[
https://issues.apache.org/jira/browse/DIRSERVER-2299?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17032286#comment-17032286
]
Emmanuel Lécharny commented on DIRSERVER-2299:
----------------------------------------------
Yes, we process every configured indexes.
The way teh repair process works is the following :
* we delete all the indexes, ie all the files that ends with *.db* that are
not the master table, based on the configuration. We don't delete *.lg* files
though, but they may be deleted when we recreate the index (to be double
checked, though)
* we then process with the master table, where we read every entry
** for each entry, we read its *RDN* and add it to the new *ParentidAndRdn*
index
** we update the new *ObjectClass, alias, entrcyCSN, administrativeRole* and
*presence* indexes
** we then iterate on all the attributes, and for each attribute being
indexed, we update the associated newly created index
At this point, I need to double check the implementation to see if it really
does what it is supposed to do...
> Missing Indexes after Restart
> -----------------------------
>
> Key: DIRSERVER-2299
> URL: https://issues.apache.org/jira/browse/DIRSERVER-2299
> Project: Directory ApacheDS
> Issue Type: Bug
> Components: index, ldap
> Affects Versions: 2.0.0.AM25
> Reporter: Rashid Mahmood
> Priority: Major
> Attachments: after-repair.png, image-2020-02-06-16-53-48-839.png,
> index-config.png
>
>
> We have decided to use ApacheDS (version 2.0.0.AM25) for Authentication in
> our product. Ldap Server and Directory Service are being used in Embedded
> Mode.
>
> During our Tests we found that after Restart, some of the Entries via Indexed
> Attribute Search Filter are not comming in search result. After considering
> various possibilities we found that it is Indexing problem. We used Repair
> commnd and then we see those entries in search results.
>
> Why some of the indexes were missing even it was a gracefull shutdown?
>
> We want to avoid repair command after every restart unless it is really
> necessary. Our product will serve in future more than 50K users, we have fear
> that it would cost us a lot if we encounter those unexpected problems.
>
> These are two attributes used in our search filters
> dn: ads-indexAttributeId=mail,ou=indexes,ads-partitionId=tsi,ou=part
> itions,ads-directoryServiceId=default,ou=config
> ads-indexattributeid: mail
> ads-indexHasReverse: FALSE
> ads-indexcachesize: 1000
> objectclass: ads-index
> objectclass: ads-jdbmIndex
> objectclass: ads-base
> objectclass: top
> ads-enabled: TRUE
> dn: ads-indexAttributeId=telID,ou=indexes,ads-partitionId=tsi,ou=part
> itions,ads-directoryServiceId=default,ou=config
> ads-indexattributeid: telID
> ads-indexHasReverse: FALSE
> ads-indexcachesize: 1000
> objectclass: ads-index
> objectclass: ads-jdbmIndex
> objectclass: ads-base
> objectclass: top
> ads-enabled: TRUE
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
