Re: SOLr 6.2.1, dealing with the redirected SOLr web admin

[Shawn Heisey <ap...@elyograg.org>]

On 3/31/2017 1:42 PM, Stewart, Scott A. CTR OSD/DoDEA wrote:
> It seems to be working once I created a dummy core... 

As you may have already figured out, and Alexandre discussed:

The admin UI does not run inside the Solr server.  It runs in your
browser.  When you use a URL in a browser with the # character, that
character and everything that follows it are *NOT* sent to the web
server.  Those are used by the admin UI javascript that is running in
your browser.

URLs with # in them should never be used in a program context.  You need
to talk to API endpoints that do not contain that character.

Possible global URLs you could use to verify that the server is at least
UP (but do not necessarily guarantee full functionality):

http://server:port/solr/admin/info/system
http://server:port/solr/admin/cores
http://server:port/solr/admin/collections (only if running in cloud mode)

What you'll probably want is to check full functionality of one core:

http://server:port/solr/mycollection/admin/ping

The ping handler can be configured in solrconfig.xml, and can include a
healthcheck file which allows the handler to be disabled so it fails the
check even if it's perfectly fine.  I am using the ping handler as the
healthcheck URL in my load balancer (haproxy).  The core that I'm using
for my healthcheck is an aggregator core for a distributed (sharded)
index, so it actually checks the health of multiple cores on multiple
servers with a single URL call.

Thanks,
Shawn

Re: SOLr 6.2.1, dealing with the redirected SOLr web admin

[Shawn Heisey <ap...@elyograg.org>]

On 4/1/2017 9:24 AM, Alexandre Rafalovitch wrote:
> Actually I think the ping handler is now one of the implicit handlers
> and does not need configuration.

This is true.  I was saying that they could configure it beyond the
defaults.  Which I believe is required if the healthcheck file is
desired.  I have not tried, though.

Thanks,
Shawn

Re: SOLr 6.2.1, dealing with the redirected SOLr web admin

[Alexandre Rafalovitch <ar...@gmail.com>]

Actually I think the ping handler is now one of the implicit handlers and
does not need configuration.

Regards,
    Alex

On 1 Apr 2017 10:35 AM, "Shawn Heisey" <ap...@elyograg.org> wrote:

> On 3/31/2017 1:42 PM, Stewart, Scott A. CTR OSD/DoDEA wrote:
> > It seems to be working once I created a dummy core...
>
> As you may have already figured out, and Alexandre discussed:
>
> The admin UI does not run inside the Solr server.  It runs in your
> browser.  When you use a URL in a browser with the # character, that
> character and everything that follows it are *NOT* sent to the web
> server.  Those are used by the admin UI javascript that is running in
> your browser.
>
> URLs with # in them should never be used in a program context.  You need
> to talk to API endpoints that do not contain that character.
>
> Possible global URLs you could use to verify that the server is at least
> UP (but do not necessarily guarantee full functionality):
>
> http://server:port/solr/admin/info/system
> http://server:port/solr/admin/cores
> http://server:port/solr/admin/collections (only if running in cloud mode)
>
> What you'll probably want is to check full functionality of one core:
>
> http://server:port/solr/mycollection/admin/ping
>
> The ping handler can be configured in solrconfig.xml, and can include a
> healthcheck file which allows the handler to be disabled so it fails the
> check even if it's perfectly fine.  I am using the ping handler as the
> healthcheck URL in my load balancer (haproxy).  The core that I'm using
> for my healthcheck is an aggregator core for a distributed (sharded)
> index, so it actually checks the health of multiple cores on multiple
> servers with a single URL call.
>
> Thanks,
> Shawn
>
>