In migrating to Hazelcast 2.4.x, replace deprecated methods. Use Hazelcast.newHazelcastInstance() rather than Hazelcast.init(). For other deprecated static methods, use the HazelcastInstance equivalent calls.
In CNodeService.updateReplicationMetadata(), we are setting the replicaVerifiedDate() when we update or wholesale add a new replica. However, in setReplicationStatus(), we only do so when there's a new entry. Change setReplicationStatus() to also update the replicaVerifiedDate on updates of existing entries to be more consistent with other changes. This affects node prioritization based on this date timestamp. Thanks to Skye for pointing this out.
To attempt to address performance and stability WRT Hazelcast communication, we're upgrading to the 2.x series of Hazelcast. remove the 1.9.x jar files, and add the 2.4.1-SNAPSHOT jars. Modify HazelcastService to handle the minor change in the ItemListener interface (now passes ItemEvent<Identifier> as an argument)....
implement query description for pathquery -- only tells callers about the pre-indexed paths we have in Metacat since there are an infinite number of "fields" when storing arbitrary XML, but we really don't want people using non-indexed paths for performance reasons anyway. I've typed all the fields as String, even though some are not just strings and can be used for numeric or data comparisons.
Implement MNQuery for "pathquery" engine. Optionally include guid in the pathquery results (https://redmine.dataone.org/issues/3083)
use ObjectFormatInfo libclient utility to look up mimeType and filename extension during get() calls. Configurable mapping file is deployed by default to /var/metacat/dataone where it can then be augmented as needed. This location is controlled in the metacat.properties file (which is injected into the DataONE Settings values during weapp intitialization)....
add count for the total processed pids (from ISet iterator)
handle /object?count=0 queries using simpler (quicker) sql https://redmine.dataone.org/issues/3065
allow getlog action to use docid parameters that do not include revision. In these cases, the latest revision will be used.
handle case where we do not have a pathexpr to checkhttp://bugzilla.ecoinformatics.org/show_bug.cgi?id=5696
simplify the xml_access query, and instead use guid to check for permission. Now the docid/rev join (to get most recent version for search results) happens "higher up" in the query.http://bugzilla.ecoinformatics.org/show_bug.cgi?id=5696
pass parameters to the getLog action for rendering in xslt
remove morpho.jar -- moved needed classes into shared utilities project. (currently building form utilities trunk -- be sure to 'ant fullclean' to get the latest utilities.jar built)
Update d1_common_java and d1_libclient_java to the newest jar files. Add methods to CNodeService to throw NotImplemented exceptions for query(), listQueryEngines(), and getQueryEngineDescription() since these API calls are handled outside of metacat.
do not allow updates to orphan another branch of revision history. https://redmine.dataone.org/issues/3338
Change the set and get methods for the replication verified date to use java.sql.Timestamp rather than java.util.Date via setTimestamp(), not setDate(). The hh:mm:ss.sss was previously getting truncated.
include the subjects we are testing for authentication.https://redmine.dataone.org/issues/2778
remove the max(rev) clause in favor of a more straight-forward join to xml_documents (that will have the max rev). http://bugzilla.ecoinformatics.org/show_bug.cgi?id=5696
include inverted sendParameters() method that uses the keys as values, and the values as keys so that multiple docid parameters can be specified for the zip download. This was a regression when moving to standard httpclient rather than the roll-your-own version we had been using. http://bugzilla.ecoinformatics.org/show_bug.cgi?id=5718
shorten the systemmetadata* table names for Oracle's 30 character limit. move version to 2.0.5. http://bugzilla.ecoinformatics.org/show_bug.cgi?id=5717
use correct docid format when checking for existing mappings.
use CDATA for docname field in docInfo so that XML parser ignores the content that can contain characters like "&
use SchemaLocationResolver to fetch remote entries for the xml_catalog -- we want to be able to fetch included xsd files as well as use any error handling it provides for checking the schemas.
when performing query, make sure we are using the access rules of the latest revision of a given docid, otherwise we may include documents that used to be public but have been made private in subsequent revisions.http://bugzilla.ecoinformatics.org/show_bug.cgi?id=5696
correct the number of prepared statement parameters when inserting to xml_revisions table. Errors like the following were showing in the replication log file:knb 20120831-19:42:38: [ERROR]: DocumentImpl.writeReplication - Failed to create access rule for package: john.15950.1 because The column index is out of range: 12, number of columns: 11. [ReplicationLogging]
include WHERE in the sql where clause - encountered by SAEON's node admin, Alex Niehaus.
create docid-guid mapping during replication if it does not exist. we were [incorrectly] assuming that there would be SM coming with the document info that would fill this information in, but for traditional non-MN Metacat deployments there is no SM to provide a mapping. In this case we use the docid as the guid.
stream the replication "update" response rather than building up a complete list in a stringbuffer. prompted by findings on t he CN: https://redmine.dataone.org/issues/3141
make sure data objects correctly use force replicate with action "insert" https://redmine.dataone.org/issues/3138
when updating a document on a remote server, we still need to use the previous docid to check that the user has permissions to do so (rather than the new id that is obsoleting the old id). This was discovered by M Servilla at LTER.
remove unused "dataonelogger"
allow SM resynch to be executed any time, not just during start up.https://redmine.dataone.org/issues/3116
change to debug log level when processing shared/local pids)
only lock the missing pid event if we know we have it locally to contribute.https://redmine.dataone.org/issues/3117
Add locking to the itemAdded() method so ideally only one CN will respond to the request for a 'wanted' pid from the cluster. The lock is on a string, not the pid, and so won't conflict with system metadata locking. The string is based on the pid, with "missing-" as a prefix.
only publish to the missing pid "wanted list" when resynching system metadata. we were seeing redundant entry added/updated events when looking up the shared systemmetadata first.
print the missing pid count, not the total shared pid count so we know how many will be processed.
change the system metadata resynch approach: nodes will publish PIDs that they are missing after inspecting the shared identifier set. other nodes will be listening for the "wanted" pids and will put their local copy of SystemMetadata on the shared SM map. This should dramatically decrease the hazelcast chatter during a resynch and targets only the pids that are missing from any of the various nodes.
logging for processing identifier set on restart.
remove possibility for infinite loop in case data replication is not configured for the server and a data file is encountered (yikes!)
added logging debug statements to see where the replication timeout might be occurring.
check for null archived flag in ORE SMhttps://redmine.dataone.org/issues/3046
check if the caller is the Node admin (the member node calling itself) as well as the existing check for the CN calling the service. Both of those callers should be given full admin rights.
use local Set processing to determine which pids (if any) should be contributed to the shared set by this node during the resync. Should save time rather than checking each and every pid against the shared set.
move the hzIdentifiers initialization into the resync thread so that it does not affect start up time. cleaned up unused methods and superfluous code.
only load local pids into hzIdentifiers if t hey do not already exist in the shared set. increase logging severity and detail of messages emitted during this process to get a better sense of what is taking so long.
utility methods to update/reserialize existing ORE maps that were generated with older foresite (and included bad dateTime strings).https://redmine.dataone.org/issues/3046
On the coordinating Nodes, we often get McdbDocNotFoundExceptions for data (doctype == 'BIN') documents because they are not synchronized to the CNs. Change the logging to only print the stack trace during load() and loadAll() when log debug is enabled.
check for invalid (!) pids. thanks, M. Reyes for catching thishttps://redmine.dataone.org/issues/3047
only look up the client timeout property once, not every time we make a callhttps://redmine.dataone.org/issues/3078
improve content type handling during the get() callshttps://redmine.dataone.org/issues/3070
check for whitespace in identifiers during create() and update()https://redmine.dataone.org/issues/3047
configurable replication client timeouthttps://redmine.dataone.org/issues/3078
order the listObjects() results by identifier to mitigate random paged resultshttps://redmine.dataone.org/issues/3065
correct the parameter/value setting in the prepared statements for retrieving log information.
use docid, not the guid when returning the accesscontrol block
include dataone.ore.downloaddata as a configurable property in case MNs (like LTER) want to have the process download externally-stored data files described in an EML data package.
set date SM modified when we are setting obsoletes/obsoletedBy/archived values. This way the CN can actualy pick up the changes in revision history.
log error when looking up non-existent local SM rather than completely bombing out of the resynch thread.
look up docid using mapped guid when checking permission on described data fileAddresses: http://support.nceas.ucsb.edu/rt/Ticket/Display.html?id=7490
use docid (not guid) when instantiating the PermissionController. Was getting an error with DOI-ified identifier and the metacat getaccesscontrol action:https://knb.ecoinformatics.org/knb/metacat?action=getaccesscontrol&docid=Collinge.3.28<error>AccessControlForSingleFile.getACL() - MCDB error when getting ACL: No guid registered for docid doi:10.5063/AA/Collinge.3.28...
make sure we have non-null values where jibx serialization expects them for LogEntry
use secure Metacat context URL for D1 registrationhttps://redmine.dataone.org/issues/3030
first pass: DataONE-specific log retrieval to avoid java-based post-processing.
set archived flag (true) when we set the obsoletedBy value in the ORE system metadata
use the localId for obsoletes/obsoletedBy ORE system metadata (https://redmine.dataone.org/issues/2964)
Print the stack trace when the MMP cannot be resolved.
report errors during XML->HTML transformhttp://bugzilla.ecoinformatics.org/show_bug.cgi?id=5618
Oops, previous commit suffered from a happy trigger finger. During deleteReplicationMetadata(), don't delete the replica on the replica Member Node. Call CN.delete() for that functionality. This call just updates sytem metadata (according to the API description).
Minor logging change.
Add debug logging to delete() to understand why we're getting InsufficientKarmaException.
Since we already have determined access via isAuthorized() and isAdminAuthorized(), act as the Metacat administrator during calls to DocumentImpl.delete() in archive(), passing in null username and group.
restrict getLogRecrods (both MN and CN) to be called only by admin users (the CN)https://redmine.dataone.org/issues/2855
In setReplicationStatus() and UpdateReplicationMetadata(), don't allow a status state change from COMPLETED to anything other than INVALIDATED. This prevents the completed status from being overwritten due to race conditions.
use metacat.properties to specify the default checksum algorithm to use -- this way it will be easy for us to switch to whatever DataONE decrees. https://redmine.dataone.org/issues/2834
put(sm) for every pid we have a SM value for so that all members receive the entry event and can save locally.
Throw an exception when NOT allowed, not when allowed =).
ignore partition owner -- always attempt to look up form local store if we were unable to get the SM from the shared map.
do not check if this CN has a "perfect" copy of the SM identifiers -- we need any CN coming online to contribute the records that they have locally so that in the event that all three CNs have a partial view of things they all eventually share each others' SM entries.
Also get the list size, which may throw an NPE.
Only add an AccessPolicy to SystemMetadata during generation when the AccessPolicy is not empty. We've had some scenarios where IdentifierManager.getaccessPolicy() is returning an empty policy because of an empty permission list coming from the db. This was causing InvalidSystemMetadata exceptions during MN to MN replication.
push SystemMetadata entries from the CN that has them all to the shared map where other nodes may not have all entries. The CN with the complete copy only pushes SM entries that it does not own and that return as null because those are the ones that are missing on the other, non-complete CNs....
trace level log for looping over EVERY pid in the system.
meant to log the guids (source) not the pids (target)
logging for each step of shared identifiers loading.
remove pause/resume - seemed to make metacat just hang on SM retrieval. Add more logging when returned SM is null -- want to make sure it is becuase the local node "owns" the pid key even though there is no value for it.
due to hudson build issue, did not actually end up testing pause/resume -- trying that again
pause/resume was not enough. trying shutdown/restart
experiment with lifecycle pause/resume. hopefully it prevents our node from taking ownership of any keys before we are sure we have them all.
increase logging and add back in the call to saveLocally() in case the SM object has already been loaded into the shared map but before this node came back online.
no need to call saveLocally explicitly since loading from the shared store triggers that behavior locally because of the configured listeners.use an iterator over the shared identifiers in case this set is constantly changing.
make only one DB call to look up local pids - no need to do a pstmt for every single shared pid.
on init (start up) launch a synchronization thread that ensures all shared identifier entries have a corresponding local System Metadata entry.
fix NPE (logMetacat object was not initialized) that was occurring during store()
stack trace the HZ put exception during CN-CN replication
additional debugging statements for CONCURRENT_MAP_PUT error during CN-CN replication.
Don't set the replication status to failed for an object when it is called by a public user. Just throw the NotAuthorized exception. This prevents this node from being de-prioritized because of public calls to the method.
share the same dbConnection when inserting and then updating SystemMetadata objects in the backing store.any errors encountered during the update will rollback the entire transaction and the SM record will not exist, even in part.