[ https://issues.apache.org/jira/browse/NUTCH-2015?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14547213#comment-14547213 ]
Sebastian Nagel commented on NUTCH-2015: ---------------------------------------- Ok. Ev. this could be changed to make it more clear: a check isRunning() inside a loop suggests that the state of the server can change between running and stopped. That's (currently) not the case and a boolean field variable seems to more verbose. In case the server could be stopped it should be: {code} if (NutchServer.getInstance().isRunning()) this.fetchNode = new FetchNode(); else this.fetchNode = null; {code} Since currently fetchNode is only used with a parsing fetcher this could be also checked, e.g. {code} if (parsing && NutchServer.getInstance().isRunning()) { reportToNutchServer = true; } {code} > Make FetchNodeDb optional (off by default) if NutchServer is not used > --------------------------------------------------------------------- > > Key: NUTCH-2015 > URL: https://issues.apache.org/jira/browse/NUTCH-2015 > Project: Nutch > Issue Type: Sub-task > Components: fetcher, REST_api > Reporter: Sujen Shah > Labels: memex > Fix For: 1.11 > > > Currently, the FetchNodes are created even if the NutchServer is not used > causing memory exceptions. This patch makes the fetcher report to the > FetchNodeDb only if the crawl is invoked from the REST service (ie > NutchServer) -- This message was sent by Atlassian JIRA (v6.3.4#6332)