[jira] [Commented] (CASSANDRA-10524) Add ability to skip TIME_WAIT sockets on port check on Windows startup
[ https://issues.apache.org/jira/browse/CASSANDRA-10524?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14964073#comment-14964073 ] Joshua McKenzie commented on CASSANDRA-10524: - Committed as {{728f94e798662bf8e0e4ad0f4d236394868cb7d6}} > Add ability to skip TIME_WAIT sockets on port check on Windows startup > -- > > Key: CASSANDRA-10524 > URL: https://issues.apache.org/jira/browse/CASSANDRA-10524 > Project: Cassandra > Issue Type: Improvement >Reporter: Joshua McKenzie >Assignee: Joshua McKenzie >Priority: Trivial > Labels: Windows > Fix For: 3.0.0 > > Attachments: 10524_v2.txt, win_aggressive_startup.txt > > > C* sockets are often staying TIME_WAIT for up to 120 seconds (2x max segment > lifetime) for me in my dev environment on Windows. This is rather obnoxious > since it means I can't launch C* for up to 2 minutes after stopping it. > Attaching a patch that adds a simple -a for aggressive startup to the launch > scripts to ignore duplicate port check from netstat if it's TIME_WAIT. Also > snuck in some more liberal interpretation of help strings in the .ps1. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (CASSANDRA-10524) Add ability to skip TIME_WAIT sockets on port check on Windows startup
[ https://issues.apache.org/jira/browse/CASSANDRA-10524?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14961115#comment-14961115 ] Andy Tolbert commented on CASSANDRA-10524: -- looks good to me! > Add ability to skip TIME_WAIT sockets on port check on Windows startup > -- > > Key: CASSANDRA-10524 > URL: https://issues.apache.org/jira/browse/CASSANDRA-10524 > Project: Cassandra > Issue Type: Improvement >Reporter: Joshua McKenzie >Assignee: Joshua McKenzie >Priority: Trivial > Labels: Windows > Fix For: 3.0.0 rc2, 2.2.4 > > Attachments: 10524_v2.txt, win_aggressive_startup.txt > > > C* sockets are often staying TIME_WAIT for up to 120 seconds (2x max segment > lifetime) for me in my dev environment on Windows. This is rather obnoxious > since it means I can't launch C* for up to 2 minutes after stopping it. > Attaching a patch that adds a simple -a for aggressive startup to the launch > scripts to ignore duplicate port check from netstat if it's TIME_WAIT. Also > snuck in some more liberal interpretation of help strings in the .ps1. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (CASSANDRA-10524) Add ability to skip TIME_WAIT sockets on port check on Windows startup
[ https://issues.apache.org/jira/browse/CASSANDRA-10524?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14960986#comment-14960986 ] Philip Thompson commented on CASSANDRA-10524: - I'm +1 on the code. [~andrew.tolbert]? > Add ability to skip TIME_WAIT sockets on port check on Windows startup > -- > > Key: CASSANDRA-10524 > URL: https://issues.apache.org/jira/browse/CASSANDRA-10524 > Project: Cassandra > Issue Type: Improvement >Reporter: Joshua McKenzie >Assignee: Joshua McKenzie >Priority: Trivial > Labels: Windows > Fix For: 3.0.0 rc2, 2.2.4 > > Attachments: 10524_v2.txt, win_aggressive_startup.txt > > > C* sockets are often staying TIME_WAIT for up to 120 seconds (2x max segment > lifetime) for me in my dev environment on Windows. This is rather obnoxious > since it means I can't launch C* for up to 2 minutes after stopping it. > Attaching a patch that adds a simple -a for aggressive startup to the launch > scripts to ignore duplicate port check from netstat if it's TIME_WAIT. Also > snuck in some more liberal interpretation of help strings in the .ps1. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (CASSANDRA-10524) Add ability to skip TIME_WAIT sockets on port check on Windows startup
[ https://issues.apache.org/jira/browse/CASSANDRA-10524?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14960828#comment-14960828 ] Andy Tolbert commented on CASSANDRA-10524: -- I have 12 nodes constantly rolling and things are running pretty reliable except I caught another state {{SYN_SENT}} on 9042 shortly after shutdown, although it could be that I'm just cycling the nodes too quickly as I wouldn't expect to see this state hang around. > Add ability to skip TIME_WAIT sockets on port check on Windows startup > -- > > Key: CASSANDRA-10524 > URL: https://issues.apache.org/jira/browse/CASSANDRA-10524 > Project: Cassandra > Issue Type: Improvement >Reporter: Joshua McKenzie >Assignee: Joshua McKenzie >Priority: Trivial > Labels: Windows > Fix For: 3.0.0 rc2, 2.2.4 > > Attachments: win_aggressive_startup.txt > > > C* sockets are often staying TIME_WAIT for up to 120 seconds (2x max segment > lifetime) for me in my dev environment on Windows. This is rather obnoxious > since it means I can't launch C* for up to 2 minutes after stopping it. > Attaching a patch that adds a simple -a for aggressive startup to the launch > scripts to ignore duplicate port check from netstat if it's TIME_WAIT. Also > snuck in some more liberal interpretation of help strings in the .ps1. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (CASSANDRA-10524) Add ability to skip TIME_WAIT sockets on port check on Windows startup
[ https://issues.apache.org/jira/browse/CASSANDRA-10524?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14959059#comment-14959059 ] Andy Tolbert commented on CASSANDRA-10524: -- Another one I've seen: {{LAST_ACK}} > Add ability to skip TIME_WAIT sockets on port check on Windows startup > -- > > Key: CASSANDRA-10524 > URL: https://issues.apache.org/jira/browse/CASSANDRA-10524 > Project: Cassandra > Issue Type: Improvement >Reporter: Joshua McKenzie >Assignee: Joshua McKenzie >Priority: Trivial > Labels: Windows > Fix For: 3.0.0 rc2, 2.2.4 > > Attachments: win_aggressive_startup.txt > > > C* sockets are often staying TIME_WAIT for up to 120 seconds (2x max segment > lifetime) for me in my dev environment on Windows. This is rather obnoxious > since it means I can't launch C* for up to 2 minutes after stopping it. > Attaching a patch that adds a simple -a for aggressive startup to the launch > scripts to ignore duplicate port check from netstat if it's TIME_WAIT. Also > snuck in some more liberal interpretation of help strings in the .ps1. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (CASSANDRA-10524) Add ability to skip TIME_WAIT sockets on port check on Windows startup
[ https://issues.apache.org/jira/browse/CASSANDRA-10524?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14957715#comment-14957715 ] Andy Tolbert commented on CASSANDRA-10524: -- Just noticed this in my environment too while doing some rolling tests. Beyond {{TIME_WAIT}} i've also seen it bail out on {{FIN_WAIT_2}}, can this be updated to handle more wait states? These are the only two states I've seen thus far. > Add ability to skip TIME_WAIT sockets on port check on Windows startup > -- > > Key: CASSANDRA-10524 > URL: https://issues.apache.org/jira/browse/CASSANDRA-10524 > Project: Cassandra > Issue Type: Improvement >Reporter: Joshua McKenzie >Assignee: Joshua McKenzie >Priority: Trivial > Labels: Windows > Fix For: 3.0.0 rc2, 2.2.4 > > Attachments: win_aggressive_startup.txt > > > C* sockets are often staying TIME_WAIT for up to 120 seconds (2x max segment > lifetime) for me in my dev environment on Windows. This is rather obnoxious > since it means I can't launch C* for up to 2 minutes after stopping it. > Attaching a patch that adds a simple -a for aggressive startup to the launch > scripts to ignore duplicate port check from netstat if it's TIME_WAIT. Also > snuck in some more liberal interpretation of help strings in the .ps1. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (CASSANDRA-10524) Add ability to skip TIME_WAIT sockets on port check on Windows startup
[ https://issues.apache.org/jira/browse/CASSANDRA-10524?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14957466#comment-14957466 ] Philip Thompson commented on CASSANDRA-10524: - +1 > Add ability to skip TIME_WAIT sockets on port check on Windows startup > -- > > Key: CASSANDRA-10524 > URL: https://issues.apache.org/jira/browse/CASSANDRA-10524 > Project: Cassandra > Issue Type: Improvement >Reporter: Joshua McKenzie >Assignee: Joshua McKenzie >Priority: Trivial > Labels: Windows > Fix For: 2.2.x, 3.0.x > > Attachments: win_aggressive_startup.txt > > > C* sockets are often staying TIME_WAIT for up to 120 seconds (2x max segment > lifetime) for me in my dev environment on Windows. This is rather obnoxious > since it means I can't launch C* for up to 2 minutes after stopping it. > Attaching a patch that adds a simple -a for aggressive startup to the launch > scripts to ignore duplicate port check from netstat if it's TIME_WAIT. Also > snuck in some more liberal interpretation of help strings in the .ps1. -- This message was sent by Atlassian JIRA (v6.3.4#6332)