Yes unix. It was an amazing moment.
On Mon, Jun 4, 2018, 11:28 PM Erick Erickson <erickerick...@gmail.com> wrote: > bq. To be clear I deleted the actual index files out from under the > running master > > I'm assuming *nix here since Windows won't let you delete a file that > has an open file handle... > > Did you then restart the master? Aside from any checks about refusing > to replicate an empty index, just deleting the files without reloading > the core (or restarting Solr) won't do it. Files disappear all the > time out from under a running Solr, specifically during merges. > However on linux machines those files are still available until the > last file handle is closed, and Solr still points to those files until > a hard commit or restart. More technically, until a new Searcher is > opened. > > Here's a check: query the _master_ after deleting the index. Absent a > commit of some sort, you'll probably get results. > > Mind you you still may not see a replication of an empty index, this > is just another complication that you need to be aware of. > > Best, > Erick > > > > On Mon, Jun 4, 2018 at 5:44 PM, Walter Underwood <wun...@wunderwood.org> > wrote: > > Check the logs. I bet it says something like “refusing to fetch empty > index.” > > > > wunder > > Walter Underwood > > wun...@wunderwood.org > > http://observer.wunderwood.org/ (my blog) > > > >> On Jun 4, 2018, at 1:41 PM, Jeff Courtade <courtadej...@gmail.com> > wrote: > >> > >> I am thankful for that! > >> > >> Could you point me at something that explains this maybe? > >> > >> J > >> > >> On Mon, Jun 4, 2018, 4:31 PM Shawn Heisey <apa...@elyograg.org> wrote: > >> > >>> On 6/4/2018 12:15 PM, Jeff Courtade wrote: > >>>> This was strange as I would have thought the replica would have > >>> replicated > >>>> an empty index from the master. > >>> > >>> Solr actually has protections in place to specifically PREVENT index > >>> replication when the master has an empty index. This is so that a > >>> accidental mistake made on the master that clears its index cannot > >>> completely delete the index on the slave. > >>> > >>> After you index some data into the new master index, if change meets > the > >>> "replicateAfter" settings, then the next poll made by the slave should > >>> replicate the new index. > >>> > >>> Thanks, > >>> Shawn > >>> > >>> -- > >> > >> Jeff Courtade > >> M: 240.507.6116 > > > -- Jeff Courtade M: 240.507.6116
