Conifer 2.4 upgrade status
25 views
Skip to first unread message
Dan Scott
Jun 17, 2013, 3:25:10 PM6/17/13
to Conifer Discussion List
Hi all:
So here's a run-down of things we've run into, or have been reported, and their status / workarounds, as of today:
* Staff client upgrades automatically, but reports horrible errors on trying to connect. We suspect that this is likely to be an issue with older workstations that have staff clients that have been upgraded multiple times. Workaround: uninstall the staff client, and reinstall a fresh staff client from http://clients.concat.ca
* Serials holdings are no longer visible on the record details page. Status: The holdings are visible, but they are under a twisty near the bottom of the page called "Issues Held". Workaround: Click the twisty. Longer term, we'll want to make those immediately visible under the copy listing, as we did before.
* Holds were not being retargeted properly. Cause: a date comparison in the recall-processing code was unhappy with the format of the date that was being supplied. Status: resolved as of 13:20.
* Bibliographic imports failed this morning. Cause: A fix that I created last week to address bloated search indexes unfortunately contained an SQL error that wasn't evident until new records were imported. Status: resolved as of 9:20.
* The reports interface showed only partial objects and fields. Cause: The translatable reports IDL file was missing one or more entities; as a short-term workaround, I replaced it with the hard-coded English (US) reports IDL file around 9:20am. Staff clients that downloaded the corrupted IDL file likely kept a cached version of that corrupted IDL file rather than checking the server for a newer version; to date I have not been able to reproduce the problem that was reported earlier. Status: resolved as of 9:20.
* Some searches (such as searching for terms or phrases that contain apostrophes) may incorrectly return 0 or incomplete results. Cause: Search indexes are being regenerated. One of the updates in the 2.4.0 release notes mentioned improvements to searching that fixed many long-standing problems. The cost is that we have to reindex all of our records. Status: I've taken a very simplistic but robust approach to the reindexing, and it will probably take 10 days to run through. Simple is good right now, so I'm okay with taking the extra time. In the meantime, most searches continue to work fine.
* Searching the Library of Congress as a Z39.50 import target always returned 0 results. Cause: the Library of Congress Z39.50 server deprecated their old hostname in favour of a new hostname (lx2.loc.gov); the update pointed our Z39.50 target at the new hostname. Status: resolved. We needed to open up access to that hostname through the firewall on our servers at Guelph.
* (University of Sudbury): The 659 field for records used to be masked in the catalogue, but is now being displayed. Cause: the upgrade modified some of the templates (for example, to add graphic 880 support to the display) and the old method of masking a field needs to be updated. Workaround: none. I expect to be able to create a fix for this relatively quickly but was focusing on system-wide issues as a priority.
* Local settings for printers disappeared: One site reported that their settings for printers had disappeared. I unfortunately have not yet had time to investigate this problem, and am not clear whether this is a problem for printing receipts, hold slips, call number labels, in particular, or for all printers. I'm also not sure if this is a widespread problem.
* Local admin disappeared: One site reported that "Local admin permissions disappeared". Again, I haven't had time to investigate this report in depth, but I do know that the Admin menu in the top right hand corner should include options like "Local administration" and "Server administration". Some of the sub-menu items will be greyed out if you do not have individual permissions, for example, creating MARC facet/search classes is not something that we want local administrators to be able to alter. However, other actions might simply require us to add the corresponding permission to the "Local administrator" group (or "Circ administrator", or "Acquisitions administrator", etc). If you see something that you feel you should have access to as a local administrator, please let us know!
* Fines not being generated: There was a report that fines were not being generated after the upgrade, and I was given a specific example of one item for one user that was checked in (backdated) with no fines showing. However, I can see that fines are being generated in the system. Still need to investigate whether the "Amnesty" modifier was checked when the check-in occurred (in which case no fines would be expected), or determine if there is an issue with backdating checkins that cause fines to be wiped, or determine if there is some combination of patron & item type that causes no fines to be generated.
If you have reported something else, and I have failed to mention it on this list, I apologize; I'm a little tired. But please respond and add it to the list so we can continue to track the known remaining issues!
So here's a run-down of things we've run into, or have been reported, and their status / workarounds, as of today:
* Staff client upgrades automatically, but reports horrible errors on trying to connect. We suspect that this is likely to be an issue with older workstations that have staff clients that have been upgraded multiple times. Workaround: uninstall the staff client, and reinstall a fresh staff client from http://clients.concat.ca
* Serials holdings are no longer visible on the record details page. Status: The holdings are visible, but they are under a twisty near the bottom of the page called "Issues Held". Workaround: Click the twisty. Longer term, we'll want to make those immediately visible under the copy listing, as we did before.
* Holds were not being retargeted properly. Cause: a date comparison in the recall-processing code was unhappy with the format of the date that was being supplied. Status: resolved as of 13:20.
* Bibliographic imports failed this morning. Cause: A fix that I created last week to address bloated search indexes unfortunately contained an SQL error that wasn't evident until new records were imported. Status: resolved as of 9:20.
* The reports interface showed only partial objects and fields. Cause: The translatable reports IDL file was missing one or more entities; as a short-term workaround, I replaced it with the hard-coded English (US) reports IDL file around 9:20am. Staff clients that downloaded the corrupted IDL file likely kept a cached version of that corrupted IDL file rather than checking the server for a newer version; to date I have not been able to reproduce the problem that was reported earlier. Status: resolved as of 9:20.
* Some searches (such as searching for terms or phrases that contain apostrophes) may incorrectly return 0 or incomplete results. Cause: Search indexes are being regenerated. One of the updates in the 2.4.0 release notes mentioned improvements to searching that fixed many long-standing problems. The cost is that we have to reindex all of our records. Status: I've taken a very simplistic but robust approach to the reindexing, and it will probably take 10 days to run through. Simple is good right now, so I'm okay with taking the extra time. In the meantime, most searches continue to work fine.
* Searching the Library of Congress as a Z39.50 import target always returned 0 results. Cause: the Library of Congress Z39.50 server deprecated their old hostname in favour of a new hostname (lx2.loc.gov); the update pointed our Z39.50 target at the new hostname. Status: resolved. We needed to open up access to that hostname through the firewall on our servers at Guelph.
* (University of Sudbury): The 659 field for records used to be masked in the catalogue, but is now being displayed. Cause: the upgrade modified some of the templates (for example, to add graphic 880 support to the display) and the old method of masking a field needs to be updated. Workaround: none. I expect to be able to create a fix for this relatively quickly but was focusing on system-wide issues as a priority.
* Local settings for printers disappeared: One site reported that their settings for printers had disappeared. I unfortunately have not yet had time to investigate this problem, and am not clear whether this is a problem for printing receipts, hold slips, call number labels, in particular, or for all printers. I'm also not sure if this is a widespread problem.
* Local admin disappeared: One site reported that "Local admin permissions disappeared". Again, I haven't had time to investigate this report in depth, but I do know that the Admin menu in the top right hand corner should include options like "Local administration" and "Server administration". Some of the sub-menu items will be greyed out if you do not have individual permissions, for example, creating MARC facet/search classes is not something that we want local administrators to be able to alter. However, other actions might simply require us to add the corresponding permission to the "Local administrator" group (or "Circ administrator", or "Acquisitions administrator", etc). If you see something that you feel you should have access to as a local administrator, please let us know!
* Fines not being generated: There was a report that fines were not being generated after the upgrade, and I was given a specific example of one item for one user that was checked in (backdated) with no fines showing. However, I can see that fines are being generated in the system. Still need to investigate whether the "Amnesty" modifier was checked when the check-in occurred (in which case no fines would be expected), or determine if there is an issue with backdating checkins that cause fines to be wiped, or determine if there is some combination of patron & item type that causes no fines to be generated.
If you have reported something else, and I have failed to mention it on this list, I apologize; I'm a little tired. But please respond and add it to the list so we can continue to track the known remaining issues!
Reply all
Reply to author
Forward
0 new messages