[jira] [Commented] (TS-3922) Add independent WebSocket timeouts
[ https://issues.apache.org/jira/browse/TS-3922?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15267973#comment-15267973 ] Brian Geffon commented on TS-3922: -- [~psudaemon] Yes. Sorry about the mess. > Add independent WebSocket timeouts > -- > > Key: TS-3922 > URL: https://issues.apache.org/jira/browse/TS-3922 > Project: Traffic Server > Issue Type: Improvement > Components: Core >Reporter: Brian Geffon >Assignee: Brian Geffon > Fix For: 7.0.0 > > > Right now WebSockets fall under the active/inactive transaction timeouts. > Obviously the inactivate timeout isn't a problem but the active timeout is. > We need at minimum a new timeout for active WS connections, I'll likely add a > new timeout for inactive state too. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (TS-3922) Add independent WebSocket timeouts
[ https://issues.apache.org/jira/browse/TS-3922?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15267958#comment-15267958 ] Phil Sorber commented on TS-3922: - Is this complete? > Add independent WebSocket timeouts > -- > > Key: TS-3922 > URL: https://issues.apache.org/jira/browse/TS-3922 > Project: Traffic Server > Issue Type: Improvement > Components: Core >Reporter: Brian Geffon >Assignee: Brian Geffon > Fix For: 7.0.0 > > > Right now WebSockets fall under the active/inactive transaction timeouts. > Obviously the inactivate timeout isn't a problem but the active timeout is. > We need at minimum a new timeout for active WS connections, I'll likely add a > new timeout for inactive state too. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (TS-3922) Add independent WebSocket timeouts
[ https://issues.apache.org/jira/browse/TS-3922?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15253228#comment-15253228 ] ASF subversion and git services commented on TS-3922: - Commit 4fdf8ac21e4dcac97297f6afb03c50a5c997f996 in trafficserver's branch refs/heads/master from [~briang] [ https://git-wip-us.apache.org/repos/asf?p=trafficserver.git;h=4fdf8ac ] TS-3922: Websocket Timeouts, transaction overrideable & clang format > Add independent WebSocket timeouts > -- > > Key: TS-3922 > URL: https://issues.apache.org/jira/browse/TS-3922 > Project: Traffic Server > Issue Type: Improvement > Components: Core >Reporter: Brian Geffon >Assignee: Brian Geffon > Fix For: 7.0.0 > > > Right now WebSockets fall under the active/inactive transaction timeouts. > Obviously the inactivate timeout isn't a problem but the active timeout is. > We need at minimum a new timeout for active WS connections, I'll likely add a > new timeout for inactive state too. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (TS-3922) Add independent WebSocket timeouts
[ https://issues.apache.org/jira/browse/TS-3922?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15251371#comment-15251371 ] ASF subversion and git services commented on TS-3922: - Commit 518e829c90dbbd01ac424e8f693955102e9cde93 in trafficserver's branch refs/heads/master from [~briang] [ https://git-wip-us.apache.org/repos/asf?p=trafficserver.git;h=518e829 ] TS-3922: clang format > Add independent WebSocket timeouts > -- > > Key: TS-3922 > URL: https://issues.apache.org/jira/browse/TS-3922 > Project: Traffic Server > Issue Type: Improvement > Components: Core >Reporter: Brian Geffon >Assignee: Brian Geffon > Fix For: 7.0.0 > > > Right now WebSockets fall under the active/inactive transaction timeouts. > Obviously the inactivate timeout isn't a problem but the active timeout is. > We need at minimum a new timeout for active WS connections, I'll likely add a > new timeout for inactive state too. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (TS-3922) Add independent WebSocket timeouts
[ https://issues.apache.org/jira/browse/TS-3922?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15251315#comment-15251315 ] ASF subversion and git services commented on TS-3922: - Commit 25e650b95c0e2644b7fb1a161a1f43e5827d858f in trafficserver's branch refs/heads/master from [~briang] [ https://git-wip-us.apache.org/repos/asf?p=trafficserver.git;h=25e650b ] TS-3922: Fix format warning > Add independent WebSocket timeouts > -- > > Key: TS-3922 > URL: https://issues.apache.org/jira/browse/TS-3922 > Project: Traffic Server > Issue Type: Improvement > Components: Core >Reporter: Brian Geffon >Assignee: Brian Geffon > Fix For: 7.0.0 > > > Right now WebSockets fall under the active/inactive transaction timeouts. > Obviously the inactivate timeout isn't a problem but the active timeout is. > We need at minimum a new timeout for active WS connections, I'll likely add a > new timeout for inactive state too. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (TS-3922) Add independent WebSocket timeouts
[ https://issues.apache.org/jira/browse/TS-3922?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15251244#comment-15251244 ] Brian Geffon commented on TS-3922: -- [~psudaemon] please consider for backport to 6.2.x > Add independent WebSocket timeouts > -- > > Key: TS-3922 > URL: https://issues.apache.org/jira/browse/TS-3922 > Project: Traffic Server > Issue Type: Improvement > Components: Core >Reporter: Brian Geffon >Assignee: Brian Geffon > Fix For: 7.0.0 > > > Right now WebSockets fall under the active/inactive transaction timeouts. > Obviously the inactivate timeout isn't a problem but the active timeout is. > We need at minimum a new timeout for active WS connections, I'll likely add a > new timeout for inactive state too. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (TS-3922) Add independent WebSocket timeouts
[ https://issues.apache.org/jira/browse/TS-3922?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15251242#comment-15251242 ] ASF GitHub Bot commented on TS-3922: Github user asfgit closed the pull request at: https://github.com/apache/trafficserver/pull/590 > Add independent WebSocket timeouts > -- > > Key: TS-3922 > URL: https://issues.apache.org/jira/browse/TS-3922 > Project: Traffic Server > Issue Type: Improvement > Components: Core >Reporter: Brian Geffon >Assignee: Brian Geffon > Fix For: 7.0.0 > > > Right now WebSockets fall under the active/inactive transaction timeouts. > Obviously the inactivate timeout isn't a problem but the active timeout is. > We need at minimum a new timeout for active WS connections, I'll likely add a > new timeout for inactive state too. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (TS-3922) Add independent WebSocket timeouts
[ https://issues.apache.org/jira/browse/TS-3922?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15251243#comment-15251243 ] ASF GitHub Bot commented on TS-3922: Github user bgaff commented on the pull request: https://github.com/apache/trafficserver/pull/590#issuecomment-212738312 Merged. > Add independent WebSocket timeouts > -- > > Key: TS-3922 > URL: https://issues.apache.org/jira/browse/TS-3922 > Project: Traffic Server > Issue Type: Improvement > Components: Core >Reporter: Brian Geffon >Assignee: Brian Geffon > Fix For: 7.0.0 > > > Right now WebSockets fall under the active/inactive transaction timeouts. > Obviously the inactivate timeout isn't a problem but the active timeout is. > We need at minimum a new timeout for active WS connections, I'll likely add a > new timeout for inactive state too. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (TS-3922) Add independent WebSocket timeouts
[ https://issues.apache.org/jira/browse/TS-3922?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15251241#comment-15251241 ] ASF subversion and git services commented on TS-3922: - Commit 1466ccadd2c3cee31de9835249bf24f29a1d9ea1 in trafficserver's branch refs/heads/master from [~briang] [ https://git-wip-us.apache.org/repos/asf?p=trafficserver.git;h=1466cca ] TS-3922: Add independent websocket timeouts. This closes #590 > Add independent WebSocket timeouts > -- > > Key: TS-3922 > URL: https://issues.apache.org/jira/browse/TS-3922 > Project: Traffic Server > Issue Type: Improvement > Components: Core >Reporter: Brian Geffon >Assignee: Brian Geffon > Fix For: 7.0.0 > > > Right now WebSockets fall under the active/inactive transaction timeouts. > Obviously the inactivate timeout isn't a problem but the active timeout is. > We need at minimum a new timeout for active WS connections, I'll likely add a > new timeout for inactive state too. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (TS-3922) Add independent WebSocket timeouts
[ https://issues.apache.org/jira/browse/TS-3922?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15251189#comment-15251189 ] ASF GitHub Bot commented on TS-3922: Github user zizhong commented on a diff in the pull request: https://github.com/apache/trafficserver/pull/590#discussion_r6052 --- Diff: iocore/net/UnixNet.cc --- @@ -96,8 +97,8 @@ class InactivityCop : public Continuation NET_SUM_DYN_STAT(keep_alive_queue_timeout_total_stat, diff); NET_INCREMENT_DYN_STAT(keep_alive_queue_timeout_count_stat); } -Debug("inactivity_cop_verbose", "vc: %p now: %" PRId64 " timeout at: %" PRId64 " timeout in: %" PRId64, vc, now, - vc->next_inactivity_timeout_at, vc->inactivity_timeout_in); +Debug("inactivity_cop_verbose_2", "vc: %p now: %" PRId64 " timeout at: %" PRId64 " timeout in: %" PRId64, vc, --- End diff -- Docs were updated. Tag name was fixed. > Add independent WebSocket timeouts > -- > > Key: TS-3922 > URL: https://issues.apache.org/jira/browse/TS-3922 > Project: Traffic Server > Issue Type: Improvement > Components: Core >Reporter: Brian Geffon >Assignee: Brian Geffon > Fix For: 7.0.0 > > > Right now WebSockets fall under the active/inactive transaction timeouts. > Obviously the inactivate timeout isn't a problem but the active timeout is. > We need at minimum a new timeout for active WS connections, I'll likely add a > new timeout for inactive state too. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (TS-3922) Add independent WebSocket timeouts
[ https://issues.apache.org/jira/browse/TS-3922?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15251172#comment-15251172 ] James Peach commented on TS-3922: - Yup, just asking if you can take a look and maybe write up Jiras :) > Add independent WebSocket timeouts > -- > > Key: TS-3922 > URL: https://issues.apache.org/jira/browse/TS-3922 > Project: Traffic Server > Issue Type: Improvement > Components: Core >Reporter: Brian Geffon >Assignee: Brian Geffon > Fix For: 7.0.0 > > > Right now WebSockets fall under the active/inactive transaction timeouts. > Obviously the inactivate timeout isn't a problem but the active timeout is. > We need at minimum a new timeout for active WS connections, I'll likely add a > new timeout for inactive state too. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (TS-3922) Add independent WebSocket timeouts
[ https://issues.apache.org/jira/browse/TS-3922?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15251157#comment-15251157 ] Brian Geffon commented on TS-3922: -- [~jpe...@apache.org] I think that's a bigger discussion we should set aside some time to discuss at the next summit. Personally I think if we're going to pull the trigger on config naming consistency it needs to happen w/ 7.0.0. [~zwoop], thoughts? Brian > Add independent WebSocket timeouts > -- > > Key: TS-3922 > URL: https://issues.apache.org/jira/browse/TS-3922 > Project: Traffic Server > Issue Type: Improvement > Components: Core >Reporter: Brian Geffon >Assignee: Brian Geffon > Fix For: 7.0.0 > > > Right now WebSockets fall under the active/inactive transaction timeouts. > Obviously the inactivate timeout isn't a problem but the active timeout is. > We need at minimum a new timeout for active WS connections, I'll likely add a > new timeout for inactive state too. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (TS-3922) Add independent WebSocket timeouts
[ https://issues.apache.org/jira/browse/TS-3922?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15251154#comment-15251154 ] ASF GitHub Bot commented on TS-3922: Github user bgaff commented on a diff in the pull request: https://github.com/apache/trafficserver/pull/590#discussion_r60519710 --- Diff: iocore/net/UnixNet.cc --- @@ -96,8 +97,8 @@ class InactivityCop : public Continuation NET_SUM_DYN_STAT(keep_alive_queue_timeout_total_stat, diff); NET_INCREMENT_DYN_STAT(keep_alive_queue_timeout_count_stat); } -Debug("inactivity_cop_verbose", "vc: %p now: %" PRId64 " timeout at: %" PRId64 " timeout in: %" PRId64, vc, now, - vc->next_inactivity_timeout_at, vc->inactivity_timeout_in); +Debug("inactivity_cop_verbose_2", "vc: %p now: %" PRId64 " timeout at: %" PRId64 " timeout in: %" PRId64, vc, --- End diff -- I just realized that in the patch I pasted in TS-3922 I left this debug tag name changed as inactivity_cop_verbose_2, @zizhong can you please correct that for me. > Add independent WebSocket timeouts > -- > > Key: TS-3922 > URL: https://issues.apache.org/jira/browse/TS-3922 > Project: Traffic Server > Issue Type: Improvement > Components: Core >Reporter: Brian Geffon >Assignee: Brian Geffon > Fix For: 7.0.0 > > > Right now WebSockets fall under the active/inactive transaction timeouts. > Obviously the inactivate timeout isn't a problem but the active timeout is. > We need at minimum a new timeout for active WS connections, I'll likely add a > new timeout for inactive state too. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (TS-3922) Add independent WebSocket timeouts
[ https://issues.apache.org/jira/browse/TS-3922?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15251153#comment-15251153 ] ASF GitHub Bot commented on TS-3922: Github user bgaff commented on the pull request: https://github.com/apache/trafficserver/pull/590#issuecomment-212705977 @zizhong we need to add docs too, can you please update the records.config documentation. Thanks. > Add independent WebSocket timeouts > -- > > Key: TS-3922 > URL: https://issues.apache.org/jira/browse/TS-3922 > Project: Traffic Server > Issue Type: Improvement > Components: Core >Reporter: Brian Geffon >Assignee: Brian Geffon > Fix For: 7.0.0 > > > Right now WebSockets fall under the active/inactive transaction timeouts. > Obviously the inactivate timeout isn't a problem but the active timeout is. > We need at minimum a new timeout for active WS connections, I'll likely add a > new timeout for inactive state too. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (TS-3922) Add independent WebSocket timeouts
[ https://issues.apache.org/jira/browse/TS-3922?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15249287#comment-15249287 ] James Peach commented on TS-3922: - [~briang] Can you take a look at the timeout settings and see what can be done to make them consistent? {{keep_alive}}, {{transaction}}, {{accept}} differ from the pattern you propose. (I think your proposal is correct, BTW). {code} proxy.config.lm.pserver_timeout proxy.config.lm.pserver_timeout proxy.config.cluster.peer_timeout proxy.config.cluster.mc_poll_timeout proxy.config.cluster.startup_timeout proxy.config.process_manager.timeout proxy.config.vmap.down_up_timeout proxy.config.ping.timeout proxy.config.http.congestion_control.default.live_os_conn_timeout proxy.config.http.congestion_control.default.dead_os_conn_timeout proxy.config.http.parent_proxy.connect_attempts_timeout proxy.config.http.keep_alive_no_activity_timeout proxy.config.http.keep_alive_no_activity_timeout proxy.config.http.transaction_no_activity_timeout proxy.config.http.transaction_no_activity_timeout proxy.config.http.transaction_active_timeout proxy.config.http.transaction_active_timeout proxy.config.http.accept_no_activity_timeout proxy.config.http.background_fill_active_timeout proxy.config.http.connect_attempts_timeout proxy.config.http.post_connect_attempts_timeout proxy.config.socks.socks_timeout proxy.config.socks.server_connect_timeout proxy.config.socks.server_retry_timeout proxy.config.net.poll_timeout proxy.config.net.default_inactivity_timeout proxy.config.dns.lookup_timeout proxy.config.hostdb.lookup_timeout proxy.config.hostdb.timeout proxy.config.hostdb.fail.timeout proxy.config.ssl.session_cache.timeout proxy.config.ssl.handshake_timeout proxy.config.ssl.ocsp.cache_timeout proxy.config.ssl.ocsp.request_timeout proxy.config.icp.query_timeout proxy.config.spdy.no_activity_timeout proxy.config.spdy.accept_no_activity_timeout proxy.config.http2.accept_no_activity_timeout proxy.config.http2.no_activity_timeout proxy.config.http2.active_timeout {code} > Add independent WebSocket timeouts > -- > > Key: TS-3922 > URL: https://issues.apache.org/jira/browse/TS-3922 > Project: Traffic Server > Issue Type: Improvement > Components: Core >Reporter: Brian Geffon >Assignee: Brian Geffon > Fix For: 7.0.0 > > > Right now WebSockets fall under the active/inactive transaction timeouts. > Obviously the inactivate timeout isn't a problem but the active timeout is. > We need at minimum a new timeout for active WS connections, I'll likely add a > new timeout for inactive state too. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (TS-3922) Add independent WebSocket timeouts
[ https://issues.apache.org/jira/browse/TS-3922?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15249214#comment-15249214 ] Brian Geffon commented on TS-3922: -- The following patch was based off an old version of master: {code} diff --git a/iocore/net/UnixNet.cc b/iocore/net/UnixNet.cc index 131a12f..64c3dcb 100644 --- a/iocore/net/UnixNet.cc +++ b/iocore/net/UnixNet.cc @@ -86,7 +86,7 @@ public: vc->set_inactivity_timeout(HRTIME_SECONDS(default_inactivity_timeout)); NET_INCREMENT_DYN_STAT(default_inactivity_timeout_stat); } else { -Debug("inactivity_cop_verbose", "vc: %p now: %" PRId64 " timeout at: %" PRId64 " timeout in: %" PRId64, vc, now, +Debug("inactivity_cop_verbose", "vc: %p now: %" PRId64 " timeout at: %" PRId64 " timeout in: %" PRId64, vc, ink_hrtime_to_sec(now), ink_hrtime_to_sec(vc->next_inactivity_timeout_at), ink_hrtime_to_sec(vc->inactivity_timeout_in)); } @@ -97,7 +97,7 @@ public: NET_SUM_DYN_STAT(keep_alive_queue_timeout_total_stat, diff); NET_INCREMENT_DYN_STAT(keep_alive_queue_timeout_count_stat); } -Debug("inactivity_cop_verbose", "vc: %p now: %" PRId64 " timeout at: %" PRId64 " timeout in: %" PRId64, vc, now, +Debug("inactivity_cop_verbose_2", "vc: %p now: %" PRId64 " timeout at: %" PRId64 " timeout in: %" PRId64, vc, ink_hrtime_to_sec(now), vc->next_inactivity_timeout_at, vc->inactivity_timeout_in); vc->handleEvent(EVENT_IMMEDIATE, e); } diff --git a/mgmt/RecordsConfig.cc b/mgmt/RecordsConfig.cc index 66c5ee6..b46ee7a 100644 --- a/mgmt/RecordsConfig.cc +++ b/mgmt/RecordsConfig.cc @@ -525,6 +525,10 @@ static const RecordElement RecordsConfig[] = , {RECT_CONFIG, "proxy.config.http.transaction_no_activity_timeout_out", RECD_INT, "30", RECU_DYNAMIC, RR_NULL, RECC_STR, "^[0-9]+$", RECA_NULL} , + {RECT_CONFIG, "proxy.config.websocket.no_activity_timeout", RECD_INT, "600", RECU_DYNAMIC, RR_NULL, RECC_STR, "^[0-9]+$", RECA_NULL} + , + {RECT_CONFIG, "proxy.config.websocket.active_timeout", RECD_INT, "3600", RECU_DYNAMIC, RR_NULL, RECC_STR, "^[0-9]+$", RECA_NULL} + , {RECT_CONFIG, "proxy.config.http.transaction_active_timeout_in", RECD_INT, "900", RECU_DYNAMIC, RR_NULL, RECC_STR, "^[0-9]+$", RECA_NULL} , {RECT_CONFIG, "proxy.config.http.transaction_active_timeout_out", RECD_INT, "0", RECU_DYNAMIC, RR_NULL, RECC_STR, "^[0-9]+$", RECA_NULL} diff --git a/proxy/http/HttpConfig.cc b/proxy/http/HttpConfig.cc index 96fe448..0100cdb 100644 --- a/proxy/http/HttpConfig.cc +++ b/proxy/http/HttpConfig.cc @@ -947,6 +947,11 @@ HttpConfig::startup() "proxy.config.http.transaction_no_activity_timeout_in"); HttpEstablishStaticConfigLongLong(c.oride.transaction_no_activity_timeout_out, "proxy.config.http.transaction_no_activity_timeout_out"); + HttpEstablishStaticConfigLongLong(c.oride.websocket_active_timeout, + "proxy.config.websocket.active_timeout"); + HttpEstablishStaticConfigLongLong(c.oride.websocket_inactive_timeout, + "proxy.config.websocket.no_activity_timeout"); + HttpEstablishStaticConfigLongLong(c.transaction_active_timeout_in, "proxy.config.http.transaction_active_timeout_in"); HttpEstablishStaticConfigLongLong(c.oride.transaction_active_timeout_out, "proxy.config.http.transaction_active_timeout_out"); HttpEstablishStaticConfigLongLong(c.accept_no_activity_timeout, "proxy.config.http.accept_no_activity_timeout"); @@ -1215,6 +1220,8 @@ HttpConfig::reconfigure() params->oride.keep_alive_no_activity_timeout_out = m_master.oride.keep_alive_no_activity_timeout_out; params->oride.transaction_no_activity_timeout_in = m_master.oride.transaction_no_activity_timeout_in; params->oride.transaction_no_activity_timeout_out = m_master.oride.transaction_no_activity_timeout_out; + params->oride.websocket_active_timeout = m_master.oride.websocket_active_timeout; + params->oride.websocket_inactive_timeout = m_master.oride.websocket_inactive_timeout; params->transaction_active_timeout_in = m_master.transaction_active_timeout_in; params->oride.transaction_active_timeout_out = m_master.oride.transaction_active_timeout_out; params->accept_no_activity_timeout = m_master.accept_no_activity_timeout; diff --git a/proxy/http/HttpConfig.h b/proxy/http/HttpConfig.h index 62cc4a4..0ebc808 100644 --- a/proxy/http/HttpConfig.h +++ b/proxy/http/HttpConfig.h @@ -378,6 +378,7 @@ struct OverridableHttpConfigParams { cache_heuristic_min_lifetime(3600), cache_heuristic_max_lifetime(86400), cache_guaranteed_min_lifetime(0), cache_guaranteed_max_lifetime(31536000), cache_max_stale_age(604800), keep_alive_no_activity_timeout_in(115), keep_alive_no_activity_timeout_out(120),
[jira] [Commented] (TS-3922) Add independent WebSocket timeouts
[ https://issues.apache.org/jira/browse/TS-3922?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15249145#comment-15249145 ] Brian Geffon commented on TS-3922: -- [~zhangzizhong] will open a pull request for this soon. > Add independent WebSocket timeouts > -- > > Key: TS-3922 > URL: https://issues.apache.org/jira/browse/TS-3922 > Project: Traffic Server > Issue Type: Improvement > Components: Core >Reporter: Brian Geffon >Assignee: Brian Geffon > Fix For: 7.0.0 > > > Right now WebSockets fall under the active/inactive transaction timeouts. > Obviously the inactivate timeout isn't a problem but the active timeout is. > We need at minimum a new timeout for active WS connections, I'll likely add a > new timeout for inactive state too. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (TS-3922) Add independent WebSocket timeouts
[ https://issues.apache.org/jira/browse/TS-3922?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15095223#comment-15095223 ] Brian Geffon commented on TS-3922: -- Well I did this but never committed it upstream, landing today. > Add independent WebSocket timeouts > -- > > Key: TS-3922 > URL: https://issues.apache.org/jira/browse/TS-3922 > Project: Traffic Server > Issue Type: Improvement > Components: Core >Reporter: Brian Geffon >Assignee: Brian Geffon > Fix For: 6.2.0 > > > Right now WebSockets fall under the active/inactive transaction timeouts. > Obviously the inactivate timeout isn't a problem but the active timeout is. > We need at minimum a new timeout for active WS connections, I'll likely add a > new timeout for inactive state too. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (TS-3922) Add independent WebSocket timeouts
[ https://issues.apache.org/jira/browse/TS-3922?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15061199#comment-15061199 ] Brian Geffon commented on TS-3922: -- Actually I'm pretty sure this was done and committed. I need to figure out why this ticket i still open. > Add independent WebSocket timeouts > -- > > Key: TS-3922 > URL: https://issues.apache.org/jira/browse/TS-3922 > Project: Traffic Server > Issue Type: Improvement > Components: Core >Reporter: Brian Geffon >Assignee: Brian Geffon > Fix For: sometime > > > Right now WebSockets fall under the active/inactive transaction timeouts. > Obviously the inactivate timeout isn't a problem but the active timeout is. > We need at minimum a new timeout for active WS connections, I'll likely add a > new timeout for inactive state too. -- This message was sent by Atlassian JIRA (v6.3.4#6332)