X-Git-Url: https://git.arvados.org/arvados.git/blobdiff_plain/972f9d0c4c32082e873bcbe29d5f7ac87b66af4a..cfd1df694f8add55ea34324407208a1f100aea42:/tools/salt-install/provision.sh

diff --git a/tools/salt-install/provision.sh b/tools/salt-install/provision.sh
index 5f6c389b76..4f9c209ef0 100755
--- a/tools/salt-install/provision.sh
+++ b/tools/salt-install/provision.sh
@@ -41,13 +41,13 @@ usage() {
   echo >&2 "  -h, --help                                  Display this help and exit"
   echo >&2 "  --dump-config <dest_dir>                    Dumps the pillars and states to a directory"
   echo >&2 "                                              This parameter does not perform any installation at all. It's"
-  echo >&2 "                                              intended to give you a parsed sot of configuration files so"
+  echo >&2 "                                              intended to give you a parsed set of configuration files so"
   echo >&2 "                                              you can inspect them or use them in you Saltstack infrastructure."
   echo >&2 "                                              It"
   echo >&2 "                                                - parses the pillar and states templates,"
   echo >&2 "                                                - downloads the helper formulas with their desired versions,"
   echo >&2 "                                                - prepares the 'top.sls' files both for pillars and states"
-  echo >&2 "                                                  for the selected role/s"
+  echo >&2 "                                                  for the selected role(s)"
   echo >&2 "                                                - writes the resulting files into <dest_dir>"
   echo >&2 "  -v, --vagrant                               Run in vagrant and use the /vagrant shared dir"
   echo >&2 "  --development                               Run in dev mode, using snakeoil certs"
@@ -189,7 +189,7 @@ WORKBENCH2_EXT_SSL_PORT=3001
 
 SSL_MODE="self-signed"
 USE_LETSENCRYPT_ROUTE53="no"
-CUSTOM_CERTS_DIR="${SCRIPT_DIR}/certs"
+CUSTOM_CERTS_DIR="${SCRIPT_DIR}/local_config_dir/certs"
 
 ## These are ARVADOS-related parameters
 # For a stable release, change RELEASE "production" and VERSION to the
@@ -206,12 +206,17 @@ VERSION="latest"
 # ARVADOS_TAG="2.2.0"
 # BRANCH="main"
 
+# We pin the salt version to avoid potential incompatibilities when a new
+# stable version is released.
+SALT_VERSION="3004"
+
 # Other formula versions we depend on
 POSTGRES_TAG="v0.44.0"
 NGINX_TAG="v2.8.1"
 DOCKER_TAG="v2.4.2"
 LOCALE_TAG="v0.3.4"
 LETSENCRYPT_TAG="v2.1.0"
+LOGROTATE_TAG="v0.14.0"
 
 # Salt's dir
 DUMP_SALT_CONFIG_DIR=""
@@ -226,6 +231,8 @@ T_DIR="/tmp/cluster_tests"
 
 arguments ${@}
 
+declare -A NODES
+
 if [ -s ${CONFIG_FILE} ]; then
   source ${CONFIG_FILE}
 else
@@ -244,14 +251,14 @@ if [ ! -d ${CONFIG_DIR} ]; then
   exit 1
 fi
 
-if grep -q 'fixme_or_this_wont_work' ${CONFIG_FILE} ; then
+if grep -rni 'fixme' ${CONFIG_FILE} ${CONFIG_DIR} ; then
   echo >&2 "The config file ${CONFIG_FILE} has some parameters that need to be modified."
   echo >&2 "Please, fix them and re-run the provision script."
   exit 1
 fi
 
 if ! grep -qE '^[[:alnum:]]{5}$' <<<${CLUSTER} ; then
-  echo >&2 "ERROR: <CLUSTER> must be exactly 5 alphanumeric characters long"
+  echo >&2 "ERROR: <CLUSTER> must be exactly 5 lowercase alphanumeric characters long"
   echo >&2 "Fix the cluster name in the 'local.params' file and re-run the provision script"
   exit 1
 fi
@@ -263,7 +270,7 @@ if [ ! -z "${HOSTNAME_EXT}" ] ; then
   # Make sure that the value configured as IP_INT is a real IP on the system.
   # If we don't error out early here when there is a mismatch, the formula will
   # fail with hard to interpret nginx errors later on.
-  ip addr list |grep -q "${IP_INT}/"
+  ip addr list |grep "${IP_INT}/" >/dev/null
   if [[ $? -ne 0 ]]; then
     echo "Unable to find the IP_INT address '${IP_INT}' on the system, please correct the value in local.params. Exiting..."
     exit 1
@@ -291,7 +298,10 @@ else
       yum install -y  curl git jq
       ;;
     "debian"|"ubuntu")
-      DEBIAN_FRONTEND=noninteractive apt update
+      # Wait 2 minutes for any apt locks to clear
+      # This option is supported from apt 1.9.1 and ignored in older apt versions.
+      # Cf. https://blog.sinjakli.co.uk/2021/10/25/waiting-for-apt-locks-without-the-hacky-bash-scripts/
+      DEBIAN_FRONTEND=noninteractive apt -o DPkg::Lock::Timeout=120 update
       DEBIAN_FRONTEND=noninteractive apt install -y curl git jq
       ;;
   esac
@@ -300,7 +310,7 @@ else
     echo "Salt already installed"
   else
     curl -L https://bootstrap.saltstack.com -o /tmp/bootstrap_salt.sh
-    sh /tmp/bootstrap_salt.sh -XdfP -x python3
+    sh /tmp/bootstrap_salt.sh -XdfP -x python3 stable ${SALT_VERSION}
     /bin/systemctl stop salt-minion.service
     /bin/systemctl disable salt-minion.service
   fi
@@ -326,34 +336,43 @@ mkdir -p ${S_DIR} ${F_DIR} ${P_DIR} ${T_DIR}
 # Get the formula and dependencies
 cd ${F_DIR} || exit 1
 echo "Cloning formulas"
-rm -rf ${F_DIR}/* || exit 1
-git clone --quiet https://github.com/saltstack-formulas/docker-formula.git ${F_DIR}/docker
-( cd docker && git checkout --quiet tags/"${DOCKER_TAG}" -b "${DOCKER_TAG}" )
+test -d docker && ( cd docker && git fetch ) \
+  || git clone --quiet https://github.com/saltstack-formulas/docker-formula.git ${F_DIR}/docker
+( cd docker && git checkout --quiet tags/"${DOCKER_TAG}" )
 
 echo "...locale"
-git clone --quiet https://github.com/saltstack-formulas/locale-formula.git ${F_DIR}/locale
-( cd locale && git checkout --quiet tags/"${LOCALE_TAG}" -b "${LOCALE_TAG}" )
+test -d locale && ( cd locale && git fetch ) \
+  || git clone --quiet https://github.com/saltstack-formulas/locale-formula.git ${F_DIR}/locale
+( cd locale && git checkout --quiet tags/"${LOCALE_TAG}" )
 
 echo "...nginx"
-git clone --quiet https://github.com/saltstack-formulas/nginx-formula.git ${F_DIR}/nginx
-( cd nginx && git checkout --quiet tags/"${NGINX_TAG}" -b "${NGINX_TAG}" )
+test -d nginx && ( cd nginx && git fetch ) \
+  || git clone --quiet https://github.com/saltstack-formulas/nginx-formula.git ${F_DIR}/nginx
+( cd nginx && git checkout --quiet tags/"${NGINX_TAG}" )
 
 echo "...postgres"
-git clone --quiet https://github.com/saltstack-formulas/postgres-formula.git ${F_DIR}/postgres
-( cd postgres && git checkout --quiet tags/"${POSTGRES_TAG}" -b "${POSTGRES_TAG}" )
+test -d postgres && ( cd postgres && git fetch ) \
+  || git clone --quiet https://github.com/saltstack-formulas/postgres-formula.git ${F_DIR}/postgres
+( cd postgres && git checkout --quiet tags/"${POSTGRES_TAG}" )
 
 echo "...letsencrypt"
-git clone --quiet https://github.com/saltstack-formulas/letsencrypt-formula.git ${F_DIR}/letsencrypt
-( cd letsencrypt && git checkout --quiet tags/"${LETSENCRYPT_TAG}" -b "${LETSENCRYPT_TAG}" )
+test -d letsencrypt && ( cd letsencrypt && git fetch ) \
+  || git clone --quiet https://github.com/saltstack-formulas/letsencrypt-formula.git ${F_DIR}/letsencrypt
+( cd letsencrypt && git checkout --quiet tags/"${LETSENCRYPT_TAG}" )
+
+echo "...logrotate"
+test -d logrotate && ( cd logrotate && git fetch ) \
+  || git clone --quiet https://github.com/saltstack-formulas/logrotate-formula.git ${F_DIR}/logrotate
+( cd logrotate && git checkout --quiet tags/"${LOGROTATE_TAG}" )
 
 echo "...arvados"
-git clone --quiet https://git.arvados.org/arvados-formula.git ${F_DIR}/arvados
+test -d arvados || git clone --quiet https://git.arvados.org/arvados-formula.git ${F_DIR}/arvados
 
 # If we want to try a specific branch of the formula
 if [ "x${BRANCH}" != "x" ]; then
   ( cd ${F_DIR}/arvados && git checkout --quiet -t origin/"${BRANCH}" -b "${BRANCH}" )
 elif [ "x${ARVADOS_TAG}" != "x" ]; then
-( cd ${F_DIR}/arvados && git checkout --quiet tags/"${ARVADOS_TAG}" -b "${ARVADOS_TAG}" )
+  ( cd ${F_DIR}/arvados && git checkout --quiet tags/"${ARVADOS_TAG}" -b "${ARVADOS_TAG}" )
 fi
 
 if [ "x${VAGRANT}" = "xyes" ]; then
@@ -417,30 +436,34 @@ for f in $(ls "${SOURCE_PILLARS_DIR}"/*); do
   "${f}" > "${P_DIR}"/$(basename "${f}")
 done
 
-if [ "x${TEST}" = "xyes" ] && [ ! -d "${SOURCE_TESTS_DIR}" ]; then
-  echo "You requested to run tests, but ${SOURCE_TESTS_DIR} does not exist or is not a directory. Exiting."
-  exit 1
-fi
-mkdir -p ${T_DIR}
-# Replace cluster and domain name in the test files
-for f in $(ls "${SOURCE_TESTS_DIR}"/*); do
-  FILTERS="s#__CLUSTER__#${CLUSTER}#g;
-       s#__CONTROLLER_EXT_SSL_PORT__#${CONTROLLER_EXT_SSL_PORT}#g;
-       s#__DOMAIN__#${DOMAIN}#g;
-       s#__IP_INT__#${IP_INT}#g;
-       s#__INITIAL_USER_EMAIL__#${INITIAL_USER_EMAIL}#g;
-       s#__INITIAL_USER_PASSWORD__#${INITIAL_USER_PASSWORD}#g
-       s#__INITIAL_USER__#${INITIAL_USER}#g;
-       s#__DATABASE_PASSWORD__#${DATABASE_PASSWORD}#g;
-       s#__SYSTEM_ROOT_TOKEN__#${SYSTEM_ROOT_TOKEN}#g"
-  if [ "$USE_SINGLE_HOSTNAME" = "yes" ]; then
-    FILTERS="s#__CLUSTER__.__DOMAIN__#${HOSTNAME_EXT}#g;
-       $FILTERS"
+if [ ! -d "${SOURCE_TESTS_DIR}" ]; then
+  echo "WARNING: The tests directory was not copied to \"${SOURCE_TESTS_DIR}\"."
+  if [ "x${TEST}" = "xyes" ]; then
+    echo "WARNING: Disabling tests for this installation."
   fi
-  sed "$FILTERS" \
-    "${f}" > ${T_DIR}/$(basename "${f}")
-done
-chmod 755 ${T_DIR}/run-test.sh
+  TEST="no"
+else
+  mkdir -p ${T_DIR}
+  # Replace cluster and domain name in the test files
+  for f in $(ls "${SOURCE_TESTS_DIR}"/*); do
+    FILTERS="s#__CLUSTER__#${CLUSTER}#g;
+         s#__CONTROLLER_EXT_SSL_PORT__#${CONTROLLER_EXT_SSL_PORT}#g;
+         s#__DOMAIN__#${DOMAIN}#g;
+         s#__IP_INT__#${IP_INT}#g;
+         s#__INITIAL_USER_EMAIL__#${INITIAL_USER_EMAIL}#g;
+         s#__INITIAL_USER_PASSWORD__#${INITIAL_USER_PASSWORD}#g
+         s#__INITIAL_USER__#${INITIAL_USER}#g;
+         s#__DATABASE_PASSWORD__#${DATABASE_PASSWORD}#g;
+         s#__SYSTEM_ROOT_TOKEN__#${SYSTEM_ROOT_TOKEN}#g"
+    if [ "$USE_SINGLE_HOSTNAME" = "yes" ]; then
+      FILTERS="s#__CLUSTER__.__DOMAIN__#${HOSTNAME_EXT}#g;
+         $FILTERS"
+    fi
+    sed "$FILTERS" \
+      "${f}" > ${T_DIR}/$(basename "${f}")
+  done
+  chmod 755 ${T_DIR}/run-test.sh
+fi
 
 # Replace helper state files that differ from the formula's examples
 if [ -d "${SOURCE_STATES_DIR}" ]; then
@@ -534,25 +557,29 @@ if [ -z "${ROLES}" ]; then
     fi
     grep -q "letsencrypt"     ${S_DIR}/top.sls || echo "    - letsencrypt" >> ${S_DIR}/top.sls
   else
-    # Use custom certs, as both bring-your-own and self-signed are copied using this state
-    # Copy certs to formula extra/files
-    # In dev mode, the files will be created and put in the destination directory by the
-    # snakeoil_certs.sls state file
     mkdir -p /srv/salt/certs
-    cp -rv ${CUSTOM_CERTS_DIR}/* /srv/salt/certs/
-    # We add the custom_certs state
-    grep -q "custom_certs"    ${S_DIR}/top.sls || echo "    - extra.custom_certs" >> ${S_DIR}/top.sls
+    if [ "${SSL_MODE}" = "bring-your-own" ]; then
+      # Copy certs to formula extra/files
+      cp -rv ${CUSTOM_CERTS_DIR}/* /srv/salt/certs/
+      # We add the custom_certs state
+      grep -q "custom_certs"    ${S_DIR}/top.sls || echo "    - extra.custom_certs" >> ${S_DIR}/top.sls
+    fi
+    # In self-signed mode, the certificate files will be created and put in the
+    # destination directory by the snakeoil_certs.sls state file
   fi
 
   echo "    - postgres" >> ${S_DIR}/top.sls
+  echo "    - logrotate" >> ${S_DIR}/top.sls
   echo "    - docker.software" >> ${S_DIR}/top.sls
   echo "    - arvados" >> ${S_DIR}/top.sls
   echo "    - extra.shell_sudo_passwordless" >> ${S_DIR}/top.sls
   echo "    - extra.shell_cron_add_login_sync" >> ${S_DIR}/top.sls
+  echo "    - extra.passenger_rvm" >> ${S_DIR}/top.sls
 
   # Pillars
   echo "    - docker" >> ${P_DIR}/top.sls
   echo "    - nginx_api_configuration" >> ${P_DIR}/top.sls
+  echo "    - logrotate_api" >> ${P_DIR}/top.sls
   echo "    - nginx_controller_configuration" >> ${P_DIR}/top.sls
   echo "    - nginx_keepproxy_configuration" >> ${P_DIR}/top.sls
   echo "    - nginx_keepweb_configuration" >> ${P_DIR}/top.sls
@@ -561,6 +588,7 @@ if [ -z "${ROLES}" ]; then
   echo "    - nginx_webshell_configuration" >> ${P_DIR}/top.sls
   echo "    - nginx_workbench2_configuration" >> ${P_DIR}/top.sls
   echo "    - nginx_workbench_configuration" >> ${P_DIR}/top.sls
+  echo "    - logrotate_wb1" >> ${P_DIR}/top.sls
   echo "    - postgresql" >> ${P_DIR}/top.sls
 
   # We need to tweak the Nginx's pillar depending whether we want plan nginx or nginx+passenger
@@ -573,11 +601,27 @@ if [ -z "${ROLES}" ]; then
     fi
     grep -q "letsencrypt" ${P_DIR}/top.sls || echo "    - letsencrypt" >> ${P_DIR}/top.sls
 
-    # As the pillar differ whether we use LE or custom certs, we need to do a final edition on them
-    for c in controller websocket workbench workbench2 webshell download collections keepproxy; do
-      sed -i "s/__CERT_REQUIRES__/cmd: create-initial-cert-${c}.${CLUSTER}.${DOMAIN}*/g;
-              s#__CERT_PEM__#/etc/letsencrypt/live/${c}.${CLUSTER}.${DOMAIN}/fullchain.pem#g;
-              s#__CERT_KEY__#/etc/letsencrypt/live/${c}.${CLUSTER}.${DOMAIN}/privkey.pem#g" \
+    hosts=("controller" "websocket" "workbench" "workbench2" "webshell" "keepproxy")
+    if [ ${USE_SINGLE_HOSTNAME} = "no" ]; then
+      hosts+=("download" "collections")
+    else
+      hosts+=("keepweb")
+    fi
+
+    for c in "${hosts[@]}"; do
+      # Are we in a single-host-single-hostname env?
+      if [ "${USE_SINGLE_HOSTNAME}" = "yes" ]; then
+        # Are we in a single-host-single-hostname env?
+        CERT_NAME=${HOSTNAME_EXT}
+      else
+        # We are in a multiple-hostnames env
+        CERT_NAME=${c}.${CLUSTER}.${DOMAIN}
+      fi
+
+      # As the pillar differs whether we use LE or custom certs, we need to do a final edition on them
+      sed -i "s/__CERT_REQUIRES__/cmd: create-initial-cert-${CERT_NAME}*/g;
+              s#__CERT_PEM__#/etc/letsencrypt/live/${CERT_NAME}/fullchain.pem#g;
+              s#__CERT_KEY__#/etc/letsencrypt/live/${CERT_NAME}/privkey.pem#g" \
       ${P_DIR}/nginx_${c}_configuration.sls
     done
   else
@@ -587,7 +631,7 @@ if [ -z "${ROLES}" ]; then
     echo "extra_custom_certs_dir: /srv/salt/certs" > ${P_DIR}/extra_custom_certs.sls
     echo "extra_custom_certs:" >> ${P_DIR}/extra_custom_certs.sls
 
-    for c in controller websocket workbench workbench2 webshell keepweb keepproxy; do
+    for c in controller websocket workbench workbench2 webshell keepweb keepproxy shell; do
       # Are we in a single-host-single-hostname env?
       if [ "${USE_SINGLE_HOSTNAME}" = "yes" ]; then
         # Are we in a single-host-single-hostname env?
@@ -634,13 +678,13 @@ else
       ;;
       "api")
         # States
-        # FIXME: https://dev.arvados.org/issues/17352
-        grep -q "postgres.client" ${S_DIR}/top.sls || echo "    - postgres.client" >> ${S_DIR}/top.sls
+        grep -q "    - logrotate" ${S_DIR}/top.sls || echo "    - logrotate" >> ${S_DIR}/top.sls
         if grep -q "    - nginx.*$" ${S_DIR}/top.sls; then
           sed -i s/"^    - nginx.*$"/"    - nginx.passenger"/g ${S_DIR}/top.sls
         else
           echo "    - nginx.passenger" >> ${S_DIR}/top.sls
         fi
+        echo "    - extra.passenger_rvm" >> ${S_DIR}/top.sls
         ### If we don't install and run LE before arvados-api-server, it fails and breaks everything
         ### after it. So we add this here as we are, after all, sharing the host for api and controller
         if [ "${SSL_MODE}" = "lets-encrypt" ]; then
@@ -657,6 +701,7 @@ else
         fi
         grep -q "arvados.${R}" ${S_DIR}/top.sls    || echo "    - arvados.${R}" >> ${S_DIR}/top.sls
         # Pillars
+        grep -q "logrotate_api" ${P_DIR}/top.sls            || echo "    - logrotate_api" >> ${P_DIR}/top.sls
         grep -q "aws_credentials" ${P_DIR}/top.sls          || echo "    - aws_credentials" >> ${P_DIR}/top.sls
         grep -q "postgresql" ${P_DIR}/top.sls               || echo "    - postgresql" >> ${P_DIR}/top.sls
         grep -q "nginx_passenger" ${P_DIR}/top.sls          || echo "    - nginx_passenger" >> ${P_DIR}/top.sls
@@ -669,6 +714,7 @@ else
       "controller" | "websocket" | "workbench" | "workbench2" | "webshell" | "keepweb" | "keepproxy")
         # States
         if [ "${R}" = "workbench" ]; then
+          grep -q "    - logrotate" ${S_DIR}/top.sls || echo "    - logrotate" >> ${S_DIR}/top.sls
           NGINX_INSTALL_SOURCE="install_from_phusionpassenger"
           if grep -q "    - nginx$" ${S_DIR}/top.sls; then
             sed -i s/"^    - nginx.*$"/"    - nginx.passenger"/g ${S_DIR}/top.sls
@@ -701,6 +747,9 @@ else
           grep -q "arvados.${R}" ${S_DIR}/top.sls || echo "    - arvados.${R}" >> ${S_DIR}/top.sls
         fi
         # Pillars
+        if [ "${R}" = "workbench" ]; then
+          grep -q "logrotate_wb1" ${P_DIR}/top.sls || echo "    - logrotate_wb1" >> ${P_DIR}/top.sls
+        fi
         grep -q "nginx_passenger" ${P_DIR}/top.sls          || echo "    - nginx_passenger" >> ${P_DIR}/top.sls
         grep -q "nginx_${R}_configuration" ${P_DIR}/top.sls || echo "    - nginx_${R}_configuration" >> ${P_DIR}/top.sls
         # Special case for keepweb