broker, some > queues stopped consuming due to the following logs: > {code} > AMQ222033: Page file 7.page had incomplete records at position > 39,795,399 at record number 13,952 > target message no.16146 not found from start offset 46032883 and start > message number 16146: java.lang.RuntimeException: target message no.16146 not > found from start offset 46032883 and start message number 16146 > {code} > It wasnt possible to recover from this state but deleting the paging > directory. > h3. Expected Situation > I would expect that the paging mechanism is resilient to any errors. > h3. Scenario Setup > Master configuration: > {code:xml} > > > > true > > > > > > 256Mb > 64Mb > > 10 > PAGE > > {code} > An extract of the monitoring of the Performance Test is attached to this > issue. > h3. Workaround > Right now we disabled paging at all and only use our Heap. However, the heap > is exhausted at 5 million messages which is in our use case better than > loosing any of them. -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (ARTEMIS-2678) Incomplete records for pages under high load
[ https://issues.apache.org/jira/browse/ARTEMIS-2678?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17065475#comment-17065475 ] Ansgar J. Sachs commented on ARTEMIS-2678: -- 1) The broker is running on all our stages with the same version. I guess that using the latest upstream is nothing I would want to run on any setup but a local one. 2) The OpenWire protocol is something caused by our old Jboss EAP which prevents us from upgrading to a more up-to-date RA (like Artemis) 3) I will checkout the read-whole-page configuration and let you know about the results 4) So what do you suggest for max-size-bytes and page-size-bytes? The default is using 50% of max heap and 20Mb for page. However, at least this 50% boundary is way to high if you have more than 1 queue or do I misunderstand this setting? > Incomplete records for pages under high load > > > Key: ARTEMIS-2678 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2678 > Project: ActiveMQ Artemis > Issue Type: Bug > Components: Broker > Environment: Linux >Reporter: Ansgar J. Sachs >Priority: Major > Attachments: Bildschirmfoto 2020-03-23 um 17.35.41.png > > > {quote}As developer, I expect paging to be resource saving and resilient to > high load{quote} > h3. Current Situation > During a performance test with a throughput of ~25.000 messages per second > that run mulitple hours, some consumers were too slow to consume and messages > piled up on the broker. For this reason, the broker started to page the > messages of growing queues. When we reduced the load from the broker, some > queues stopped consuming due to the following logs: > {code} > AMQ222033: Page file 7.page had incomplete records at position > 39,795,399 at record number 13,952 > target message no.16146 not found from start offset 46032883 and start > message number 16146: java.lang.RuntimeException: target message no.16146 not > found from start offset 46032883 and start message number 16146 > {code} > It wasnt possible to recover from this state but deleting the paging > directory. > h3. Expected Situation > I would expect that the paging mechanism is resilient to any errors. > h3. Scenario Setup > Master configuration: > {code:xml} > > > > true > > > > > > 256Mb > 64Mb > > 10 > PAGE > > {code} > An extract of the monitoring of the Performance Test is attached to this > issue. > h3. Workaround > Right now we disabled paging at all and only use our Heap. However, the heap > is exhausted at 5 million messages which is in our use case better than > loosing any of them. -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (ARTEMIS-2678) Incomplete records for pages under high load
[ https://issues.apache.org/jira/browse/ARTEMIS-2678?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17065440#comment-17065440 ] Francesco Nigro commented on ARTEMIS-2678: -- I suggests several things to do: * try with the latest master upstream * try switch to core protocol (if you can do it) * try configure read-whole-page true (default is false) Other (separate) questions: why you've chosen 64Mb of page size and just 256Mb of max size bytes? > Incomplete records for pages under high load > > > Key: ARTEMIS-2678 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2678 > Project: ActiveMQ Artemis > Issue Type: Bug > Components: Broker > Environment: Linux >Reporter: Ansgar J. Sachs >Priority: Major > Attachments: Bildschirmfoto 2020-03-23 um 17.35.41.png > > > {quote}As developer, I expect paging to be resource saving and resilient to > high load{quote} > h3. Current Situation > During a performance test with a throughput of ~25.000 messages per second > that run mulitple hours, some consumers were too slow to consume and messages > piled up on the broker. For this reason, the broker started to page the > messages of growing queues. When we reduced the load from the broker, some > queues stopped consuming due to the following logs: > {code} > AMQ222033: Page file 7.page had incomplete records at position > 39,795,399 at record number 13,952 > target message no.16146 not found from start offset 46032883 and start > message number 16146: java.lang.RuntimeException: target message no.16146 not > found from start offset 46032883 and start message number 16146 > {code} > It wasnt possible to recover from this state but deleting the paging > directory. > h3. Expected Situation > I would expect that the paging mechanism is resilient to any errors. > h3. Scenario Setup > Master configuration: > {code:xml} > > > > true > > > > > > 256Mb > 64Mb > > 10 > PAGE > > {code} > An extract of the monitoring of the Performance Test is attached to this > issue. > h3. Workaround > Right now we disabled paging at all and only use our Heap. However, the heap > is exhausted at 5 million messages which is in our use case better than > loosing any of them. -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (ARTEMIS-2678) Incomplete records for pages under high load
[ https://issues.apache.org/jira/browse/ARTEMIS-2678?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17064970#comment-17064970 ] Ansgar J. Sachs commented on ARTEMIS-2678: -- - I am using version 2.11.0 of Artemis with OpenWire as protocol. - We are using a shared-store after we had some troubles with Replication and XA-Transactions. - For our test setup we had two JVMs running on one node - so just separated by process > Incomplete records for pages under high load > > > Key: ARTEMIS-2678 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2678 > Project: ActiveMQ Artemis > Issue Type: Bug > Components: Broker > Environment: Linux >Reporter: Ansgar J. Sachs >Priority: Major > Attachments: Bildschirmfoto 2020-03-23 um 17.35.41.png > > > {quote}As developer, I expect paging to be resource saving and resilient to > high load{quote} > h3. Current Situation > During a performance test with a throughput of ~25.000 messages per second > that run mulitple hours, some consumers were too slow to consume and messages > piled up on the broker. For this reason, the broker started to page the > messages of growing queues. When we reduced the load from the broker, some > queues stopped consuming due to the following logs: > {code} > AMQ222033: Page file 7.page had incomplete records at position > 39,795,399 at record number 13,952 > target message no.16146 not found from start offset 46032883 and start > message number 16146: java.lang.RuntimeException: target message no.16146 not > found from start offset 46032883 and start message number 16146 > {code} > It wasnt possible to recover from this state but deleting the paging > directory. > h3. Expected Situation > I would expect that the paging mechanism is resilient to any errors. > h3. Scenario Setup > Master configuration: > {code:xml} > > > > true > > > > > > 256Mb > 64Mb > > 10 > PAGE > > {code} > An extract of the monitoring of the Performance Test is attached to this > issue. > h3. Workaround > Right now we disabled paging at all and only use our Heap. However, the heap > is exhausted at 5 million messages which is in our use case better than > loosing any of them. -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (ARTEMIS-2678) Incomplete records for pages under high load
[ https://issues.apache.org/jira/browse/ARTEMIS-2678?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17064946#comment-17064946 ] Francesco Nigro commented on ARTEMIS-2678: -- * Which version of artemis? * I see you're using shared-store, are you sure is the right choice if high throughput are needed? * Which protocol of messages is being used? Core, AMQP, etc etc > Incomplete records for pages under high load > > > Key: ARTEMIS-2678 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2678 > Project: ActiveMQ Artemis > Issue Type: Bug > Components: Broker > Environment: Linux >Reporter: Ansgar J. Sachs >Priority: Major > Attachments: Bildschirmfoto 2020-03-23 um 17.35.41.png > > > {quote}As developer, I expect paging to be resource saving and resilient to > high load{quote} > h3. Current Situation > During a performance test with a throughput of ~25.000 messages per second > that run mulitple hours, some consumers were too slow to consume and messages > piled up on the broker. For this reason, the broker started to page the > messages of growing queues. When we reduced the load from the broker, some > queues stopped consuming due to the following logs: > {code} > AMQ222033: Page file 7.page had incomplete records at position > 39,795,399 at record number 13,952 > target message no.16146 not found from start offset 46032883 and start > message number 16146: java.lang.RuntimeException: target message no.16146 not > found from start offset 46032883 and start message number 16146 > {code} > It wasnt possible to recover from this state but deleting the paging > directory. > h3. Expected Situation > I would expect that the paging mechanism is resilient to any errors. > h3. Scenario Setup > Master configuration: > {code:xml} > > > > true > > > > > > 256Mb > 64Mb > > 10 > PAGE > > {code} > An extract of the monitoring of the Performance Test is attached to this > issue. > h3. Workaround > Right now we disabled paging at all and only use our Heap. However, the heap > is exhausted at 5 million messages which is in our use case better than > loosing any of them. -- This message was sent by Atlassian Jira (v8.3.4#803005)