Connecting redis on a site knocks another site offline

I eventually figured this out. But I’m posting this in case it’s useful to others.
I probably got into this pickle because I’m not deeply familiar with redis yet.

After migrating a WordPress site from localdev to remote I installed the redis-cache plugin, activated it and connected the plugin to the redis server.

The site immediately whitescreened. Subsequent attempts to browse to the site yielded redirects to another site on the same host.

I’d seen this before when experimenting, and the solution is in this plugin threads sticky: I flushed the cache from the command line. After this the site works as normal.

However, the other site immediately went offline.

Flushing the cache brought the old site back online. But knocked out the newly installed site: it yielded whitescreens and redirects again.

After cycling between flushing and activating/deactivating the plugin it seemed that I could only have one site up at a time for more than a few seconds.
(In those few seconds I’d be making browser requests to both sites)

On the commandline I navigated to the old site.
I discovered that I had a deactivated install of the redis-cache plugin.

So I then went to the old sites wp-content directory and found the file object-cache.php.

I deleted it. Then went to the new site in my browser, reactivated the redis-cache plugin and made sure it was connected to redis.

This seems to have solved the problem. Both sites now work. The newly installed site has the redis-cache plugin activated. The old site does not.

My next step will be to see if I get problems if the plugin is activate on both sites. But for now the problem is solved.