Helpful Information for Solr Support Resources

See UsingMailingLists for more basic information.

Think of the information that this page describes as a "support package". It will have a very large amount of information that can be helpful to people on the mailing list or in the IRC channel who are trying to assist with a problem.

One of the most important pieces of information that gets left out of support requests is the version. Complete log entries are also useful, and you may want to actually include the entire solr logfile.

Sharing large pieces of information

The mailing list typically will not accept attachments. IRC has no built-in file-transferring mechanisms. For this reason, it is important to utilize other resources for sharing large amounts of information or binary data like images. Many services exist for this. These services are typically utilized by uploading the information to a website, which will then provide a URL for accessing what was uploaded.

Apache runs a paste website that can accept large amounts of text, or binary files like images that are up to 1MB in size. If you use this service, be sure to specify a one month expiration, so that the service will not get overloaded and require maintenance. Maintenance typically involves deleting all the stored information.

There are also many other paste websites and file sharing services. One service that is particularly good for sharing files is Dropbox.

Info Gathering Script for UNIX-like Operating Systems

This script assumes that you have the curl utility installed and on the basic system path:

#!/bin/sh

######################################################
#- Modify these variables to suit local requirements:

SOLR_HOST_PORT="http://server:port"
SOLR_SUPPORT_TMP_DIR=solrsupportpackage
SOLR_ARCHIVE_PREFIX=solrsupportpackage

#####################
#- No changes below.

OLD_PWD=`pwd`
SOLR_ARCHIVE=${OLD_PWD}/${SOLR_ARCHIVE_PREFIX}.tar.gz
mkdir -p ${SOLR_SUPPORT_TMP_DIR}
cd ${SOLR_SUPPORT_TMP_DIR}

rm -f system.info.txt
rm -f core.info.txt
rm -f zk.info.txt
curl -o system.info.txt "${SOLR_HOST_PORT}/solr/admin/info/system?wt=json&indent=true"
curl -o core.info.txt "${SOLR_HOST_PORT}/solr/admin/info/system?wt=json&indent=true"
curl -o zk.info.txt "${SOLR_HOST_PORT}/solr/admin/zookeeper?detail=true&wt=json"

rm -f ${SOLR_ARCHIVE}
tar czf ${SOLR_ARCHIVE} *
cd ${OLD_PWD}
rm -rf ${SOLR_SUPPORT_TMP_DIR}

#######
#- END

This script will create an archive in the current directory, named "solrsupportpackage.tar.gz" by default, which you can share with those who are trying to help you. See an earlier section of this page for information about sharing files.

If you're running on Windows, there is no equivalent script. You can instead grab the info from these three URLs and share it using methods discussed above.

http://server:port/solr/admin/info/system?wt=json&indent=true http://server:port/solr/admin/info/system?wt=json&indent=true http://server:port/solr/admin/zookeeper?detail=true&wt=json

Other information that can be extremely useful

Below are other pieces of information that can be helpful. You might provide these pieces of information separately from what's described above, or you might include them in a single archive.

Screenshot of a process list sorted by memory usage

Once the correct information is shown on your screen, take a screen capture (or even take a picture of the screen with a cell phone or other digital camera), then share the image as discussed earlier.

The solrconfig.xml and schema files

It is very helpful to have the config and schema from the core/collection you are trying to use. The config file is solrconfig.xml, and the schema will usually be named either managed-schema or schema.xml, depending on the Solr version and whether or not solrconfig.xml has been changed.

Info being indexed into Solr

It may be necessary to see the exact information that Solr is receiving from the indexing source, to understand how the schema might change this information.

Precise queries being made

Having an exact copy of the query request that is being made, including the URL path and all parameters, can be vital to figuring out why a query isn't working as expected.

Longevity of shared URLs

Anybody replying to a message with sharing URLs in them should keep in mind that messages to the Solr mailing list remain in the archive forever. URLs generated by paste websites and file sharing services probably will NOT live forever. For that reason, it will be critical in all replies to quote the pieces of information seen in the shared info that were seen as problematic, so that people who find historical messages in the archive will know what information was necessary to figure out the problem, even if the shared URL no longer functions.

HelpfulInfoForSupport (last edited 2017-03-27 20:19:40 by ShawnHeisey)