1 <chapter id="administration">
2 <title>Administrating &zebra;</title>
3 <!-- ### It's a bit daft that this chapter (which describes half of
4 the configuration-file formats) is separated from
5 "recordmodel-grs.xml" (which describes the other half) by the
6 instructions on running zebraidx and zebrasrv. Some careful
7 re-ordering is required here.
11 Unlike many simpler retrieval systems, &zebra; supports safe, incremental
12 updates to an existing index.
16 Normally, when &zebra; modifies the index it reads a number of records
18 Depending on your specifications and on the contents of each record
19 one the following events take place for each record:
26 The record is indexed as if it never occurred before.
27 Either the &zebra; system doesn't know how to identify the record or
28 &zebra; can identify the record but didn't find it to be already indexed.
36 The record has already been indexed.
37 In this case either the contents of the record or the location
38 (file) of the record indicates that it has been indexed before.
46 The record is deleted from the index. As in the
47 update-case it must be able to identify the record.
55 Please note that in both the modify- and delete- case the &zebra;
56 indexer must be able to generate a unique key that identifies the record
57 in question (more on this below).
61 To administrate the &zebra; retrieval system, you run the
62 <literal>zebraidx</literal> program.
63 This program supports a number of options which are preceded by a dash,
64 and a few commands (not preceded by dash).
68 Both the &zebra; administrative tool and the &acro.z3950; server share a
69 set of index files and a global configuration file.
70 The name of the configuration file defaults to
71 <literal>zebra.cfg</literal>.
72 The configuration file includes specifications on how to index
73 various kinds of records and where the other configuration files
74 are located. <literal>zebrasrv</literal> and <literal>zebraidx</literal>
75 <emphasis>must</emphasis> be run in the directory where the
76 configuration file lives unless you indicate the location of the
77 configuration file by option <literal>-c</literal>.
80 <sect1 id="record-types">
81 <title>Record Types</title>
84 Indexing is a per-record process, in which either insert/modify/delete
85 will occur. Before a record is indexed search keys are extracted from
86 whatever might be the layout the original record (sgml,html,text, etc..).
87 The &zebra; system currently supports two fundamental types of records:
88 structured and simple text.
89 To specify a particular extraction process, use either the
90 command line option <literal>-t</literal> or specify a
91 <literal>recordType</literal> setting in the configuration file.
96 <sect1 id="zebra-cfg">
97 <title>The &zebra; Configuration File</title>
100 The &zebra; configuration file, read by <literal>zebraidx</literal> and
101 <literal>zebrasrv</literal> defaults to <literal>zebra.cfg</literal>
102 unless specified by <literal>-c</literal> option.
106 You can edit the configuration file with a normal text editor.
107 parameter names and values are separated by colons in the file. Lines
108 starting with a hash sign (<literal>#</literal>) are
113 If you manage different sets of records that share common
114 characteristics, you can organize the configuration settings for each
116 When <literal>zebraidx</literal> is run and you wish to address a
117 given group you specify the group name with the <literal>-g</literal>
119 In this case settings that have the group name as their prefix
120 will be used by <literal>zebraidx</literal>.
121 If no <literal>-g</literal> option is specified, the settings
122 without prefix are used.
126 In the configuration file, the group name is placed before the option
127 name itself, separated by a dot (.). For instance, to set the record type
128 for group <literal>public</literal> to <literal>grs.sgml</literal>
129 (the &acro.sgml;-like format for structured records) you would write:
134 public.recordType: grs.sgml
139 To set the default value of the record type to <literal>text</literal>
150 The available configuration settings are summarized below. They will be
151 explained further in the following sections.
155 FIXME - Didn't Adam make something to have multiple databases in multiple dirs...
163 <emphasis>group</emphasis>
164 .recordType[<emphasis>.name</emphasis>]:
165 <replaceable>type</replaceable>
169 Specifies how records with the file extension
170 <emphasis>name</emphasis> should be handled by the indexer.
171 This option may also be specified as a command line option
172 (<literal>-t</literal>). Note that if you do not specify a
173 <emphasis>name</emphasis>, the setting applies to all files.
174 In general, the record type specifier consists of the elements (each
175 element separated by dot), <emphasis>fundamental-type</emphasis>,
176 <emphasis>file-read-type</emphasis> and arguments. Currently, two
177 fundamental types exist, <literal>text</literal> and
178 <literal>grs</literal>.
183 <term><emphasis>group</emphasis>.recordId:
184 <replaceable>record-id-spec</replaceable></term>
187 Specifies how the records are to be identified when updated. See
188 <xref linkend="locating-records"/>.
193 <term><emphasis>group</emphasis>.database:
194 <replaceable>database</replaceable></term>
197 Specifies the &acro.z3950; database name.
198 <!-- FIXME - now we can have multiple databases in one server. -H -->
203 <term><emphasis>group</emphasis>.storeKeys:
204 <replaceable>boolean</replaceable></term>
207 Specifies whether key information should be saved for a given
208 group of records. If you plan to update/delete this type of
209 records later this should be specified as 1; otherwise it
210 should be 0 (default), to save register space.
211 <!-- ### this is the first mention of "register" -->
212 See <xref linkend="file-ids"/>.
217 <term><emphasis>group</emphasis>.storeData:
218 <replaceable>boolean</replaceable></term>
221 Specifies whether the records should be stored internally
222 in the &zebra; system files.
223 If you want to maintain the raw records yourself,
224 this option should be false (0).
225 If you want &zebra; to take care of the records for you, it
231 <!-- ### probably a better place to define "register" -->
232 <term>register: <replaceable>register-location</replaceable></term>
235 Specifies the location of the various register files that &zebra; uses
236 to represent your databases.
237 See <xref linkend="register-location"/>.
242 <term>shadow: <replaceable>register-location</replaceable></term>
245 Enables the <emphasis>safe update</emphasis> facility of &zebra;, and
246 tells the system where to place the required, temporary files.
247 See <xref linkend="shadow-registers"/>.
252 <term>lockDir: <replaceable>directory</replaceable></term>
255 Directory in which various lock files are stored.
260 <term>keyTmpDir: <replaceable>directory</replaceable></term>
263 Directory in which temporary files used during zebraidx's update
269 <term>setTmpDir: <replaceable>directory</replaceable></term>
272 Specifies the directory that the server uses for temporary result sets.
273 If not specified <literal>/tmp</literal> will be used.
278 <term>profilePath: <replaceable>path</replaceable></term>
281 Specifies a path of profile specification files.
282 The path is composed of one or more directories separated by
283 colon. Similar to <literal>PATH</literal> for UNIX systems.
289 <term>modulePath: <replaceable>path</replaceable></term>
292 Specifies a path of record filter modules.
293 The path is composed of one or more directories separated by
294 colon. Similar to <literal>PATH</literal> for UNIX systems.
295 The 'make install' procedure typically puts modules in
296 <filename>/usr/local/lib/idzebra-2.0/modules</filename>.
302 <term>index: <replaceable>filename</replaceable></term>
305 Defines the filename which holds fields structure
306 definitions. If omitted, the file <filename>default.idx</filename>
308 Refer to <xref linkend="default-idx-file"/> for
315 <term>staticrank: <replaceable>integer</replaceable></term>
318 Enables whether static ranking is to be enabled (1) or
319 disabled (0). If omitted, it is disabled - corresponding
321 Refer to <xref linkend="administration-ranking-static"/> .
328 <term>estimatehits:: <replaceable>integer</replaceable></term>
331 Controls whether &zebra; should calculate approximite hit counts and
332 at which hit count it is to be enabled.
333 A value of 0 disables approximiate hit counts.
334 For a positive value approximaite hit count is enabled
335 if it is known to be larger than <replaceable>integer</replaceable>.
338 Approximate hit counts can also be triggered by a particular
339 attribute in a query.
340 Refer to <xref linkend="querymodel-zebra-global-attr-limit"/>.
346 <term>attset: <replaceable>filename</replaceable></term>
349 Specifies the filename(s) of attribute set files for use in
350 searching. In many configurations <filename>bib1.att</filename>
351 is used, but that is not required. If Classic Explain
352 attributes is to be used for searching,
353 <filename>explain.att</filename> must be given.
354 The path to att-files in general can be given using
355 <literal>profilePath</literal> setting.
356 See also <xref linkend="attset-files"/>.
361 <term>memMax: <replaceable>size</replaceable></term>
364 Specifies <replaceable>size</replaceable> of internal memory
365 to use for the zebraidx program.
366 The amount is given in megabytes - default is 4 (4 MB).
367 The more memory, the faster large updates happen, up to about
368 half the free memory available on the computer.
373 <term>tempfiles: <replaceable>Yes/Auto/No</replaceable></term>
376 Tells zebra if it should use temporary files when indexing. The
377 default is Auto, in which case zebra uses temporary files only
378 if it would need more that <replaceable>memMax</replaceable>
379 megabytes of memory. This should be good for most uses.
385 <term>root: <replaceable>dir</replaceable></term>
388 Specifies a directory base for &zebra;. All relative paths
389 given (in profilePath, register, shadow) are based on this
390 directory. This setting is useful if your &zebra; server
391 is running in a different directory from where
392 <literal>zebra.cfg</literal> is located.
398 <term>passwd: <replaceable>file</replaceable></term>
401 Specifies a file with description of user accounts for &zebra;.
402 The format is similar to that known to Apache's htpasswd files
403 and UNIX' passwd files. Non-empty lines not beginning with
404 # are considered account lines. There is one account per-line.
405 A line consists of fields separate by a single colon character.
406 First field is username, second is password.
412 <term>passwd.c: <replaceable>file</replaceable></term>
415 Specifies a file with description of user accounts for &zebra;.
416 File format is similar to that used by the passwd directive except
417 that the password are encrypted. Use Apache's htpasswd or similar
424 <term>perm.<replaceable>user</replaceable>:
425 <replaceable>permstring</replaceable></term>
428 Specifies permissions (priviledge) for a user that are allowed
429 to access &zebra; via the passwd system. There are two kinds
430 of permissions currently: read (r) and write(w). By default
431 users not listed in a permission directive are given the read
432 privilege. To specify permissions for a user with no
433 username, or &acro.z3950; anonymous style use
434 <literal>anonymous</literal>. The permstring consists of
435 a sequence of characters. Include character <literal>w</literal>
436 for write/update access, <literal>r</literal> for read access and
437 <literal>a</literal> to allow anonymous access through this account.
443 <term>dbaccess <replaceable>accessfile</replaceable></term>
446 Names a file which lists database subscriptions for individual users.
447 The access file should consists of lines of the form <literal>username:
448 dbnames</literal>, where dbnames is a list of database names, seprated by
449 '+'. No whitespace is allowed in the database list.
459 <sect1 id="locating-records">
460 <title>Locating Records</title>
463 The default behavior of the &zebra; system is to reference the
464 records from their original location, i.e. where they were found when you
465 run <literal>zebraidx</literal>.
466 That is, when a client wishes to retrieve a record
467 following a search operation, the files are accessed from the place
468 where you originally put them - if you remove the files (without
469 running <literal>zebraidx</literal> again, the server will return
470 diagnostic number 14 (``System error in presenting records'') to
475 If your input files are not permanent - for example if you retrieve
476 your records from an outside source, or if they were temporarily
477 mounted on a CD-ROM drive,
478 you may want &zebra; to make an internal copy of them. To do this,
479 you specify 1 (true) in the <literal>storeData</literal> setting. When
480 the &acro.z3950; server retrieves the records they will be read from the
481 internal file structures of the system.
486 <sect1 id="simple-indexing">
487 <title>Indexing with no Record IDs (Simple Indexing)</title>
490 If you have a set of records that are not expected to change over time
491 you may can build your database without record IDs.
492 This indexing method uses less space than the other methods and
497 To use this method, you simply omit the <literal>recordId</literal> entry
498 for the group of files that you index. To add a set of records you use
499 <literal>zebraidx</literal> with the <literal>update</literal> command. The
500 <literal>update</literal> command will always add all of the records that it
501 encounters to the index - whether they have already been indexed or
502 not. If the set of indexed files change, you should delete all of the
503 index files, and build a new index from scratch.
507 Consider a system in which you have a group of text files called
508 <literal>simple</literal>.
509 That group of records should belong to a &acro.z3950; database called
510 <literal>textbase</literal>.
511 The following <literal>zebra.cfg</literal> file will suffice:
516 profilePath: /usr/local/idzebra/tab
518 simple.recordType: text
519 simple.database: textbase
525 Since the existing records in an index can not be addressed by their
526 IDs, it is impossible to delete or modify records when using this method.
531 <sect1 id="file-ids">
532 <title>Indexing with File Record IDs</title>
535 If you have a set of files that regularly change over time: Old files
536 are deleted, new ones are added, or existing files are modified, you
537 can benefit from using the <emphasis>file ID</emphasis>
538 indexing methodology.
539 Examples of this type of database might include an index of WWW
540 resources, or a USENET news spool area.
541 Briefly speaking, the file key methodology uses the directory paths
542 of the individual records as a unique identifier for each record.
543 To perform indexing of a directory with file keys, again, you specify
544 the top-level directory after the <literal>update</literal> command.
545 The command will recursively traverse the directories and compare
546 each one with whatever have been indexed before in that same directory.
547 If a file is new (not in the previous version of the directory) it
548 is inserted into the registers; if a file was already indexed and
549 it has been modified since the last update, the index is also
550 modified; if a file has been removed since the last
551 visit, it is deleted from the index.
555 The resulting system is easy to administrate. To delete a record you
556 simply have to delete the corresponding file (say, with the
557 <literal>rm</literal> command). And to add records you create new
558 files (or directories with files). For your changes to take effect
559 in the register you must run <literal>zebraidx update</literal> with
560 the same directory root again. This mode of operation requires more
561 disk space than simpler indexing methods, but it makes it easier for
562 you to keep the index in sync with a frequently changing set of data.
563 If you combine this system with the <emphasis>safe update</emphasis>
564 facility (see below), you never have to take your server off-line for
565 maintenance or register updating purposes.
569 To enable indexing with pathname IDs, you must specify
570 <literal>file</literal> as the value of <literal>recordId</literal>
571 in the configuration file. In addition, you should set
572 <literal>storeKeys</literal> to <literal>1</literal>, since the &zebra;
573 indexer must save additional information about the contents of each record
574 in order to modify the indexes correctly at a later time.
578 FIXME - There must be a simpler way to do this with Adams string tags -H
582 For example, to update records of group <literal>esdd</literal>
584 <literal>/data1/records/</literal> you should type:
586 $ zebraidx -g esdd update /data1/records
591 The corresponding configuration file includes:
594 esdd.recordType: grs.sgml
600 <para>You cannot start out with a group of records with simple
601 indexing (no record IDs as in the previous section) and then later
602 enable file record Ids. &zebra; must know from the first time that you
604 the files should be indexed with file record IDs.
609 You cannot explicitly delete records when using this method (using the
610 <literal>delete</literal> command to <literal>zebraidx</literal>. Instead
611 you have to delete the files from the file system (or move them to a
613 and then run <literal>zebraidx</literal> with the
614 <literal>update</literal> command.
616 <!-- ### what happens if a file contains multiple records? -->
619 <sect1 id="generic-ids">
620 <title>Indexing with General Record IDs</title>
623 When using this method you construct an (almost) arbitrary, internal
624 record key based on the contents of the record itself and other system
625 information. If you have a group of records that explicitly associates
626 an ID with each record, this method is convenient. For example, the
627 record format may contain a title or a ID-number - unique within the group.
628 In either case you specify the &acro.z3950; attribute set and use-attribute
629 location in which this information is stored, and the system looks at
630 that field to determine the identity of the record.
634 As before, the record ID is defined by the <literal>recordId</literal>
635 setting in the configuration file. The value of the record ID specification
636 consists of one or more tokens separated by whitespace. The resulting
637 ID is represented in the index by concatenating the tokens and
638 separating them by ASCII value (1).
642 There are three kinds of tokens:
646 <term>Internal record info</term>
649 The token refers to a key that is
650 extracted from the record. The syntax of this token is
651 <literal>(</literal> <emphasis>set</emphasis> <literal>,</literal>
652 <emphasis>use</emphasis> <literal>)</literal>,
653 where <emphasis>set</emphasis> is the
654 attribute set name <emphasis>use</emphasis> is the
655 name or value of the attribute.
660 <term>System variable</term>
663 The system variables are preceded by
668 and immediately followed by the system variable name, which
681 <term>database</term>
684 Current database specified.
701 <term>Constant string</term>
704 A string used as part of the ID — surrounded
705 by single- or double quotes.
713 For instance, the sample GILS records that come with the &zebra;
714 distribution contain a unique ID in the data tagged Control-Identifier.
715 The data is mapped to the &acro.bib1; use attribute Identifier-standard
716 (code 1007). To use this field as a record id, specify
717 <literal>(bib1,Identifier-standard)</literal> as the value of the
718 <literal>recordId</literal> in the configuration file.
719 If you have other record types that uses the same field for a
720 different purpose, you might add the record type
721 (or group or database name) to the record id of the gils
722 records as well, to prevent matches with other types of records.
723 In this case the recordId might be set like this:
726 gils.recordId: $type (bib1,Identifier-standard)
732 (see <xref linkend="grs"/>
733 for details of how the mapping between elements of your records and
734 searchable attributes is established).
738 As for the file record ID case described in the previous section,
739 updating your system is simply a matter of running
740 <literal>zebraidx</literal>
741 with the <literal>update</literal> command. However, the update with general
742 keys is considerably slower than with file record IDs, since all files
743 visited must be (re)read to discover their IDs.
747 As you might expect, when using the general record IDs
748 method, you can only add or modify existing records with the
749 <literal>update</literal> command.
750 If you wish to delete records, you must use the,
751 <literal>delete</literal> command, with a directory as a parameter.
752 This will remove all records that match the files below that root
758 <sect1 id="register-location">
759 <title>Register Location</title>
762 Normally, the index files that form dictionaries, inverted
763 files, record info, etc., are stored in the directory where you run
764 <literal>zebraidx</literal>. If you wish to store these, possibly large,
765 files somewhere else, you must add the <literal>register</literal>
766 entry to the <literal>zebra.cfg</literal> file.
767 Furthermore, the &zebra; system allows its file
768 structures to span multiple file systems, which is useful for
769 managing very large databases.
773 The value of the <literal>register</literal> setting is a sequence
774 of tokens. Each token takes the form:
777 <emphasis>dir</emphasis><literal>:</literal><emphasis>size</emphasis>
780 The <emphasis>dir</emphasis> specifies a directory in which index files
781 will be stored and the <emphasis>size</emphasis> specifies the maximum
782 size of all files in that directory. The &zebra; indexer system fills
783 each directory in the order specified and use the next specified
784 directories as needed.
785 The <emphasis>size</emphasis> is an integer followed by a qualifier
787 <literal>b</literal> for bytes,
788 <literal>k</literal> for kilobytes.
789 <literal>M</literal> for megabytes,
790 <literal>G</literal> for gigabytes.
791 Specifying a negative value disables the checking (it still needs the unit,
792 use <literal>-1b</literal>).
796 For instance, if you have allocated three disks for your register, and
797 the first disk is mounted
798 on <literal>/d1</literal> and has 2GB of free space, the
799 second, mounted on <literal>/d2</literal> has 3.6 GB, and the third,
800 on which you have more space than you bother to worry about, mounted on
801 <literal>/d3</literal> you could put this entry in your configuration file:
804 register: /d1:2G /d2:3600M /d3:-1b
809 Note that &zebra; does not verify that the amount of space specified is
810 actually available on the directory (file system) specified - it is
811 your responsibility to ensure that enough space is available, and that
812 other applications do not attempt to use the free space. In a large
813 production system, it is recommended that you allocate one or more
814 file system exclusively to the &zebra; register files.
819 <sect1 id="shadow-registers">
820 <title>Safe Updating - Using Shadow Registers</title>
822 <sect2 id="shadow-registers-description">
823 <title>Description</title>
826 The &zebra; server supports <emphasis>updating</emphasis> of the index
827 structures. That is, you can add, modify, or remove records from
828 databases managed by &zebra; without rebuilding the entire index.
829 Since this process involves modifying structured files with various
830 references between blocks of data in the files, the update process
831 is inherently sensitive to system crashes, or to process interruptions:
832 Anything but a successfully completed update process will leave the
833 register files in an unknown state, and you will essentially have no
834 recourse but to re-index everything, or to restore the register files
835 from a backup medium.
836 Further, while the update process is active, users cannot be
837 allowed to access the system, as the contents of the register files
838 may change unpredictably.
842 You can solve these problems by enabling the shadow register system in
844 During the updating procedure, <literal>zebraidx</literal> will temporarily
845 write changes to the involved files in a set of "shadow
846 files", without modifying the files that are accessed by the
847 active server processes. If the update procedure is interrupted by a
848 system crash or a signal, you simply repeat the procedure - the
849 register files have not been changed or damaged, and the partially
850 written shadow files are automatically deleted before the new updating
855 At the end of the updating procedure (or in a separate operation, if
856 you so desire), the system enters a "commit mode". First,
857 any active server processes are forced to access those blocks that
858 have been changed from the shadow files rather than from the main
859 register files; the unmodified blocks are still accessed at their
860 normal location (the shadow files are not a complete copy of the
861 register files - they only contain those parts that have actually been
862 modified). If the commit process is interrupted at any point during the
863 commit process, the server processes will continue to access the
864 shadow files until you can repeat the commit procedure and complete
865 the writing of data to the main register files. You can perform
866 multiple update operations to the registers before you commit the
867 changes to the system files, or you can execute the commit operation
868 at the end of each update operation. When the commit phase has
869 completed successfully, any running server processes are instructed to
870 switch their operations to the new, operational register, and the
871 temporary shadow files are deleted.
876 <sect2 id="shadow-registers-how-to-use">
877 <title>How to Use Shadow Register Files</title>
880 The first step is to allocate space on your system for the shadow
882 You do this by adding a <literal>shadow</literal> entry to the
883 <literal>zebra.cfg</literal> file.
884 The syntax of the <literal>shadow</literal> entry is exactly the
885 same as for the <literal>register</literal> entry
886 (see <xref linkend="register-location"/>).
887 The location of the shadow area should be
888 <emphasis>different</emphasis> from the location of the main register
889 area (if you have specified one - remember that if you provide no
890 <literal>register</literal> setting, the default register area is the
891 working directory of the server and indexing processes).
895 The following excerpt from a <literal>zebra.cfg</literal> file shows
896 one example of a setup that configures both the main register
897 location and the shadow file area.
898 Note that two directories or partitions have been set aside
899 for the shadow file area. You can specify any number of directories
900 for each of the file areas, but remember that there should be no
901 overlaps between the directories used for the main registers and the
902 shadow files, respectively.
908 shadow: /scratch1:100M /scratch2:200M
914 When shadow files are enabled, an extra command is available at the
915 <literal>zebraidx</literal> command line.
916 In order to make changes to the system take effect for the
917 users, you'll have to submit a "commit" command after a
918 (sequence of) update operation(s).
924 $ zebraidx update /d1/records
931 Or you can execute multiple updates before committing the changes:
937 $ zebraidx -g books update /d1/records /d2/more-records
938 $ zebraidx -g fun update /d3/fun-records
945 If one of the update operations above had been interrupted, the commit
946 operation on the last line would fail: <literal>zebraidx</literal>
947 will not let you commit changes that would destroy the running register.
948 You'll have to rerun all of the update operations since your last
949 commit operation, before you can commit the new changes.
953 Similarly, if the commit operation fails, <literal>zebraidx</literal>
954 will not let you start a new update operation before you have
955 successfully repeated the commit operation.
956 The server processes will keep accessing the shadow files rather
957 than the (possibly damaged) blocks of the main register files
958 until the commit operation has successfully completed.
962 You should be aware that update operations may take slightly longer
963 when the shadow register system is enabled, since more file access
964 operations are involved. Further, while the disk space required for
965 the shadow register data is modest for a small update operation, you
966 may prefer to disable the system if you are adding a very large number
967 of records to an already very large database (we use the terms
968 <emphasis>large</emphasis> and <emphasis>modest</emphasis>
969 very loosely here, since every application will have a
970 different perception of size).
971 To update the system without the use of the the shadow files,
972 simply run <literal>zebraidx</literal> with the <literal>-n</literal>
973 option (note that you do not have to execute the
974 <emphasis>commit</emphasis> command of <literal>zebraidx</literal>
975 when you temporarily disable the use of the shadow registers in
977 Note also that, just as when the shadow registers are not enabled,
978 server processes will be barred from accessing the main register
979 while the update procedure takes place.
987 <sect1 id="administration-ranking">
988 <title>Relevance Ranking and Sorting of Result Sets</title>
990 <sect2 id="administration-overview">
991 <title>Overview</title>
993 The default ordering of a result set is left up to the server,
994 which inside &zebra; means sorting in ascending document ID order.
995 This is not always the order humans want to browse the sometimes
996 quite large hit sets. Ranking and sorting comes to the rescue.
1000 In cases where a good presentation ordering can be computed at
1001 indexing time, we can use a fixed <literal>static ranking</literal>
1002 scheme, which is provided for the <literal>alvis</literal>
1003 indexing filter. This defines a fixed ordering of hit lists,
1004 independently of the query issued.
1008 There are cases, however, where relevance of hit set documents is
1009 highly dependent on the query processed.
1010 Simply put, <literal>dynamic relevance ranking</literal>
1011 sorts a set of retrieved records such that those most likely to be
1012 relevant to your request are retrieved first.
1013 Internally, &zebra; retrieves all documents that satisfy your
1014 query, and re-orders the hit list to arrange them based on
1015 a measurement of similarity between your query and the content of
1020 Finally, there are situations where hit sets of documents should be
1021 <literal>sorted</literal> during query time according to the
1022 lexicographical ordering of certain sort indexes created at
1028 <sect2 id="administration-ranking-static">
1029 <title>Static Ranking</title>
1032 &zebra; uses internally inverted indexes to look up term occurencies
1033 in documents. Multiple queries from different indexes can be
1034 combined by the binary boolean operations <literal>AND</literal>,
1035 <literal>OR</literal> and/or <literal>NOT</literal> (which
1036 is in fact a binary <literal>AND NOT</literal> operation).
1037 To ensure fast query execution
1038 speed, all indexes have to be sorted in the same order.
1041 The indexes are normally sorted according to document
1042 <literal>ID</literal> in
1043 ascending order, and any query which does not invoke a special
1044 re-ranking function will therefore retrieve the result set in
1046 <literal>ID</literal>
1054 directive in the main core &zebra; configuration file, the internal document
1055 keys used for ordering are augmented by a preceding integer, which
1056 contains the static rank of a given document, and the index lists
1058 first by ascending static rank,
1059 then by ascending document <literal>ID</literal>.
1061 is the ``best'' rank, as it occurs at the
1062 beginning of the list; higher numbers represent worse scores.
1065 The experimental <literal>alvis</literal> filter provides a
1066 directive to fetch static rank information out of the indexed &acro.xml;
1067 records, thus making <emphasis>all</emphasis> hit sets ordered
1068 after <emphasis>ascending</emphasis> static
1069 rank, and for those doc's which have the same static rank, ordered
1070 after <emphasis>ascending</emphasis> doc <literal>ID</literal>.
1071 See <xref linkend="record-model-alvisxslt"/> for the gory details.
1076 <sect2 id="administration-ranking-dynamic">
1077 <title>Dynamic Ranking</title>
1079 In order to fiddle with the static rank order, it is necessary to
1080 invoke additional re-ranking/re-ordering using dynamic
1081 ranking or score functions. These functions return positive
1082 integer scores, where <emphasis>highest</emphasis> score is
1084 hit sets are sorted according to <emphasis>descending</emphasis>
1086 to the index lists which are sorted according to
1087 ascending rank number and document ID).
1090 Dynamic ranking is enabled by a directive like one of the
1091 following in the zebra configuration file (use only one of these a time!):
1093 rank: rank-1 # default TDF-IDF like
1094 rank: rank-static # dummy do-nothing
1099 Dynamic ranking is done at query time rather than
1100 indexing time (this is why we
1101 call it ``dynamic ranking'' in the first place ...)
1102 It is invoked by adding
1103 the &acro.bib1; relation attribute with
1104 value ``relevance'' to the &acro.pqf; query (that is,
1105 <literal>@attr 2=102</literal>, see also
1106 <ulink url="&url.z39.50;bib1.html">
1107 The &acro.bib1; Attribute Set Semantics</ulink>, also in
1108 <ulink url="&url.z39.50.attset.bib1;">HTML</ulink>).
1109 To find all articles with the word <literal>Eoraptor</literal> in
1110 the title, and present them relevance ranked, issue the &acro.pqf; query:
1112 @attr 2=102 @attr 1=4 Eoraptor
1116 <sect3 id="administration-ranking-dynamic-rank1">
1117 <title>Dynamically ranking using &acro.pqf; queries with the 'rank-1'
1121 The default <literal>rank-1</literal> ranking module implements a
1122 TF/IDF (Term Frequecy over Inverse Document Frequency) like
1123 algorithm. In contrast to the usual defintion of TF/IDF
1124 algorithms, which only considers searching in one full-text
1125 index, this one works on multiple indexes at the same time.
1127 &zebra; does boolean queries and searches in specific addressed
1128 indexes (there are inverted indexes pointing from terms in the
1129 dictionary to documents and term positions inside documents).
1133 <term>Query Components</term>
1136 First, the boolean query is dismantled into its principal components,
1137 i.e. atomic queries where one term is looked up in one index.
1138 For example, the query
1140 @attr 2=102 @and @attr 1=1010 Utah @attr 1=1018 Springer
1142 is a boolean AND between the atomic parts
1144 @attr 2=102 @attr 1=1010 Utah
1148 @attr 2=102 @attr 1=1018 Springer
1150 which gets processed each for itself.
1156 <term>Atomic hit lists</term>
1159 Second, for each atomic query, the hit list of documents is
1163 In this example, two hit lists for each index
1164 <literal>@attr 1=1010</literal> and
1165 <literal>@attr 1=1018</literal> are computed.
1171 <term>Atomic scores</term>
1174 Third, each document in the hit list is assigned a score (_if_ ranking
1175 is enabled and requested in the query) using a TF/IDF scheme.
1178 In this example, both atomic parts of the query assign the magic
1179 <literal>@attr 2=102</literal> relevance attribute, and are
1180 to be used in the relevance ranking functions.
1183 It is possible to apply dynamic ranking on only parts of the
1186 @and @attr 2=102 @attr 1=1010 Utah @attr 1=1018 Springer
1188 searches for all documents which have the term 'Utah' on the
1189 body of text, and which have the term 'Springer' in the publisher
1190 field, and sort them in the order of the relevance ranking made on
1191 the body-of-text index only.
1197 <term>Hit list merging</term>
1200 Fourth, the atomic hit lists are merged according to the boolean
1201 conditions to a final hit list of documents to be returned.
1204 This step is always performed, independently of the fact that
1205 dynamic ranking is enabled or not.
1211 <term>Document score computation</term>
1214 Fifth, the total score of a document is computed as a linear
1215 combination of the atomic scores of the atomic hit lists
1218 Ranking weights may be used to pass a value to a ranking
1219 algorithm, using the non-standard &acro.bib1; attribute type 9.
1220 This allows one branch of a query to use one value while
1221 another branch uses a different one. For example, we can search
1222 for <literal>utah</literal> in the
1223 <literal>@attr 1=4</literal> index with weight 30, as
1224 well as in the <literal>@attr 1=1010</literal> index with weight 20:
1226 @attr 2=102 @or @attr 9=30 @attr 1=4 utah @attr 9=20 @attr 1=1010 city
1230 The default weight is
1231 sqrt(1000) ~ 34 , as the &acro.z3950; standard prescribes that the top score
1232 is 1000 and the bottom score is 0, encoded in integers.
1236 The ranking-weight feature is experimental. It may change in future
1244 <term>Re-sorting of hit list</term>
1247 Finally, the final hit list is re-ordered according to scores.
1255 Still need to describe the exact TF/IDF formula. Here's the info, need -->
1256 <!--to extract it in human readable form .. MC
1258 static int calc (void *set_handle, zint sysno, zint staticrank,
1261 int i, lo, divisor, score = 0;
1262 struct rank_set_info *si = (struct rank_set_info *) set_handle;
1264 if (!si->no_rank_entries)
1265 return -1; /* ranking not enabled for any terms */
1267 for (i = 0; i < si->no_entries; i++)
1269 yaz_log(log_level, "calc: i=%d rank_flag=%d lo=%d",
1270 i, si->entries[i].rank_flag, si->entries[i].local_occur);
1271 if (si->entries[i].rank_flag && (lo = si->entries[i].local_occur))
1272 score += (8+log2_int (lo)) * si->entries[i].global_inv *
1273 si->entries[i].rank_weight;
1275 divisor = si->no_rank_entries * (8+log2_int (si->last_pos/si->no_entries));
1276 score = score / divisor;
1277 yaz_log(log_level, "calc sysno=" ZINT_FORMAT " score=%d", sysno, score);
1280 /* reset the counts for the next term */
1281 for (i = 0; i < si->no_entries; i++)
1282 si->entries[i].local_occur = 0;
1287 where lo = si->entries[i].local_occur is the local documents term-within-index frequency, si->entries[i].global_inv represents the IDF part (computed in static void *begin()), and
1288 si->entries[i].rank_weight is the weight assigner per index (default 34, or set in the @attr 9=xyz magic)
1290 Finally, the IDF part is computed as:
1292 static void *begin (struct zebra_register *reg,
1293 void *class_handle, RSET rset, NMEM nmem,
1294 TERMID *terms, int numterms)
1296 struct rank_set_info *si =
1297 (struct rank_set_info *) nmem_malloc (nmem,sizeof(*si));
1300 yaz_log(log_level, "rank-1 begin");
1301 si->no_entries = numterms;
1302 si->no_rank_entries = 0;
1304 si->entries = (struct rank_term_info *)
1305 nmem_malloc (si->nmem, sizeof(*si->entries)*numterms);
1306 for (i = 0; i < numterms; i++)
1308 zint g = rset_count(terms[i]->rset);
1309 yaz_log(log_level, "i=%d flags=%s '%s'", i,
1310 terms[i]->flags, terms[i]->name );
1311 if (!strncmp (terms[i]->flags, "rank,", 5))
1313 const char *cp = strstr(terms[i]->flags+4, ",w=");
1314 si->entries[i].rank_flag = 1;
1316 si->entries[i].rank_weight = atoi (cp+3);
1318 si->entries[i].rank_weight = 34; /* sqrroot of 1000 */
1319 yaz_log(log_level, " i=%d weight=%d g="ZINT_FORMAT, i,
1320 si->entries[i].rank_weight, g);
1321 (si->no_rank_entries)++;
1324 si->entries[i].rank_flag = 0;
1325 si->entries[i].local_occur = 0; /* FIXME */
1326 si->entries[i].global_occur = g;
1327 si->entries[i].global_inv = 32 - log2_int (g);
1328 yaz_log(log_level, " global_inv = %d g = " ZINT_FORMAT,
1329 (int) (32-log2_int (g)), g);
1330 si->entries[i].term = terms[i];
1331 si->entries[i].term_index=i;
1332 terms[i]->rankpriv = &(si->entries[i]);
1338 where g = rset_count(terms[i]->rset) is the count of all documents in this specific index hit list, and the IDF part then is
1340 si->entries[i].global_inv = 32 - log2_int (g);
1347 The <literal>rank-1</literal> algorithm
1348 does not use the static rank
1349 information in the list keys, and will produce the same ordering
1350 with or without static ranking enabled.
1355 <sect3 id="administration-ranking-dynamic-rank1">
1356 <title>Dynamically ranking &acro.pqf; queries with the 'rank-static'
1359 The dummy <literal>rank-static</literal> reranking/scoring
1360 function returns just
1361 <literal>score = max int - staticrank</literal>
1362 in order to preserve the static ordering of hit sets that would
1363 have been produced had it not been invoked.
1364 Obviously, to combine static and dynamic ranking usefully,
1366 to make a new ranking
1367 function; this is left
1368 as an exercise for the reader.
1375 <literal>Dynamic ranking</literal> is not compatible
1376 with <literal>estimated hit sizes</literal>, as all documents in
1377 a hit set must be accessed to compute the correct placing in a
1378 ranking sorted list. Therefore the use attribute setting
1379 <literal>@attr 2=102</literal> clashes with
1380 <literal>@attr 9=integer</literal>.
1385 we might want to add ranking like this:
1387 Simple BM25 Extension to Multiple Weighted Fields
1388 Stephen Robertson, Hugo Zaragoza and Michael Taylor
1392 mitaylor2microsoft.com
1397 <sect3 id="administration-ranking-dynamic-cql">
1398 <title>Dynamically ranking &acro.cql; queries</title>
1400 Dynamic ranking can be enabled during sever side &acro.cql;
1401 query expansion by adding <literal>@attr 2=102</literal>
1402 chunks to the &acro.cql; config file. For example
1404 relationModifier.relevant = 2=102
1406 invokes dynamic ranking each time a &acro.cql; query of the form
1409 Z> f alvis.text =/relevant house
1411 is issued. Dynamic ranking can also be automatically used on
1412 specific &acro.cql; indexes by (for example) setting
1414 index.alvis.text = 1=text 2=102
1416 which then invokes dynamic ranking each time a &acro.cql; query of the form
1419 Z> f alvis.text = house
1429 <sect2 id="administration-ranking-sorting">
1430 <title>Sorting</title>
1432 &zebra; sorts efficiently using special sorting indexes
1433 (type=<literal>s</literal>; so each sortable index must be known
1434 at indexing time, specified in the configuration of record
1435 indexing. For example, to enable sorting according to the &acro.bib1;
1436 <literal>Date/time-added-to-db</literal> field, one could add the line
1438 xelm /*/@created Date/time-added-to-db:s
1440 to any <literal>.abs</literal> record-indexing configuration file.
1441 Similarly, one could add an indexing element of the form
1443 <z:index name="date-modified" type="s">
1444 <xsl:value-of select="some/xpath"/>
1447 to any <literal>alvis</literal>-filter indexing stylesheet.
1450 Indexing can be specified at searching time using a query term
1451 carrying the non-standard
1452 &acro.bib1; attribute-type <literal>7</literal>. This removes the
1453 need to send a &acro.z3950; <literal>Sort Request</literal>
1454 separately, and can dramatically improve latency when the client
1455 and server are on separate networks.
1456 The sorting part of the query is separate from the rest of the
1457 query - the actual search specification - and must be combined
1461 A sorting subquery needs two attributes: an index (such as a
1462 &acro.bib1; type-1 attribute) specifying which index to sort on, and a
1463 type-7 attribute whose value is be <literal>1</literal> for
1464 ascending sorting, or <literal>2</literal> for descending. The
1465 term associated with the sorting attribute is the priority of
1466 the sort key, where <literal>0</literal> specifies the primary
1467 sort key, <literal>1</literal> the secondary sort key, and so
1470 <para>For example, a search for water, sort by title (ascending),
1471 is expressed by the &acro.pqf; query
1473 @or @attr 1=1016 water @attr 7=1 @attr 1=4 0
1475 whereas a search for water, sort by title ascending,
1476 then date descending would be
1478 @or @or @attr 1=1016 water @attr 7=1 @attr 1=4 0 @attr 7=2 @attr 1=30 1
1482 Notice the fundamental differences between <literal>dynamic
1483 ranking</literal> and <literal>sorting</literal>: there can be
1484 only one ranking function defined and configured; but multiple
1485 sorting indexes can be specified dynamically at search
1486 time. Ranking does not need to use specific indexes, so
1487 dynamic ranking can be enabled and disabled without
1488 re-indexing; whereas, sorting indexes need to be
1489 defined before indexing.
1497 <sect1 id="administration-extended-services">
1498 <title>Extended Services: Remote Insert, Update and Delete</title>
1502 Extended services are only supported when accessing the &zebra;
1503 server using the <ulink url="&url.z39.50;">&acro.z3950;</ulink>
1504 protocol. The <ulink url="&url.sru;">&acro.sru;</ulink> protocol does
1505 not support extended services.
1510 The extended services are not enabled by default in zebra - due to the
1511 fact that they modify the system. &zebra; can be configured
1513 search, and to allow only updates for a particular admin user
1514 in the main zebra configuration file <filename>zebra.cfg</filename>.
1515 For user <literal>admin</literal>, you could use:
1519 passwd: passwordfile
1521 And in the password file
1522 <filename>passwordfile</filename>, you have to specify users and
1523 encrypted passwords as colon separated strings.
1524 Use a tool like <filename>htpasswd</filename>
1525 to maintain the encrypted passwords.
1529 It is essential to configure &zebra; to store records internally,
1531 modifications and deletion of records:
1536 The general record type should be set to any record filter which
1537 is able to parse &acro.xml; records, you may use any of the two
1538 declarations (but not both simultaneously!)
1540 recordType: dom.filter_dom_conf.xml
1541 # recordType: grs.xml
1543 Notice the difference to the specific instructions
1545 recordType.xml: dom.filter_dom_conf.xml
1546 # recordType.xml: grs.xml
1548 which only work when indexing XML files from the filesystem using
1549 the <literal>*.xml</literal> naming convention.
1552 To enable transaction safe shadow indexing,
1553 which is extra important for this kind of operation, set
1555 shadow: directoryname: size (e.g. 1000M)
1557 See <xref linkend="zebra-cfg"/> for additional information on
1558 these configuration options.
1562 It is not possible to carry information about record types or
1563 similar to &zebra; when using extended services, due to
1564 limitations of the <ulink url="&url.z39.50;">&acro.z3950;</ulink>
1565 protocol. Therefore, indexing filters can not be chosen on a
1566 per-record basis. One and only one general &acro.xml; indexing filter
1568 <!-- but because it is represented as an OID, we would need some
1569 form of proprietary mapping scheme between record type strings and
1572 However, as a minimum, it would be extremely useful to enable
1573 people to use &acro.marc21;, assuming grs.marcxml.marc21 as a record
1580 <sect2 id="administration-extended-services-z3950">
1581 <title>Extended services in the &acro.z3950; protocol</title>
1584 The <ulink url="&url.z39.50;">&acro.z3950;</ulink> standard allows
1585 servers to accept special binary <emphasis>extended services</emphasis>
1586 protocol packages, which may be used to insert, update and delete
1587 records into servers. These carry control and update
1588 information to the servers, which are encoded in seven package fields:
1591 <table id="administration-extended-services-z3950-table" frame="top">
1592 <title>Extended services &acro.z3950; Package Fields</title>
1596 <entry>Parameter</entry>
1597 <entry>Value</entry>
1598 <entry>Notes</entry>
1603 <entry><literal>type</literal></entry>
1604 <entry><literal>'update'</literal></entry>
1605 <entry>Must be set to trigger extended services</entry>
1608 <entry><literal>action</literal></entry>
1609 <entry><literal>string</literal></entry>
1611 Extended service action type with
1612 one of four possible values: <literal>recordInsert</literal>,
1613 <literal>recordReplace</literal>,
1614 <literal>recordDelete</literal>,
1615 and <literal>specialUpdate</literal>
1619 <entry><literal>record</literal></entry>
1620 <entry><literal>&acro.xml; string</literal></entry>
1621 <entry>An &acro.xml; formatted string containing the record</entry>
1624 <entry><literal>syntax</literal></entry>
1625 <entry><literal>'xml'</literal></entry>
1626 <entry>XML/SUTRS/MARC. GRS-1 not supported.
1627 The default filter (record type) as given by recordType in
1628 zebra.cfg is used to parse the record.</entry>
1631 <entry><literal>recordIdOpaque</literal></entry>
1632 <entry><literal>string</literal></entry>
1634 Optional client-supplied, opaque record
1635 identifier used under insert operations.
1639 <entry><literal>recordIdNumber </literal></entry>
1640 <entry><literal>positive number</literal></entry>
1641 <entry>&zebra;'s internal system number,
1642 not allowed for <literal>recordInsert</literal> or
1643 <literal>specialUpdate</literal> actions which result in fresh
1648 <entry><literal>databaseName</literal></entry>
1649 <entry><literal>database identifier</literal></entry>
1651 The name of the database to which the extended services should be
1661 The <literal>action</literal> parameter can be any of
1662 <literal>recordInsert</literal> (will fail if the record already exists),
1663 <literal>recordReplace</literal> (will fail if the record does not exist),
1664 <literal>recordDelete</literal> (will fail if the record does not
1666 <literal>specialUpdate</literal> (will insert or update the record
1667 as needed, record deletion is not possible).
1671 During all actions, the
1672 usual rules for internal record ID generation apply, unless an
1673 optional <literal>recordIdNumber</literal> &zebra; internal ID or a
1674 <literal>recordIdOpaque</literal> string identifier is assigned.
1675 The default ID generation is
1676 configured using the <literal>recordId:</literal> from
1677 <filename>zebra.cfg</filename>.
1678 See <xref linkend="zebra-cfg"/>.
1682 Setting of the <literal>recordIdNumber</literal> parameter,
1683 which must be an existing &zebra; internal system ID number, is not
1684 allowed during any <literal>recordInsert</literal> or
1685 <literal>specialUpdate</literal> action resulting in fresh record
1690 When retrieving existing
1691 records indexed with &acro.grs1; indexing filters, the &zebra; internal
1692 ID number is returned in the field
1693 <literal>/*/id:idzebra/localnumber</literal> in the namespace
1694 <literal>xmlns:id="http://www.indexdata.dk/zebra/"</literal>,
1695 where it can be picked up for later record updates or deletes.
1699 A new element set for retrieval of internal record
1700 data has been added, which can be used to access minimal records
1701 containing only the <literal>recordIdNumber</literal> &zebra;
1702 internal ID, or the <literal>recordIdOpaque</literal> string
1703 identifier. This works for any indexing filter used.
1704 See <xref linkend="special-retrieval"/>.
1708 The <literal>recordIdOpaque</literal> string parameter
1709 is an client-supplied, opaque record
1710 identifier, which may be used under
1711 insert, update and delete operations. The
1712 client software is responsible for assigning these to
1713 records. This identifier will
1714 replace zebra's own automagic identifier generation with a unique
1715 mapping from <literal>recordIdOpaque</literal> to the
1716 &zebra; internal <literal>recordIdNumber</literal>.
1717 <emphasis>The opaque <literal>recordIdOpaque</literal> string
1719 are not visible in retrieval records, nor are
1720 searchable, so the value of this parameter is
1721 questionable. It serves mostly as a convenient mapping from
1722 application domain string identifiers to &zebra; internal ID's.
1728 <sect2 id="administration-extended-services-yaz-client">
1729 <title>Extended services from yaz-client</title>
1732 We can now start a yaz-client admin session and create a database:
1735 $ yaz-client localhost:9999 -u admin/secret
1739 Now the <literal>Default</literal> database was created,
1740 we can insert an &acro.xml; file (esdd0006.grs
1741 from example/gils/records) and index it:
1744 Z> update insert id1234 esdd0006.grs
1747 The 3rd parameter - <literal>id1234</literal> here -
1748 is the <literal>recordIdOpaque</literal> package field.
1751 Actually, we should have a way to specify "no opaque record id" for
1752 yaz-client's update command.. We'll fix that.
1755 The newly inserted record can be searched as usual:
1760 Received SearchResponse.
1761 Search was a success.
1762 Number of hits: 1, setno 1
1763 SearchResult-1: term=utah cnt=1
1770 Let's delete the beast, using the same
1771 <literal>recordIdOpaque</literal> string parameter:
1774 Z> update delete id1234
1775 No last record (update ignored)
1776 Z> update delete 1 esdd0006.grs
1777 Got extended services response
1782 Received SearchResponse.
1783 Search was a success.
1784 Number of hits: 0, setno 2
1785 SearchResult-1: term=utah cnt=0
1792 If shadow register is enabled in your
1793 <filename>zebra.cfg</filename>,
1794 you must run the adm-commit command
1800 after each update session in order write your changes from the
1801 shadow to the life register space.
1806 <sect2 id="administration-extended-services-yaz-php">
1807 <title>Extended services from yaz-php</title>
1810 Extended services are also available from the &yaz; &acro.php; client layer. An
1811 example of an &yaz;-&acro.php; extended service transaction is given here:
1814 $record = '<record><title>A fine specimen of a record</title></record>';
1816 $options = array('action' => 'recordInsert',
1818 'record' => $record,
1819 'databaseName' => 'mydatabase'
1822 yaz_es($yaz, 'update', $options);
1823 yaz_es($yaz, 'commit', array());
1826 if ($error = yaz_error($yaz))
1833 <sect2 id="administration-extended-services-debugging">
1834 <title>Extended services debugging guide</title>
1836 When debugging ES over PHP we recomment the following order of tests:
1842 Make sure you have a nice record on your filesystem, which you can
1843 index from the filesystem by use of the zebraidx command.
1844 Do it exactly as you planned, using one of the GRS-1 filters,
1845 or the DOMXML filter.
1846 When this works, proceed.
1851 Check that your server setup is OK before you even coded one single
1853 Take the same record form the file system, and send as ES via
1854 <literal>yaz-client</literal> like described in
1855 <xref linkend="administration-extended-services-yaz-client"/>,
1857 remeber the <literal>-a</literal> option which tells you what
1858 goes over the wire! Notice also the section on permissions:
1863 in <literal>zebra.cfg</literal> to make sure you do not run into
1864 permission problems (but never expose such an unsecure setup on the
1865 internet!!!). Then, make sure to set the general
1866 <literal>recordType</literal> instruction, pointing correctly
1867 to the GRS-1 filters,
1868 or the DOMXML filters.
1873 If you insist on using the <literal>sysno</literal> in the
1874 <literal>recordIdNumber</literal> setting,
1875 please make sure you do only updates and deletes. Zebra's internal
1876 system number is not allowed for
1877 <literal>recordInsert</literal> or
1878 <literal>specialUpdate</literal> actions
1879 which result in fresh record inserts.
1884 If <literal>shadow register</literal> is enabled in your
1885 <literal>zebra.cfg</literal>, you must remember running the
1894 If this works, then proceed to do the same thing in your PHP script.
1906 <!-- Keep this comment at the end of the file
1911 sgml-minimize-attributes:nil
1912 sgml-always-quote-attributes:t
1915 sgml-parent-document: "zebra.xml"
1916 sgml-local-catalogs: nil
1917 sgml-namecase-general:t