Hi Craig, As part of my manual testing for the 5.3 RC, I tried out collection admin request restriction and update restriction on a single node setup. I don't have the manual test steps documented but it wasn't too intensive I'd admit. I think the complications involved in stopping specific nodes and bringing them back up stop us from testing the node restarts as part of the automated tests but we should find a way and fix that.
I've just found another issue and opened SOLR-8355 for the same and it involves the "update" permission. As far as patching 5.3.1 go, it's involves more than just this one patch and this patch alone wouldn't help you resolve this issue. You'd certainly need the patch from SOLR-8167. Also, make sure you actually use the 'commit' and not the posted patch as the patch on SOLR-8167 is different from the commit. I don't think you'd need anything other than those patches and whatever comes from 8355 to have a patched 5.3.1. Any help in testing this out would be awesome and thanks for reporting and following up on the issues! On Tue, Dec 1, 2015 at 6:09 AM, Oakley, Craig (NIH/NLM/NCBI) [C] < craig.oak...@nih.gov> wrote: > Thank you, Anshum and Nobel, for your progress on SOLR-8326 > > I have a couple questions to tide me over until 5.4 (hoping to test > security.json a bit further while I wait). > > Given that the seven steps (tar xvzf solr-5.3.1.tgz; tar xvzf > zookeeper-3.4.6.tar.gz; zkServer.sh start zoo_sample.cfg; zkcli.sh -zkhost > localhost:2181 -cmd putfile /security.json ~/security.json; solr start -e > cloud -z localhost:2181; solr stop -p 7574 & solr start -c -p 7574 -s > "example/cloud/node2/solr" -z localhost:2181) demonstrate the problem, are > there a similar set of steps by which one can load _some_ minimal > security.json and still be able to stop & successfully restart one node of > the cluster? (I am wondering what steps were used in the original testing > of 5.3.1) > > Also, has it been verified that the SOLR-8326 patch resolves the > ADDREPLICA bug in addition to the > shutdown-&-restart-one-node-while-keeping-another-node-running bug? > > Also, would it make sense for me to download solr-5.3.1-src.tgz and (in a > test environment) make the changes described in the latest attachment to > SOLR-8326? Or would it be more advisable just to wait for 5.4? I don't know > what may be involved in compiling a new solr.war from the source code. > > Thanks again > > -----Original Message----- > From: Erick Erickson [mailto:erickerick...@gmail.com] > Sent: Tuesday, November 24, 2015 1:25 PM > To: solr-user <solr-user@lucene.apache.org> > Subject: Re: Re:Re: Implementing security.json is breaking ADDREPLICA > > bq: I don't suppose there is an ETA for 5.4? > > Actually, 5.4 is probably in the works within the next month. I'm not > the one cutting the > release, but there's some rumors that a label will be cut this week, > then the "usual" > process is a week or two (sometimes more if bugs are flushed out) before > the > official release. > > Call it the first of the year for safety's sake, but that's a guess. > > Best, > Erick > > On Tue, Nov 24, 2015 at 10:22 AM, Oakley, Craig (NIH/NLM/NCBI) [C] > <craig.oak...@nih.gov> wrote: > > Thanks for the reply, > > > > I don't suppose there is an ETA for 5.4? > > > > > > Thanks again > > > > -----Original Message----- > ... > -- Anshum Gupta
