This test reproduces the problem reported in bug #59181 ("[berlin] web
services fail to start on reboot due to anonip").
* gnu/tests/web.scm (run-webserver-test) <extra-args>: New keyword argument.
(%nginx+anonip-os, nginx-anonip-tests, %test-nginx+anonip): New variables.
Change-Id: I7bbc8fb9f2ab33ce89bf1c0945d7ddbddf26a7ea
* gnu/tests/web.scm (%test-anonip): New test.
(%anonip-os): New variables.
(run-anonip-test): New procedure.
Change-Id: Ieed210a784dbdeee8a498e74b6c0e31cb72cd9b8
Previously, it'd simply return the error arguments, which would not always
fail a 'test-assert' test, for example.
* gnu/tests/web.scm (retry-on-error): Re-throw the error when the attempts
have been exhausted.
Change-Id: Ic1468d8ba23f0f78377e90d19bccb2878dc56f0e
In Shepherd 1.0, the “running value” of processes is no longer a plain
integer; instead, it is a (process …) sexp. This commit adjusts tests
to this change in a way that works both for 1.0 and for previous
versions.
* gnu/tests/databases.scm (run-memcached-test)
(run-mysql-test): Don’t expect PID to be a number.
* gnu/tests/docker.scm (run-docker-test)
(run-docker-system-test, run-oci-container-test): Likewise.
* gnu/tests/guix.scm (run-guix-build-coordinator-test)
(run-guix-data-service-test, run-nar-herder-test)
(run-bffe-test): Likewise.
* gnu/tests/ldap.scm (run-ldap-test): Likewise.
* gnu/tests/monitoring.scm (run-prometheus-node-exporter-server-test):
Likewise.
* gnu/tests/virtualization.scm (run-libvirt-test)
(run-qemu-guest-agent-test, run-childhurd-test): Likewise.
* gnu/tests/web.scm (run-webserver-test, run-php-fpm-test)
(run-hpcguix-web-server-test, run-patchwork-test)
(run-agate-test): Likewise
* gnu/tests/ssh.scm (run-ssh-test): Accept a number, an ‘inetd-service’
sexp, or a ‘process’ sexp.
Change-Id: I8c7a37a981f0788780fbc33752a38e7f9a026437
This is a follow-up of 73eeeeafbb.
* gnu/tests/web.scm (run-tailon-test): Wait for the TCP port 8080 instead of
relying on start-service.
Signed-off-by: Mathieu Othacehe <othacehe@gnu.org>
The patchwork test sometimes fail because there's not enough available RAM in
the marionette. Increase the VM memory size to 1024MiB.
* gnu/tests/web.scm (run-patchwork-test): Increase VM memory size to 1024MiB.
The "http-get" test is sometimes failing because the Web server is not yet
initialized and returns the 500 error code.
Use the retry-or-error procedure, like in the tailon test to do a few retries.
* gnu/tests/web.scm (run-tailon-test): Move "retry-or-error" procedure to the
top level and adapt its call.
(run-patchwork-test): Use it.
Fixes <https://bugs.gnu.org/46737>.
PostgreSQL running with a different socket directory to the default one in the
package itself breaks some services, this commit restores the previous
behaviour where PostgreSQL by default will run with a socket directory that
matches the default used by PostgreSQL packaged for Guix.
Switching to a different default value can happen, but only alongside changing
the PostgreSQL package.
* gnu/services/databases.scm (<postgresql-config-file>)[socket-directory]:
Change default to #false.
* doc/guix.texi (Database Services): Update documentation, and specify a
different value for disabling connections via sockets.
* gnu/tests/guix.scm (%guix-data-service-os): Use default PostgreSQL
behaviour.
* gnu/tests/monitoring.scm (%zabbix-os): Likewise.
* gnu/tests/web.scm (patchwork-os): Likewise.
Signed-off-by: Leo Famulari <leo@famulari.name>
Currently, if the postgresql package major version changes, this is going to
break the service upon upgrade, because PostgreSQL will reject the data files
from the differing major version of the service.
Because it's important to either keep running a particular major version, or
intentionally upgrade, I think the configuration would be better with no
default. I think this is also going to be helpful when trying to assist users
upgrading PostgreSQL.
* gnu/services/databases.scm (<postgresql-configuration>): Remove default for
postgresql.
(postgresql-service-type): Remove the default value.
* gnu/tests/databases.scm (%postgresql-os): Update accordingly.
* gnu/tests/guix.scm (%guix-data-service-os): Update accordingly.
* gnu/tests/monitoring.scm (%zabbix-os): Update accordingly.
* gnu/tests/web.scm (patchwork-os): Update accordingly.
* doc/guix.texi (PostgreSQL): Update accordingly.
Record type descriptors were made private in
a143e92446, but a usage of them was
forgotten in the tests files.
* gnu/tests/web.scm (patchwork-initial-database-setup-service): Use
accessors to access field values instead of unexported type descriptor.
* gnu/service/web.scm (<patchwork-database-configuration>
<patchwork-settings-module>, <patchwork-configuration>): New record types.
(patchwork-virtualhost): New procedure.
(patchwork-service-type): New variable.
* gnu/tests/web.scm (%test-patchwork): New variable.
* doc/guix.text (Web Services): Document it.
* gnu/services/web.scm (<varnish-configuration>): New record type.
(%varnish-accounts, %varnish-service-type): New variables.
(varnish-shepherd-service): New procedure.
* gnu/tests/web.scm (%varnish-vcl, %varnish-os): New variables.
(%test-varnish): New test.
* doc/guix.texi (Web Services): Document it.
This allows (gnu services admin) to remain deeper in the module graph
and to be used by (gnu services web).
* gnu/services/admin.scm (<tailon-configuration-file>)
(tailon-configuration-files-string)
(tailon-configuration-file-compiler, <tailon-configuration>)
(tailon-shepherd-service, %tailon-accounts)
(tailon-service-type): Move to...
* gnu/services/web.scm: ... here.
* gnu/tests/admin.scm: Remove. Move test to...
* gnu/tests/web.scm (%tailon-os)
(run-tailon-test, %test-tailon): ... here.
Since commit dc7b3e5633, 'start-service' returns
the Shepherd's representation of the service as a sexp, and '#f' if the
service fails to start. Also, it doesn't throw an exception when the service
fails to start, so relying on an exception instead of relying on its return
value is a false positive.
* gnu/tests/base.scm (run-nss-mdns-test): Split tests so to check the return
value of 'start-service'.
(run-mcron-test, run-nss-mdns-test): Use 'test-assert' with the return value
of 'start-service'.
* gnu/tests/admin.scm (run-tailon-test): Idem.
* gnu/tests/dict.scm (run-dicod-test): Idem.
* gnu/tests/mail.scm (run-opensmtpd-test, run-exim-test, run-dovecot-test):
Idem.
* gnu/tests/messaging.scm (run-xmpp-test, run-bitlbee-test): Idem.
* gnu/tests/nfs.scm (run-nfs-test): Idem.
* gnu/tests/rsync.scm (run-rsync-test): Idem.
* gnu/tests/ssh.scm (run-ssh-test): Idem.
* gnu/tests/version-control.scm (run-cgit-test, run-git-http-test): Idem.
* gnu/tests/web.scm (run-php-fpm-test): Idem.
So that it can also be used for other web servers.
* gnu/tests/web.scm (%index.html-contents): Change nginx to guix.
(%make-http-root): Move the index.html file from /srv to /srv/http.
(%nginx-servers): Remove the setting of root.
(run-nginx-test, run-webserver-test): Rename run-nginx-test to
run-webserver-test and generalise its behaviour
(%test-nginx): Change to use run-webserver-test, rather than run-nginx-test.
Fixes <https://bugs.gnu.org/29769>.
Reported by Mark H Weaver <mhw@netris.org>.
This is a followup to 8b223ceac4.
* gnu/tests/web.scm (%php-fpm-nginx-server-blocks): Turn 'listen' into a
list.
If nginx is configured with a ssl-certificate file, and ssl-certificate-key,
it will fail to start unless these exist. To avoid this happening, change the
default to #f.
* gnu/services/web.scm (<nginx-server-configuration>)
[ssl-certificate,ssl-certificate-key]: Set the defaults to #f.
* gnu/tests/web.scm (%nginx-servers): Remove redundant
nginx-server-configuration fields.
* doc/guix.texi (Web Services): Update examples and documentation.
