[ https://issues.apache.org/jira/browse/SOLR-9590?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15542693#comment-15542693 ]
Shawn Heisey commented on SOLR-9590: ------------------------------------ For success, a first step on the breadcrumb trail needs to be a filename that will be identical on every installation. It must be a filename that will not ever be confused for any other purpose. The info in that file can be used directly, or as a way to locate further configuration information in other config files. Automation is the whole idea, so this file would not normally be hand-edited. Here's some ideas that I came up with for the filename: /etc/default/solr.crumbs /etc/solr.crumbs In an actual install directory, we can have a file with a name like "solr.servicename" that contains the name of the service, which is given to the installation script as the -s option. Although in most cases I would expect there to only be one service installed, the crumbs file could contain info about multiple installs. The contents might look something like this: {quote} solr.servicenames=solr,solr6,mysolr solr.solr.zkHost=localhost:9984 solr.solr.port=8984 solr.solr.install=/opt/solr solr.solr.home=/var/solr solr.solr.zkPort=9984 solr.solr6.zkHost=zoo1.example.com:2181,zoo2.example.com:2181,zoo3.example.com:2181 solr.solr6.port=8983 solr.solr6.install=/opt/solr6 solr.solr6.home=/var/solr6 solr.mysolr.port=8988 solr.mysolr.install=/opt/mysolr solr.mysolr.home=/var/mysolr {quote} > Service installation -- save breadcrumbs for other scripts to use > ----------------------------------------------------------------- > > Key: SOLR-9590 > URL: https://issues.apache.org/jira/browse/SOLR-9590 > Project: Solr > Issue Type: New Feature > Security Level: Public(Default Security Level. Issues are Public) > Components: scripts and tools > Reporter: Shawn Heisey > Priority: Minor > > When I opened SOLR-7826, I brought up the idea of installation breadcrumbs. > If we had good breadcrumb data saved in the install directory by the install > script, a number of other scripts could use the breadcrumbs to gather > relevant data about the *service* installation, for additional safety and > more automatic operation. > The "bin/solr create" command could verify that it is running as the exact > same user that installed Solr, and abort if they don't match. > What if zkcli.sh (and bin/solr zookeeper options) no longer needed to be told > where zookeeper was, because it could find its way to > /etc/default/<service>.in.sh or $SOLR_HOME/solr.xml and grab zkHost from > there? The same thing could happen for zkHost in the idea that I filed as > SOLR-9587. -- This message was sent by Atlassian JIRA (v6.3.4#6332) --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org