Re: pgadmin4 2.0 and Scram
On Fri, Oct 6, 2017 at 10:18 PM, Murtuza Zabuawala < murtuza.zabuaw...@enterprisedb.com> wrote: > On Sat, Oct 7, 2017 at 4:48 AM, Jeff Janes wrote: > >> Since pgAdmin4 2.0 is targetting v10 of PostgreSQL, I thought it would >> support scram authentication in addition to md5. In fact, I thought this >> would happen automatically simply by virtue of using the latest libpq.dll, >> without any coding changes to pgAdmin4 itself. >> >> But when I set myself a scram-hashed password (using psql) and then try >> to connect with pgAdmin4, I get the message "authentication method 10 not >> supported", which is the message you get from pre-v10 versions of libpq. >> >> Is this a known issue? Does the installer include the wrong libpq.dll? >> > pgAdmin4 uses psycopg2 <http://initd.org/psycopg/docs/> in order to > communicate with Postgres and psycopg2 is lacking support for Scram at a > moment, and I think developers are already working on adding support for > PG10 (https://github.com/psycopg/psycopg2/pull/576). > That item is about changing the password through psycopg2, which does require work from the client software (if you don't want to send the password itself to the server). But I am just talking about using an existing password to log on, which is purely a function of libpq and shouldn't require any changes from the client. When I log in from python scripts via psycopg2 or from perl with DBD::Pg, old versions of psycopg2 or perl work fine with a SCRAM password as long as the latest libpq is loaded. However, I've only tested python direct usage and perl from Linux, and only tested pgAdmin from Windows. So it could be an OS issue. I've tried using ListDLLs to see which dll pgAdmin is loading to make sure it is not finding the wrong one, but it says it doesn't have permission to inspect the DLL for that program, so I don't know what else to check from Windows. Cheers, Jeff
pgadmin4 2.0 and Scram
Since pgAdmin4 2.0 is targetting v10 of PostgreSQL, I thought it would support scram authentication in addition to md5. In fact, I thought this would happen automatically simply by virtue of using the latest libpq.dll, without any coding changes to pgAdmin4 itself. But when I set myself a scram-hashed password (using psql) and then try to connect with pgAdmin4, I get the message "authentication method 10 not supported", which is the message you get from pre-v10 versions of libpq. Is this a known issue? Does the installer include the wrong libpq.dll? Cheers, Jeff
Re: pgadmin4 ssl HELP!
On Fri, Aug 11, 2017 at 4:30 AM, Иванов Евгений Сергеевич < zhenj...@gmail.com> wrote: > I can not solve the problem. I use pgadmin4 v 1.6 through the web > (apache2). In the database connection settings in the SSL-mode > parameter, select "Verify-FULL". When I try to connect, I get an > error: "root certificate file" /var/lib/wwwrun/.postgresql/root.crt > "does not exist. Either provide the file or change sslmode to disable > server certificate verification." > > Problem: The path by which the application searches for the > "/var/lib/wwwrun/.postgresql/root.crt" certificate is the same for all > users of pgadmin4. That makes it impossible to use this application as > a multi-user application when using SSL. > If multiple users are all using the same machine, then there is already one person they all trust--the person who has root access to that machine. Why wouldn't that person sign all the certs, or arrange for them to be signed? Cheers, Jeff
Re: Windows testing required: Updated runtime
On Thu, Jul 6, 2017 at 7:59 AM, Dave Page wrote: > All, > > I've put an interim build of pgAdmin 4 at https://developer.pgadmin. > org/~dpage/pgadmin4-1.5-x86.exe for Windows users. This build uses a work > in progress patch to replace the slow Qt components with ActiveQt + the > Microsoft web browser control to significantly improve performance. > > The version also includes close to 100 other bug fixes and changes, most > noticeably performance improvements in the Query Tool that make it > significantly faster than pgAdmin 3 with large result sets, and the ability > to open the Query Tool and Debugger in new tabs that can be dragged onto > different windows and therefore displays. > It starts quite a bit faster than it used to. It used to be close to a minute, now it is 15 seconds from cold and 10 if it were recently running and then started again. I haven't really played with the features, so don't have anything to report there. Cheers, Jeff