Metacat: Issueshttps://projects.ecoinformatics.org/ecoinfo/https://projects.ecoinformatics.org/ecoinfo/ecoinfo/favicon.ico?14691340362007-11-02T20:11:30ZEcoinformatics Redmine
Redmine Bug #2993 (Resolved): Permission Order is always set to 'allowFirst'https://projects.ecoinformatics.org/ecoinfo/issues/29932007-11-02T20:11:30ZChris Barteaubarteau@nceas.ucsb.edu
<p>Permission Order is always set to 'allowFirst'. Even if you send 'denyFirst, it will save 'allowFirst' to the access table.</p> Bug #2832 (Resolved): Online Data Table display is awkwardhttps://projects.ecoinformatics.org/ecoinfo/issues/28322007-04-24T23:03:33ZCallie Bowdishbowdish@nceas.ucsb.edu
<p>Data Table and attribute information when displayed online has problems when there are greater than 10 columns and when greater than 400 characters with spaces are in the attribute description. Object Name, Text Format and Online Distribution Info are all centered in relation to the amount of columns. This makes this data hard to see because you have to scroll to the middle area that is measured by the amount of columns.</p>
<p>Data package cmccreedy.3 is a good example of these problems with the online display.</p> Bug #2807 (Resolved): Data Set Upload for SAEONhttps://projects.ecoinformatics.org/ecoinfo/issues/28072007-03-28T17:44:01ZChris Barteaubarteau@nceas.ucsb.edu
<p>Create a web interface to upload data sets and metadata files for Kruger NP.<br />Includes login, and inserting new documents into metacat.</p> Bug #2806 (Resolved): Online Edit action for a data package brings up a blank pagehttps://projects.ecoinformatics.org/ecoinfo/issues/28062007-03-23T22:33:35ZCallie Bowdishbowdish@nceas.ucsb.edu
<p>A data package can not be edited using the Online NCEAS Data Repository (<a class="external" href="http://data.nceas.ucsb.edu">http://data.nceas.ucsb.edu</a> ). I tried data packages that I had rights to. I was logged into the KNB. When I clicked the edit button on the data packages found page it brought up a blank page. There is no notification of there being a problem. Currently once a data package is saved a user would have to use Morpho to edit it. The online edit feature does not work.</p> Bug #2557 (Resolved): Metacat Performace: Rewrite the xml_access part of the queryhttps://projects.ecoinformatics.org/ecoinfo/issues/25572006-09-28T16:11:35ZMatt Jonesjones@nceas.ucsb.edu
<p>The subpart of metacat search queries that does an access control check is inefficient. It first checks the user_owner field in xml_documents (of dubious value), and then it does two queries against xml_access to determine if the user has rights to query. The query as implemented in Metacat 1.6 is attached. Note in particular that the user_owner part of the query appears to be wrong -- at no point would the user_owner ever be 'public', so this part of the query can be omitted for public searches, avoiding a large subquery.</p> Bug #2175 (Resolved): Metacat Performace: Upgrading hardware setuphttps://projects.ecoinformatics.org/ecoinfo/issues/21752005-09-05T18:37:10ZSaurabh Gargsgarg@nceas.ucsb.edu
<p>Following things have to be done in regards to the upgrading of hardware setup</p>
<p>-> Have to install a new ESA metacat server with Postgres as the db.</p>
<p>-> Upgrade KNB metacat server to ecoinfo2 with Postgres on it.</p>
<p>-> check docs on ecoinfo == docs on ecoinfo2 (This includes the revisions and <br />the deleted documents. Bug# 2136)</p>
<p>(The above are from the ESA tasklist)</p>
<p>Notes from Matt's email regarding Metacat and ldap setup:</p>
<p>Basically, we want LDAP and metacat to be a failsafe system, which means <br />eliminating single points of failure. Our proposed solution is to basically <br />make both NCEAS and LTER house identical copies of all information and services <br />and act as failover points for each other. The setup is slightly different for <br />ldap and metacat so I'll talk about both of them separately.</p>
<p>LDAP<br />------<br />ldap1.ecoinformatics.org resides at NCEAS<br /> master for the NCEAS, UCNRS, and some other trees<br /> slave for LTER trees<br /> slave for KNP tree, PISCO tree, etc<br />ldap2.ecoinformatics.org resides at LTER<br /> master for LTER<br /> slave for other domains</p>
<p>Matt and James agreed to work with Sid, Duane, and Jing to get this replication <br />and referral system working.</p>
<p>ldap.ecoinformatics.org<br />Uses DNS round-robin to point at both of those servers and redirect clients to <br />one or the other in a load-balanced and failover way. For this to work we need <br />to make sure that DNS caching won't delay failover to the other host in event <br />of a net failure. Colby and Mark will experiment with various DNS <br />failover/load balanacing techniques and propose a solution to this group.</p>
<p>Metacat<br />---------<br />Basically, same setup as now, except we need to install identical set of skins <br />and default skin on both the NCEAS and LTER metacats. The NCEAS metacat will <br />be renamed 'knb1.ecoinformatics.org' and the LTER metacat will <br />become 'knb2.ecoinformatics.org'. Both of these will be used in a round robin <br />(as for ldap) based on the address knb.ecoinformatics.org which will be <br />modified to resolve to both the knb1 and knb2 hosts (and will need to be moved <br />in a way that doesn't cause downtime).</p>
<p>In normal operation, requests for access will be distributed between both <br />hosts. In the event of a failure, DNS will redirect hosts to the alternate <br />address that remains accessible. During a failure, documents with a home <br />server on the inaccessible host will not be editable (because a lock for the <br />document will be inaccessible). Otherwise, all other functions will be present.</p>
<p>This will mean that the NCEAS and LTER servers will be treated as a single <br />integrated system, and upgrades and changes will need to be synchronized <br />carefull between them. Also, access logs will be split between them because <br />requests will be regularly routed to both servers.</p> Bug #2157 (Resolved): Metacat Performance: Optimize Postgres and Tomcat...https://projects.ecoinformatics.org/ecoinfo/issues/21572005-07-14T01:24:42ZSaurabh Gargsgarg@nceas.ucsb.edu
<p>There might be some options in Tomcat to speed things up a little bit. e.g. <br />Chris suggested about turning off autoloading of war files.</p>
<p>In Postgres.. From Matt's email...<br />It would be best to fit the whole postgres<br />image into memory so it need not be swapped to disk for queries.<br />Postgres needs to be configured to use the memory even if its on the<br />machine. Chris saw significant performance gains by upping the memory<br />for Postgres to 768MB. We have a much larger DB, so we probably want to<br />go higher. Ecoinfo2 has 2GB of RAM, which we should probably increase<br />to 4GB. New machines should have that much at least, but we run into<br />the 32bit limit so might need to consider moving to a 64bit os. Sid<br />should look at the memory size needed for the postgres image on ecoinfo2<br />to get an idea of how much we need.</p> Bug #2153 (Resolved): Metacat Performace: Add/drop indices on key columnshttps://projects.ecoinformatics.org/ecoinfo/issues/21532005-07-14T01:14:49ZSaurabh Gargsgarg@nceas.ucsb.edu
<p>From Matt's email...</p>
<p>DB indices aren't created on some key columns such as<br />xml_nodes.nodedata (originally because indices weren't supported on<br />fields that big). This means a full table scan of the 6.5M records in<br />xml_nodes for every query. We need to inspect for missing indices and<br />create them, and need to consider the effect of <abbr title="">UPPER</abbr> on these indices<br />in case-insensitive queries and be sure to create corresponding <abbr title="">UPPER</abbr><br />indices too. Once the indices are in place we need to check EXPLAIN to<br />be sure the optimizer is using the indices instead of doing table scans<br />(see <a class="external" href="http://www.postgresql.org/docs/faqs.FAQ.html#4.7">http://www.postgresql.org/docs/faqs.FAQ.html#4.7</a>). Also, indices<br />slow down inserts and updates, so any that aren't in use should be<br />dropped, which will take some analysis of the running DB (but this is<br />low priority comparatively speaking).</p> Bug #1982 (Resolved): add function to submit data with registry entitieshttps://projects.ecoinformatics.org/ecoinfo/issues/19822005-02-18T01:37:34ZMatt Jonesjones@nceas.ucsb.edu
<p>A common request is to be able to submit data to the metacat along with a<br />registry entry as it is submitted. We've not done this because of the<br />additional metadata burden that is typically required to be EML compliant<br />(entity and attribute). However, there is some utility to being able to archive<br />the data even without the detailed entity and attribute metadata. Once strategy<br />is this:</p>
<p>1) Add a form upload field to allo a user to point at a data file on disk for upload<br />2) Upon submit, upload the metadata and data, parse out the data file and submit<br />it to metacat as its own object.<br />3) Use the ID for the submitted dataset to create a metacat URL, and write that<br />metacat url into the distribution section for the overall data package, and then<br />submit the metadata.</p>
<p>This is not the traditional approach to uploading EML data, as data are<br />typically uploaded in the dataTable/distribution section. A couple of problems<br />arise. First, because the distribution url is in an atypical place in EML,<br />metacat doesn't know how to set access control rules on the data object<br />properly, so this will need to be specially handled. Second, because the<br />distribution url is in an atypical place in EML, morpho won't know how to deal<br />with the data package and will need to be modified to detect the presence of<br />these data tables and 'move' the reference to the dataTable subtree to allow the<br />user to provide entity/attribute info for the table and to view the data. This<br />will be entered as a separate bug. So, an additional feature to contemplate is:</p>
<p>4) Add a form field for specifying access restrictions on the data. This could<br />be a simple radio button like:<br /> "Make data 1) publicly available 2) private"</p>
<p>In some special cases, such as the NCEAS data registry, we might want a third<br />choice like "Group available" where the group consists of the people that are<br />members of the associated working group. THis will require being able to get<br />the list of ldap accounts associated with a group from the registry script, but<br />this could be presumably added to the admindb access script in perl.</p> Bug #1232 (Resolved): cookie needs to be written after knb page loginhttps://projects.ecoinformatics.org/ecoinfo/issues/12322003-12-06T01:26:25ZMatthew Brookebrooke@nceas.ucsb.edu
<p>this actually applies to code in the KNBWEB cvs module, but the consequences<br />apply to metacat.</p>
<p>here's the scenario:<br />1) a user logs in on knb homepage (knb.ecoinformatics.org/index.jsp)<br />2) the search form then has a hidden sessionid field added to it by the jsp<br />3) user then does a search and gets some private datasets in results<br />4) what I haven't done is to add the sessionid to the results links<br /> - so when user clicks a dataset, metacat assumes user is "public" <br />5) when you get "into" a dataset, there are even more links that would need the<br />sessionid added</p>
<p>...so, the easiest solution would probably be just to write a cookie on login.<br />This code would probably go in <knbweb_cvs>/includes/include_session_vars.jsp,<br />where there is a code block for handling login (add code to create cookie here),<br />and another for handling logout (destroy cookie).</p>
<p>Sample code for writing cookies from java currently exists in the Metacat<br />module, in the HttpMessage class</p>