From 9b111e24937df6f65e770cfb50007b43c381bbee Mon Sep 17 00:00:00 2001 From: Joshua Tauberer Date: Sat, 3 Sep 2022 07:50:42 -0400 Subject: [PATCH 1/4] Update to Nextcloud 23.0.8 (contacts 4.2.0, calendar 3.5.0) --- setup/nextcloud.sh | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) diff --git a/setup/nextcloud.sh b/setup/nextcloud.sh index 37181a83..5c42c655 100755 --- a/setup/nextcloud.sh +++ b/setup/nextcloud.sh @@ -21,8 +21,8 @@ echo "Installing Nextcloud (contacts/calendar)..." # we automatically install intermediate versions as needed. # * The hash is the SHA1 hash of the ZIP package, which you can find by just running this script and # copying it from the error message when it doesn't match what is below. -nextcloud_ver=23.0.4 -nextcloud_hash=87afec0bf90b3c66289e6fedd851867bc5a58f01 +nextcloud_ver=23.0.8 +nextcloud_hash=9d63416a0697eeecf243d09461f5881f8997f50a # Nextcloud apps # -------------- @@ -33,10 +33,10 @@ nextcloud_hash=87afec0bf90b3c66289e6fedd851867bc5a58f01 # https://github.com/nextcloud/user_external/blob/master/appinfo/info.xml # * The hash is the SHA1 hash of the ZIP package, which you can find by just running this script and # copying it from the error message when it doesn't match what is below. -contacts_ver=4.1.0 -contacts_hash=697f6b4a664e928d72414ea2731cb2c9d1dc3077 -calendar_ver=3.2.2 -calendar_hash=ce4030ab57f523f33d5396c6a81396d440756f5f +contacts_ver=4.2.0 +contacts_hash=79b506574834db5e1b6ab47aadd4041e12ad9a9c +calendar_ver=3.5.0 +calendar_hash=941381536287a015081669513f8f79f6f262508a user_external_ver=3.0.0 user_external_hash=0df781b261f55bbde73d8c92da3f99397000972f From 0a970f4bb2dba052ce1c321dd8b4e9f016a56550 Mon Sep 17 00:00:00 2001 From: Joshua Tauberer Date: Sun, 2 Oct 2022 21:20:21 -0400 Subject: [PATCH 2/4] Use nsd-control to refresh nsd after zone files are rewritten rather than 'service nsd restart' I am not sure if this was the problem but nsd didn't serve updated zonefiles on my box and 'service nsd restart' must have been used, so maybe it doesn't reload zones. --- management/dns_update.py | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/management/dns_update.py b/management/dns_update.py index 45ea94fa..a9c44f01 100755 --- a/management/dns_update.py +++ b/management/dns_update.py @@ -96,9 +96,9 @@ def do_dns_update(env, force=False): if len(updated_domains) == 0: updated_domains.append("DNS configuration") - # Kick nsd if anything changed. + # Tell nsd to reload changed zone files. if len(updated_domains) > 0: - shell('check_call', ["/usr/sbin/service", "nsd", "restart"]) + shell('check_call', ["/usr/sbin/nsd-control", "reload"]) # Write the OpenDKIM configuration tables for all of the mail domains. from mailconfig import get_mail_domains From 22a6270657ed6906a5572c2779a381a0f55f8eab Mon Sep 17 00:00:00 2001 From: Joshua Tauberer Date: Sat, 8 Oct 2022 08:16:51 -0400 Subject: [PATCH 3/4] Remove old setup step to uninstall acme library --- setup/management.sh | 12 ------------ 1 file changed, 12 deletions(-) diff --git a/setup/management.sh b/setup/management.sh index cebed8d5..1d1f4cc5 100755 --- a/setup/management.sh +++ b/setup/management.sh @@ -7,18 +7,6 @@ echo "Installing Mail-in-a-Box system management daemon..." # DEPENDENCIES -# We used to install management daemon-related Python packages -# directly to /usr/local/lib. We moved to a virtualenv because -# these packages might conflict with apt-installed packages. -# We may have a lingering version of acme that conflcits with -# certbot, which we're about to install below, so remove it -# first. Once acme is installed by an apt package, this might -# break the package version and `apt-get install --reinstall python3-acme` -# might be needed in that case. -while [ -d /usr/local/lib/python3.4/dist-packages/acme ]; do - pip3 uninstall -y acme; -done - # duplicity is used to make backups of user data. # # virtualenv is used to isolate the Python 3 packages we From 89cd9fb6113770cdda3c458e80fa601eb1569173 Mon Sep 17 00:00:00 2001 From: Joshua Tauberer Date: Sat, 8 Oct 2022 08:17:24 -0400 Subject: [PATCH 4/4] Increase gunicorn's worker timeout since some /admin commands take a long time --- setup/management.sh | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/setup/management.sh b/setup/management.sh index 1d1f4cc5..7fe4da25 100755 --- a/setup/management.sh +++ b/setup/management.sh @@ -78,6 +78,8 @@ rm -f /tmp/bootstrap.zip # Create an init script to start the management daemon and keep it # running after a reboot. +# Set a long timeout since some commands take a while to run, matching +# the timeout we set for PHP (fastcgi_read_timeout in the nginx confs). # Note: Authentication currently breaks with more than 1 gunicorn worker. cat > $inst_dir/start <