Dgraph flags

The Oracle Endeca Server starts the Dgraph process for each Endeca data store.

When you create or attach an Endeca data store, you can optionally specify that its Dgraph process start with any of the flags in the following table. These Dgraph flags allow you to adjust its configuration.

To obtain a listing of the Dgraph process arguments at the Oracle Endeca Server host and port, issue the create-ds or the attach-ds command with the --args --usage flag, as in this example for the create-ds command:
endeca-cmd create-ds --args --usage
Flag Description
? Print the help message and exit.
-v Verbose mode.

Print information about each request to stdout.

--ancestor_counts Compute counts for root managed attribute values and any intermediate managed attribute value selections.

By default, the Dgraph only computes refinement counts for proper refinements (in other words, for actual managed attribute values).

It does not compute counts for root managed attribute values or for any intermediate managed attribute value selections.

--backlog-timeout <seconds> Specify the wait limit (in seconds) for a query that has been read and queued for processing.

This is the maximum number of seconds that a query is allowed to spend waiting in the processing queue before the Dgraph responds with a timeout message.

The default value is 0 seconds.

--cmem <MB> Specify an absolute value in MB for the Dgraph cache.

When an absolute value is not specified with the --cmem flag, the default Dgraph cache size is computed as 10% of the amount of RAM available in the system.

--coordinator_host <host name> Specify the host name of the server on which the Cluster Coordinator service is running.

You specify this flag along with the --coordinator_port flag when you start the Dgraph as one of the nodes in the cluster.

--coordinator_port <num> Specify the port of the server on which the Cluster Coordinator service is running.

The Cluster Coordinator expects that you specify the port 2181 (if you specify another port, changes to the Cluster Coordinator configuration file are required).

You specify this flag along with the --coordinator_host flag when you start the Dgraph as one of the nodes in the cluster.

--disable_fast_aspell Disable fast mode for the aspell spelling module. If you disable fast mode, it decreases the performance of the spelling correction, but may allow additional queries to be corrected.

When the fast mode is enabled, it can significantly speed up applications that use spelling correction features with the aspell module. The fast mode is used by default.

--esampmin <num> Specify the minimum number of records to sample during refinement computation. The default is 0.
Tuning recommendations:
  • For most applications, larger values reduce performance without improving dynamic refinement ranking quality.
  • For some applications with extremely large, non-hierarchical managed attributes (if they cannot be avoided), larger values can meaningfully improve dynamic refinement ranking quality with minor performance cost.
--follower <name> Specify the name of the node that should serve as one of the follower nodes in the cluster. This name must be unique across the cluster, and must also be a valid directory name (characters such as slashes (/) are not allowed).

You can start more than one node in the cluster with this command, thus designating more than one follower node.

Before starting the Dgraph with this command flag, ensure that the Cluster Coordinator service is running on the server that serves as the leader node.

All nodes must be able to connect to the Cluster Coordinator. Therefore, when you specify a follower node with the --follower flag, also specify for the follower node the host name and port of the Cluster Coordinator service using the --coordinator_host and --coordinator_port commands.

For a list of operations allowed in the Dgraph when you use the --follower flag, see Deploying a Cluster of Oracle Endeca Servers.

Note: If you start a node without the --follower flag, the Cluster Coordinator assumes this is the leader node. Since there can be one and only one leader node in the cluster, the Dgraph will not start if it is asked to be the leader node when a leader node already exists.
--help Print the help message and exit.
--implicit_exact Disable approximate computation of implicit refinements.

Use of this option is not recommended.

If this option is not enabled, managed attribute values without full coverage of the current result record set may sometimes be returned as implicit refinements, although the probability of such "false" implicit refinements is minuscule.

--implicit_sample <num> Set the maximum number of records to sample when computing implicit refinements (which are a performance tuning parameter). The default value is 1024.
--latin1 Ignore character accents when handling search requests, and use ISO Latin 1 character mappings when processing search requests.
--log <path> Specify the path for the Dgraph request log file. The default log file is named datastore.reqlog and is located in the Endeca Server's logs directory.
--net-timeout <num> Specify the maximum number of seconds the Dgraph waits for the client to download data from queries across the network. The default network timeout value is 30 seconds.
--out <stdout/stderr file> Specify file path to which stdout/stderr should be remapped. The default stdout/stderr file is named datastore.out and is located in the Endeca Server's logs directory.
--pidfile <pidfile-path> Specify the file to which to write the process ID (pid). The default PID file is named datastore.pid and is located in the Endeca Server's logs directory.
--read-only Provide read-only access to the data files for the data store. If you start the Dgraph with this flag, it performs only read-only operations. Any operations that attempt to write to the data files are rejected by the Oracle Endeca Server, and return an HTTP status code 403.

This flag can be useful in implementations with multiple staging and production environments. You can also use this flag for hosting demos on public-facing web sites.

If you use the --read-only flag, you cannot use the --follower flag at the same time.

--search_max <num> Specify the maximum number of terms for text search. Default is 10.
--snip_cutoff <num> Limit the number of words in an attribute that the Dgraph evaluates to identify the snippet.

If a match is not found within <num> words, the Dgraph does not return a snippet, even if a match occurs later in the attribute value.

If the flag is not specified, or <num> is not specified, the default is 500.

--snip_disable Globally disable snippeting.
--sslcafile <CA-certfile-path> Specify the path of the eneCA.pem Certificate Authority file that the Dgraph will use to authenticate SSL communications with other components that must communicate with the Dgraph.

If not given, SSL mutual authentication is not performed.

--sslcertfile <certfile-path> Specify the path of the eneCert.pem certificate file that will be used by the Dgraph to present to any client for SSL communications. If not given, SSL is not enabled for Dgraph communications.
--sslcipher <cipher-list> Set one or more cipher names (such as RC4-SHA) that specify the minimum cryptographic algorithm that the Dgraph will use during the SSL negotiation. If multiple ciphers are specified, the names must be separated by colons.
--stat-all Enable all available dynamic attribute value characteristics. Note that this option has performance implications and is not intended for production use.
--stat-brel Create dynamic record attributes indicating the relevance rank assigned to full-text search result records.
--syslog Direct all output to syslog.
--thesaurus_cutoff <limit> Set a limit on the number of words in a user’s search query that are subject to thesaurus replacement. If more terms than this number match thesaurus entries, none of the terms are thesaurus expanded.

The default value of <limit> is 3. This means that up to 3 words in a user’s search query can be replaced with thesaurus entries.

This option is intended as a performance guard against very expensive thesaurus queries. Lower values improve thesaurus engine performance.

--thesaurus_multiword_nostem Specify that words in a multiple-word thesaurus form should be treated like phrases and should not be stemmed, which increases performance for some query loads.

Single-word terms are subject to stemming regardless of whether this flag is specified.

This flag prevents the Dgraph from expanding multi-word thesaurus forms by stemming. Thesaurus entries continue to match any stemmed form in the query, but multi-word expansions only include explicitly listed forms. To get the multi-word stemmed thesaurus expansions, the various forms must be listed explicitly in the thesaurus.

--threads <num> Specify the number of threads in the Dgraph threading pool.

The value of <num> must be a positive integer (that is, 1 or greater).

The default for num is 2.

The recommended number of threads for the Dgraph is typically equal to the number of cores on the host machine.

--unctrct Specify to the Dgraph not to compute implicit managed attributes, and to only compute and present explicitly specified managed attributes, when displaying refinements in navigation results.

Specifying this flag does not reduce the size of the resulting record set that is being displayed; however, it improves run-time performance of the Dgraph process.

Be aware that if you use this flag, in order to receive meaningful navigation refinements, you need to make top-level precedence rules work for ALL outbound queries.

--validate_data Validate that all processed data loads and then exit.
--version Print version information and then exit. This includes both the Oracle Endeca Server version and the internal Dgraph identifier.
--wildcard_max <count> Specify the maximum number of terms that can match a wildcard term in a wildcard query that contains punctuation, such as ab*c.def*. The default is 100.
--whymatch Enable computation of Why Did It Match dynamic record attributes returned as results of full-text search queries.

These dynamic attributes contain a copy of the attribute key and value that caused the match, along with query interpretation notes (spelling, thesaurus, and so on).

--whymatchConcise Similar to --whymatch, but produces more concise dynamic attribute values containing only the attribute key and query interpretation notes.

This is useful when the attribute value might include large amounts of text, such as document contents.

--wordinterp Enable computation of word interpretation dynamic supplement (or see-also) objects, which report on alternate forms of user query terms considered by the text search engine while processing full-text (record) search requests.
--xquery_fndoc <mode> Specifies the handling of the fn:doc() function within XQuery.
The following values are supported:
  • none causes all calls to fn:doc() to fail.
  • sandbox allows fn:doc(), but interprets its argument as a relative path within the XML subdirectory of the XQuery service directory.
  • open allows fn:doc() and interprets its argument as a URL. Note that open is not supported for use in deployed applications.

If not specified, defaults to sandbox.

--xquery_path <path> Specify the directory in which XQuery Web service resources are located. (This Web service is internal to the Dgraph and should not be used directly). XQuery main modules and WSDL files are loaded from this directory. If not specified, a user XQuery path is not used.