[jira] [Commented] (WAGON-567) Provide request retry strategy on transient client and server side errors
[ https://issues.apache.org/jira/browse/WAGON-567?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16969960#comment-16969960 ] Michael Osipov commented on WAGON-567: -- They aren't the best, but the retry handler uses similar names. We wanted to keep it consistent for the users. > Provide request retry strategy on transient client and server side errors > - > > Key: WAGON-567 > URL: https://issues.apache.org/jira/browse/WAGON-567 > Project: Maven Wagon > Issue Type: New Feature > Components: wagon-http >Reporter: xueqian zhang >Assignee: Michael Osipov >Priority: Minor > Fix For: 3.3.4 > > Time Spent: 20m > Remaining Estimate: 0h > > Currently wagon has a retry mechanism when there are connections issues with > repository manager like maven central, artifactory or nexus, see [wagon > source|https://github.com/apache/maven-wagon/blob/dfd22586b6b934aa5a652ca32d57ed26067432fd/wagon-providers/wagon-http-shared/src/main/java/org/apache/maven/wagon/shared/http/AbstractHttpClientWagon.java#L393] > and [httpclient > source|https://github.com/apache/httpcomponents-client/blob/4.5.x/httpclient/src/main/java/org/apache/http/impl/execchain/RetryExec.java#L90], > [httpclient default retry > handler|https://github.com/apache/httpcomponents-client/blob/4.5.x/httpclient/src/main/java/org/apache/http/impl/client/DefaultHttpRequestRetryHandler.java#L154]. > However, it only captures low level connection issues. When the repo > returns errors like 500 or 503, it does not retry. The 5xx response from the > repo might be a blip and maven should retry before giving up. -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (WAGON-567) Provide request retry strategy on transient client and server side errors
[ https://issues.apache.org/jira/browse/WAGON-567?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16969958#comment-16969958 ] Herve Boutemy commented on WAGON-567: - thanks for the explanation: now I better understand how these ids came in :) > Provide request retry strategy on transient client and server side errors > - > > Key: WAGON-567 > URL: https://issues.apache.org/jira/browse/WAGON-567 > Project: Maven Wagon > Issue Type: New Feature > Components: wagon-http >Reporter: xueqian zhang >Assignee: Michael Osipov >Priority: Minor > Fix For: 3.3.4 > > Time Spent: 20m > Remaining Estimate: 0h > > Currently wagon has a retry mechanism when there are connections issues with > repository manager like maven central, artifactory or nexus, see [wagon > source|https://github.com/apache/maven-wagon/blob/dfd22586b6b934aa5a652ca32d57ed26067432fd/wagon-providers/wagon-http-shared/src/main/java/org/apache/maven/wagon/shared/http/AbstractHttpClientWagon.java#L393] > and [httpclient > source|https://github.com/apache/httpcomponents-client/blob/4.5.x/httpclient/src/main/java/org/apache/http/impl/execchain/RetryExec.java#L90], > [httpclient default retry > handler|https://github.com/apache/httpcomponents-client/blob/4.5.x/httpclient/src/main/java/org/apache/http/impl/client/DefaultHttpRequestRetryHandler.java#L154]. > However, it only captures low level connection issues. When the repo > returns errors like 500 or 503, it does not retry. The 5xx response from the > repo might be a blip and maven should retry before giving up. -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (WAGON-567) Provide request retry strategy on transient client and server side errors
[
https://issues.apache.org/jira/browse/WAGON-567?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16969527#comment-16969527
]
Michael Osipov commented on WAGON-567:
--
[~hboutemy], looking into your question: no default behavior change has been
applied. If you don't specifiy
{{maven.wagon.http.serviceUnavailableRetryStrategy.class != none}}, nothing
will change.
I agree that the names are a bit irritating, but the refer to
{{DefaultServiceUnavailableRetryStrategy}} and
{{StandardServiceUnavailableRetryStrategy}} provided by the {{HttpClient}}, no
custom implementations. This will be merged with the retry handler as soon as
HTTPCLIENT-2019 is solved in 5.0 and we switch to 5.0.
> Provide request retry strategy on transient client and server side errors
> -
>
> Key: WAGON-567
> URL: https://issues.apache.org/jira/browse/WAGON-567
> Project: Maven Wagon
> Issue Type: New Feature
> Components: wagon-http
>Reporter: xueqian zhang
>Assignee: Michael Osipov
>Priority: Minor
> Fix For: 3.3.4
>
> Time Spent: 20m
> Remaining Estimate: 0h
>
> Currently wagon has a retry mechanism when there are connections issues with
> repository manager like maven central, artifactory or nexus, see [wagon
> source|https://github.com/apache/maven-wagon/blob/dfd22586b6b934aa5a652ca32d57ed26067432fd/wagon-providers/wagon-http-shared/src/main/java/org/apache/maven/wagon/shared/http/AbstractHttpClientWagon.java#L393]
> and [httpclient
> source|https://github.com/apache/httpcomponents-client/blob/4.5.x/httpclient/src/main/java/org/apache/http/impl/execchain/RetryExec.java#L90],
> [httpclient default retry
> handler|https://github.com/apache/httpcomponents-client/blob/4.5.x/httpclient/src/main/java/org/apache/http/impl/client/DefaultHttpRequestRetryHandler.java#L154].
> However, it only captures low level connection issues. When the repo
> returns errors like 500 or 503, it does not retry. The 5xx response from the
> repo might be a blip and maven should retry before giving up.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
[jira] [Commented] (WAGON-567) Provide request retry strategy on transient client and server side errors
[ https://issues.apache.org/jira/browse/WAGON-567?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16969350#comment-16969350 ] Herve Boutemy commented on WAGON-567: - ok, thanks > Provide request retry strategy on transient client and server side errors > - > > Key: WAGON-567 > URL: https://issues.apache.org/jira/browse/WAGON-567 > Project: Maven Wagon > Issue Type: New Feature > Components: wagon-http >Reporter: xueqian zhang >Assignee: Michael Osipov >Priority: Minor > Fix For: 3.3.4 > > Time Spent: 20m > Remaining Estimate: 0h > > Currently wagon has a retry mechanism when there are connections issues with > repository manager like maven central, artifactory or nexus, see [wagon > source|https://github.com/apache/maven-wagon/blob/dfd22586b6b934aa5a652ca32d57ed26067432fd/wagon-providers/wagon-http-shared/src/main/java/org/apache/maven/wagon/shared/http/AbstractHttpClientWagon.java#L393] > and [httpclient > source|https://github.com/apache/httpcomponents-client/blob/4.5.x/httpclient/src/main/java/org/apache/http/impl/execchain/RetryExec.java#L90], > [httpclient default retry > handler|https://github.com/apache/httpcomponents-client/blob/4.5.x/httpclient/src/main/java/org/apache/http/impl/client/DefaultHttpRequestRetryHandler.java#L154]. > However, it only captures low level connection issues. When the repo > returns errors like 500 or 503, it does not retry. The 5xx response from the > repo might be a blip and maven should retry before giving up. -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (WAGON-567) Provide request retry strategy on transient client and server side errors
[
https://issues.apache.org/jira/browse/WAGON-567?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16969308#comment-16969308
]
Michael Osipov commented on WAGON-567:
--
[~hboutemy], I will go on detail on this later this day as soon as I have
access to my dev machine. But right away, as {{none}} is the default impl no
retried are retried. Though, we have a retry handler in place for transport
errors. See also HTTPCLIENT-2019. I plan to rework this in HttpClient and have
constent handling.
> Provide request retry strategy on transient client and server side errors
> -
>
> Key: WAGON-567
> URL: https://issues.apache.org/jira/browse/WAGON-567
> Project: Maven Wagon
> Issue Type: New Feature
> Components: wagon-http
>Reporter: xueqian zhang
>Assignee: Michael Osipov
>Priority: Minor
> Fix For: 3.3.4
>
> Time Spent: 20m
> Remaining Estimate: 0h
>
> Currently wagon has a retry mechanism when there are connections issues with
> repository manager like maven central, artifactory or nexus, see [wagon
> source|https://github.com/apache/maven-wagon/blob/dfd22586b6b934aa5a652ca32d57ed26067432fd/wagon-providers/wagon-http-shared/src/main/java/org/apache/maven/wagon/shared/http/AbstractHttpClientWagon.java#L393]
> and [httpclient
> source|https://github.com/apache/httpcomponents-client/blob/4.5.x/httpclient/src/main/java/org/apache/http/impl/execchain/RetryExec.java#L90],
> [httpclient default retry
> handler|https://github.com/apache/httpcomponents-client/blob/4.5.x/httpclient/src/main/java/org/apache/http/impl/client/DefaultHttpRequestRetryHandler.java#L154].
> However, it only captures low level connection issues. When the repo
> returns errors like 500 or 503, it does not retry. The 5xx response from the
> repo might be a blip and maven should retry before giving up.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
[jira] [Commented] (WAGON-567) Provide request retry strategy on transient client and server side errors
[ https://issues.apache.org/jira/browse/WAGON-567?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16969245#comment-16969245 ] Herve Boutemy commented on WAGON-567: - [~michael-o] just had a look at this commit Can you confirm that by default (= no configuration anywhere, like in Maven core by default), no new retry will be added vs the previous behaviour? then IIUC, additionnal behaviour (for those who know...) can be configured via system properties (names can be found in code) I'm confused: 1. between "none" vs "default" vs "standard" names of the predefined strategies 2. and the default choice that looks to be "none" > Provide request retry strategy on transient client and server side errors > - > > Key: WAGON-567 > URL: https://issues.apache.org/jira/browse/WAGON-567 > Project: Maven Wagon > Issue Type: New Feature > Components: wagon-http >Reporter: xueqian zhang >Assignee: Michael Osipov >Priority: Minor > Fix For: 3.3.4 > > Time Spent: 20m > Remaining Estimate: 0h > > Currently wagon has a retry mechanism when there are connections issues with > repository manager like maven central, artifactory or nexus, see [wagon > source|https://github.com/apache/maven-wagon/blob/dfd22586b6b934aa5a652ca32d57ed26067432fd/wagon-providers/wagon-http-shared/src/main/java/org/apache/maven/wagon/shared/http/AbstractHttpClientWagon.java#L393] > and [httpclient > source|https://github.com/apache/httpcomponents-client/blob/4.5.x/httpclient/src/main/java/org/apache/http/impl/execchain/RetryExec.java#L90], > [httpclient default retry > handler|https://github.com/apache/httpcomponents-client/blob/4.5.x/httpclient/src/main/java/org/apache/http/impl/client/DefaultHttpRequestRetryHandler.java#L154]. > However, it only captures low level connection issues. When the repo > returns errors like 500 or 503, it does not retry. The 5xx response from the > repo might be a blip and maven should retry before giving up. -- This message was sent by Atlassian Jira (v8.3.4#803005)
