Looks like something recent broke search in web staff client

[Jason Stephenson]

Working with Evergreen master this morning (2017-09-02) and catalog search results appear to be broken. The OPAC seems to work OK. I'm attaching screenshots with Concerto data for more or less the same search to illustrate.

[Jason Stephenson (jstephenson)]

I just did a fresh installation on a clean VM, and it works. I don't know what the problem was before, maybe some junk from a previous installation.

I did do the npm install and grunt all before make install.

[Jason Stephenson (jstephenson)]

I'm setting this back to "New" because I'm seeing it repeatedly again today both with Firefox and with Chromium. It seems any search returning more than 1 result in the web staff client is broken. I get either an Internal Server Error or a list of "blank" results.

The OPAC works, if I search it before using the web staff client in the same browser session. If I use the web staff client in the same browser session, then OPAC searches returning more than 1 result show the same behavior.

I have seen this with the master branch as well as one where I have been working on an enhancement for placing multiple holds at once.

I am also using OpenSRF master.

The database is PostgreSQL 9.5.9 from the PostgreSQL community apt repositories.

The O/S is Ubuntu 16.04.3 with the 4.10 HWE kernel. The vm was freshly built today. I'm going to clear the apache logs and then add them when this result shows to this bug.

This happens with such frequency with 3 rebuilds of the VM that I'm confident it is not me.

[Jason Stephenson (jstephenson)]

The above three logs were generated with the web staff client logged in from Chromium as the br1mtownsend circulator doing a keyword search for Brahms with the Concerto dataset loaded. I got an internal server error.

[Jason Stephenson (jstephenson)]

I'm seeing this mainly after having shut the vm down and restarted it. I have not confirmed the behavior in the staff client with a fresh-built vm. However, it is consistent if I stop the VM and restart it, being sure to shut services off beforehand and restarting them after starting the vm up again. I used to do this with some frequency and never experienced this behavior until September 2.

[Jason Stephenson (jstephenson)]

This might be bug 1709710. I applied the fix for that and the behavior seems to have stopped. I'll give this another day (tomorrow) and see what happens.