[jira] [Commented] (PROTON-401) Ordering issue prevents credit drain from working properly
[ https://issues.apache.org/jira/browse/PROTON-401?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13814034#comment-13814034 ] Ted Ross commented on PROTON-401: - If there is a way for the Engine user to determine the drain mode of a sender, I believe I can solve my problem. Ordering issue prevents credit drain from working properly -- Key: PROTON-401 URL: https://issues.apache.org/jira/browse/PROTON-401 Project: Qpid Proton Issue Type: Bug Components: proton-c Affects Versions: 0.5 Reporter: Ken Giusti Assignee: Rafael H. Schloming Fix For: 0.6 Attachments: drain-error.patch, drain-hack.patch If the sending link calls pn_link_drained() to indicate that it has send all pending data, and afterwards it receives a Flow frame with drain=true from the peer, then the drain never completes. The ordering is the problem: if the flow frame w/drain=true is received _BEFORE_ the sender calls pn_link_drained(), then it works. -- This message was sent by Atlassian JIRA (v6.1#6144)
Re: [jira] [Commented] (PROTON-401) Ordering issue prevents credit drain from working properly
You should be able to use pn_link_get_drain(sender) to determine the drain mode of a sender. On Tue, Nov 5, 2013 at 11:50 AM, Ted Ross (JIRA) j...@apache.org wrote: [ https://issues.apache.org/jira/browse/PROTON-401?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13814034#comment-13814034] Ted Ross commented on PROTON-401: - If there is a way for the Engine user to determine the drain mode of a sender, I believe I can solve my problem. Ordering issue prevents credit drain from working properly -- Key: PROTON-401 URL: https://issues.apache.org/jira/browse/PROTON-401 Project: Qpid Proton Issue Type: Bug Components: proton-c Affects Versions: 0.5 Reporter: Ken Giusti Assignee: Rafael H. Schloming Fix For: 0.6 Attachments: drain-error.patch, drain-hack.patch If the sending link calls pn_link_drained() to indicate that it has send all pending data, and afterwards it receives a Flow frame with drain=true from the peer, then the drain never completes. The ordering is the problem: if the flow frame w/drain=true is received _BEFORE_ the sender calls pn_link_drained(), then it works. -- This message was sent by Atlassian JIRA (v6.1#6144)
[jira] [Commented] (PROTON-401) Ordering issue prevents credit drain from working properly
[ https://issues.apache.org/jira/browse/PROTON-401?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13812823#comment-13812823 ] Ted Ross commented on PROTON-401: - I propose that proton consider the 'drained' state of the application to be level-sensitive and indicated by calls to 'drained' and 'offered'. Whenever the application has at least one message to send, it will call pn_link_offered to indicate how many messages are ready to go. Once the last message has been sent, the application calls 'drained' to indicate that there are no more messages for now (and until a subsequent call to offered). Note that we could remove the pn_link_drained call altogether and replace it with an offered value of zero. Ordering issue prevents credit drain from working properly -- Key: PROTON-401 URL: https://issues.apache.org/jira/browse/PROTON-401 Project: Qpid Proton Issue Type: Bug Components: proton-c Affects Versions: 0.4 Reporter: Ken Giusti Assignee: Ken Giusti Attachments: drain-error.patch, drain-hack.patch If the sending link calls pn_link_drained() to indicate that it has send all pending data, and afterwards it receives a Flow frame with drain=true from the peer, then the drain never completes. The ordering is the problem: if the flow frame w/drain=true is received _BEFORE_ the sender calls pn_link_drained(), then it works. -- This message was sent by Atlassian JIRA (v6.1#6144)
[jira] [Commented] (PROTON-401) Ordering issue prevents credit drain from working properly
[ https://issues.apache.org/jira/browse/PROTON-401?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13812836#comment-13812836 ] Ted Ross commented on PROTON-401: - To put the proposal in the context of Rafi's scenario: If the queue implementation invokes pn_link_offered on the outbound link prior to accepting the message on the inbound link, the strict semantics will be met. Ordering issue prevents credit drain from working properly -- Key: PROTON-401 URL: https://issues.apache.org/jira/browse/PROTON-401 Project: Qpid Proton Issue Type: Bug Components: proton-c Affects Versions: 0.5 Reporter: Ken Giusti Assignee: Ken Giusti Fix For: 0.6 Attachments: drain-error.patch, drain-hack.patch If the sending link calls pn_link_drained() to indicate that it has send all pending data, and afterwards it receives a Flow frame with drain=true from the peer, then the drain never completes. The ordering is the problem: if the flow frame w/drain=true is received _BEFORE_ the sender calls pn_link_drained(), then it works. -- This message was sent by Atlassian JIRA (v6.1#6144)
[jira] [Commented] (PROTON-401) Ordering issue prevents credit drain from working properly
[ https://issues.apache.org/jira/browse/PROTON-401?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13812951#comment-13812951 ] Ted Ross commented on PROTON-401: - That works for me. I just don't want to poll the drain in case someone might care. The way it is now, a broker/intermediary that is idle with no enqueued messages needs to hammer away at pn_link_drained for each outgoing link. qpid::messaging::fetch now hangs when receiving from Dispatch because of this. When can we get a fix? Ordering issue prevents credit drain from working properly -- Key: PROTON-401 URL: https://issues.apache.org/jira/browse/PROTON-401 Project: Qpid Proton Issue Type: Bug Components: proton-c Affects Versions: 0.5 Reporter: Ken Giusti Assignee: Rafael H. Schloming Fix For: 0.6 Attachments: drain-error.patch, drain-hack.patch If the sending link calls pn_link_drained() to indicate that it has send all pending data, and afterwards it receives a Flow frame with drain=true from the peer, then the drain never completes. The ordering is the problem: if the flow frame w/drain=true is received _BEFORE_ the sender calls pn_link_drained(), then it works. -- This message was sent by Atlassian JIRA (v6.1#6144)
[jira] [Commented] (PROTON-401) Ordering issue prevents credit drain from working properly
[ https://issues.apache.org/jira/browse/PROTON-401?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13812959#comment-13812959 ] Rafael H. Schloming commented on PROTON-401: I'll take a look as soon as I wrap up the JIRA I'm currently working on. BTW, what do you mean by hammer? Is your code actually busy looping, or is it simply having to do a linear search whenever the connection has I/O? Ordering issue prevents credit drain from working properly -- Key: PROTON-401 URL: https://issues.apache.org/jira/browse/PROTON-401 Project: Qpid Proton Issue Type: Bug Components: proton-c Affects Versions: 0.5 Reporter: Ken Giusti Assignee: Rafael H. Schloming Fix For: 0.6 Attachments: drain-error.patch, drain-hack.patch If the sending link calls pn_link_drained() to indicate that it has send all pending data, and afterwards it receives a Flow frame with drain=true from the peer, then the drain never completes. The ordering is the problem: if the flow frame w/drain=true is received _BEFORE_ the sender calls pn_link_drained(), then it works. -- This message was sent by Atlassian JIRA (v6.1#6144)
[jira] [Commented] (PROTON-401) Ordering issue prevents credit drain from working properly
[ https://issues.apache.org/jira/browse/PROTON-401?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13812978#comment-13812978 ] Ted Ross commented on PROTON-401: - My code only calls pn_link_drained once when the output buffer empties. Because of this issue, subsequent drain requests (like from qpid::messaging::Receiver::fetch) are not responded to and hang (even if there is a timeout). Since I'm doing the full output processing for every IO event, it's possible that I may have a relatively easy workaround for this. Currently, I don't always re-process the connector in proton after calling pn_link_drained. Ordering issue prevents credit drain from working properly -- Key: PROTON-401 URL: https://issues.apache.org/jira/browse/PROTON-401 Project: Qpid Proton Issue Type: Bug Components: proton-c Affects Versions: 0.5 Reporter: Ken Giusti Assignee: Rafael H. Schloming Fix For: 0.6 Attachments: drain-error.patch, drain-hack.patch If the sending link calls pn_link_drained() to indicate that it has send all pending data, and afterwards it receives a Flow frame with drain=true from the peer, then the drain never completes. The ordering is the problem: if the flow frame w/drain=true is received _BEFORE_ the sender calls pn_link_drained(), then it works. -- This message was sent by Atlassian JIRA (v6.1#6144)
[jira] [Commented] (PROTON-401) Ordering issue prevents credit drain from working properly
[ https://issues.apache.org/jira/browse/PROTON-401?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13781811#comment-13781811 ] Ken Giusti commented on PROTON-401: --- I see your point - I'll update the PROTON-200 patch to call drained as you explain in your comment. Ordering issue prevents credit drain from working properly -- Key: PROTON-401 URL: https://issues.apache.org/jira/browse/PROTON-401 Project: Qpid Proton Issue Type: Bug Components: proton-c Affects Versions: 0.4 Reporter: Ken Giusti Assignee: Ken Giusti Attachments: drain-error.patch, drain-hack.patch If the sending link calls pn_link_drained() to indicate that it has send all pending data, and afterwards it receives a Flow frame with drain=true from the peer, then the drain never completes. The ordering is the problem: if the flow frame w/drain=true is received _BEFORE_ the sender calls pn_link_drained(), then it works. -- This message was sent by Atlassian JIRA (v6.1#6144)
[jira] [Commented] (PROTON-401) Ordering issue prevents credit drain from working properly
[ https://issues.apache.org/jira/browse/PROTON-401?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13746509#comment-13746509 ] Ken Giusti commented on PROTON-401: --- Reviewboard link: https://reviews.apache.org/r/13652/ Ordering issue prevents credit drain from working properly -- Key: PROTON-401 URL: https://issues.apache.org/jira/browse/PROTON-401 Project: Qpid Proton Issue Type: Bug Components: proton-c Affects Versions: 0.4 Reporter: Ken Giusti Assignee: Ken Giusti Fix For: 0.5 Attachments: drain-error.patch, drain-hack.patch If the sending link calls pn_link_drained() to indicate that it has send all pending data, and afterwards it receives a Flow frame with drain=true from the peer, then the drain never completes. The ordering is the problem: if the flow frame w/drain=true is received _BEFORE_ the sender calls pn_link_drained(), then it works. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
