Внедрение KESL на ВМ2 + замена CLAMAV на KESL

This commit is contained in:
mi
2026-09-08 01:39:37 +03:00
parent 85df788f2d
commit fdfdeaffb4
43 changed files with 2210 additions and 329 deletions
+22 -12
View File
@@ -1,10 +1,10 @@
Логирование операций
1. env:
editor /etc/han/vm1.env
изменение Release_Version
+ две переменные:
HOST_NAME=devhanapp \ prodhanapp.ihan.ru.
HOST_NAME=prodhanapp.ihan.ru \devhanapp \
TELEMETRYGEN_IMAGE=ghcr.io/open-telemetry/opentelemetry-collector-contrib/telemetrygen@sha256:9461e2c7213219467e5f775ea8652d15c990afb4d1ca94e62344f2025a7587d1
2. Обновить образы:
@@ -25,7 +25,7 @@ tar --exclude=backend/.env `
-czf "vm1-backend-$Release.tar.gz" `
-C C:/Users/MI/Documents/Assistent/HAN_chat_specification/VM1_app/codebase backend
Get-FileHash "vm1-backend-$Release.tar.gz" -Algorithm SHA256
scp "vm1-backend-$Release.tar.gz" devVM1Deploy:/var/lib/han-deploy/incoming/
scp "vm1-backend-$Release.tar.gz" prodVM1Deploy:/var/lib/han-deploy/incoming/
```
RELEASE='1.0.1'
@@ -91,14 +91,14 @@ sudo apt purge -y dos2unix
scp `
C:/Users/MI/.ssh/devDeploy.pub `
C:/Users/MI/.ssh/devAdmin.pub `
devVM1Deploy:/var/lib/han-deploy/incoming/
C:/Users/MI/.ssh/prodDeploy.pub `
C:/Users/MI/.ssh/prodAdmin.pub `
prodVM1Deploy:/var/lib/han-deploy/incoming/
install -m 0600 -o root -g root /var/lib/han-deploy/incoming/devDeploy.pub /root/bootstrap/deploy.pub
install -m 0600 -o root -g root /var/lib/han-deploy/incoming/devAdmin.pub /root/bootstrap/admin.pub
install -m 0600 -o root -g root /var/lib/han-deploy/incoming/prodDeploy.pub /root/bootstrap/deploy.pub
install -m 0600 -o root -g root /var/lib/han-deploy/incoming/prodAdmin.pub /root/bootstrap/admin.pub
rm -f /var/lib/han-deploy/incoming/devDeploy.pub /var/lib/han-deploy/incoming/devAdmin.pub
rm -f /var/lib/han-deploy/incoming/prodDeploy.pub /var/lib/han-deploy/incoming/prodAdmin.pub
cd /opt/han-chat/current/backend
@@ -109,6 +109,12 @@ ADMIN_AUTHORIZED_KEY_FILE=/root/bootstrap/admin.pub \
HARDEN_SSH=true SKIP_APT_UPGRADE=true \
/opt/han-chat/current/backend/deployment/scripts/setup-vm.sh
export OTEL_REMOTE_ENDPOINT="$(awk -F= '$1=="OTEL_REMOTE_ENDPOINT"{print substr($0,index($0,"=")+1)}' /etc/han/vm1.env)"
export OTEL_REMOTE_TLS_INSECURE="$(awk -F= '$1=="OTEL_REMOTE_TLS_INSECURE"{print substr($0,index($0,"=")+1)}' /etc/han/vm1.env)"
export HOST_NAME="$(awk -F= '$1=="HOST_NAME"{print substr($0,index($0,"=")+1)}' /etc/han/vm1.env)"
export APP_ENV="$(awk -F= '$1=="APP_ENV"{print substr($0,index($0,"=")+1)}' /etc/han/vm1.env)"
export RELEASE_VERSION="$(awk -F= '$1=="RELEASE_VERSION"{print substr($0,index($0,"=")+1)}' /etc/han/vm1.env)"
printf '%s\n' "$OTEL_REMOTE_ENDPOINT"
/usr/local/bin/otelcol-contrib validate \
--config=/opt/han-chat/current/backend/deployment/observability/otel-host-collector.yaml
@@ -116,12 +122,16 @@ chmod 0755 /opt/han-chat/current/backend/scripts/validate-env
/opt/han-chat/current/backend/deployment/preflight.sh
systemctl restart han-stack@production.service
если ошибка systemctl status han-stack@production.service
если ошибка
journalctl -u han-stack@production.service --no-pager -n 80 --output=cat
### !!! (образы могут устареть в будущем)
docker pull cr.selcloud.ru/han-images/nginx@sha256:2cd7d32967b400c15bb6aaa18e97e99d429b2b6f2bfd54b965fc7d42a91a228e
docker pull cr.selcloud.ru/han-images/sms-service@sha256:404685c45b7391a78f95c014ce8fdff64a84ccd7ae5b280e7a65b524c109d898
systemctl start han-host-otel-collector@production.service
rm -f /root/bootstrap/deploy.pub /root/bootstrap/admin.pub
Canary test: CONFIG_FILE=/etc/han/vm1.env ./deployment/scripts/verify-observability.sh
Canary test:
chmod 0755 /opt/han-chat/current/backend/deployment/scripts/verify-observability.sh
CONFIG_FILE=/etc/han/vm1.env ./deployment/scripts/verify-observability.sh
rm -f /root/bootstrap/deploy.pub /root/bootstrap/admin.pub
+647
View File
@@ -0,0 +1,647 @@
# Ставим KES на ВМ1
# Ставим KES на ВМ2
Готовим новый образ Message Safety
MESSAGE_SAFETY_IMAGE=cr.selcloud.ru/han-images/han-message-safety@sha256:225808bd111be03112ab6693b5a10e3d97ae8cdbe5370a5d0aa5a8e2278a3044
Копируем проект
```powershell
$Release = "1.0.1"
tar --exclude=services/.env `
--exclude='services/**/__pycache__' `
--exclude='services/**/.pytest_cache' `
--exclude='services/**/.ruff_cache' `
-czf "vm2-services-$Release.tar.gz" -C C:\Users\MI\Documents\Assistent\HAN_chat_specification\VM2_services\codebase services
Get-FileHash "vm2-services-$Release.tar.gz" -Algorithm SHA256
scp "vm2-services-$Release.tar.gz" prodVM2Deploy:/var/lib/han-deploy/incoming/
scp "vm2-services-$Release.tar.gz" devVM2Deploy:/var/lib/han-deploy/incoming/
```
Проверить размер диска
RELEASE='1.0.1'
EXPECTED_SHA256='1030DE6A50BA945FAEC5A45E1D42C47A01808E737AA1955871F3D2ED089F8879'
ARCHIVE="/var/lib/han-deploy/incoming/vm2-services-${RELEASE}.tar.gz"
printf '%s %s\n' "$EXPECTED_SHA256" "$ARCHIVE" | sha256sum --check -
tar -tvzf "$ARCHIVE"
if tar -tzf "$ARCHIVE" | grep -Eq '(^/|(^|/)\.\.(/|$)|^services/\.env$)'; then
echo 'ОШИБКА: архив содержит небезопасный путь или .env' >&2
exit 1
fi
if tar -tzf "$ARCHIVE" | grep -Ev '^services(/|$)' | grep -q .; then
echo 'ОШИБКА: архив содержит файлы вне каталога services' >&2
exit 1
fi
if tar -tvzf "$ARCHIVE" | awk '$1 ~ /^[lh]/ { found=1 } END { exit !found }'; then
echo 'ОШИБКА: архив содержит symlink или hardlink' >&2
exit 1
fi
install -d -m 0755 -o root -g root /opt/han-chat/services
STAGING="$(mktemp -d /opt/han-chat/.vm2-release.XXXXXX)"
tar --extract --gzip --file "$ARCHIVE" \
--directory "$STAGING" --no-same-owner --no-same-permissions
test -f "$STAGING/services/docker-compose.yml"
rsync -a --delete --exclude=.env \
--chown=root:root --chmod=D755,F644 \
"$STAGING/services/" /opt/han-chat/services/
rm -rf -- "$STAGING"
chmod 0755 /opt/han-chat/services/deployment/preflight.sh
cd /opt/han-chat/services/
sudo apt update
sudo apt install -y dos2unix
find . -type f \( \
-name '*.sh' -o -name '*.py' -o -name '*.service' -o -name '*.sudoers' -o -name 'han-compose' -o -name 'han-secrets' -o -name 'han-message-safety-mode' -o -name '*.yml' -o -name '*.conf*' -o -name 'preflight.sh' -o -name '*.yaml' -o -name '*.md' -o -name '*.toml' -o -name 'Dockerfile' -o -name '*.acl*' \
\) -exec dos2unix {} +
find -type f -exec file {} \; | grep -i 'CRLF'
sudo apt remove -y dos2unix
sudo apt purge -y dos2unix
Копируем образ на ВМ:
scp "C:\Users\MI\Documents\job\HAN_new_life\HANapp\docs\kes 12.4\049-16-d-01.iso" "C:\Users\MI\Documents\job\HAN_new_life\HANapp\docs\kes 12.4\049-16-d-01.sig" prodVM2Deploy:/var/lib/han-deploy/incoming/
Проверка файла:
apt install -y rhash
ISO='/var/lib/han-deploy/incoming/049-16-d-01.iso'
command -v rhash
rhash --gost "$ISO"
Должно быть 6a94b16afad211e8b9be5ec86f5379184f2b3a9e5869843fe763e2796e5ac1d3
sudo apt remove -y rhash
sudo apt purge -y rhash
Если совпадёт, извлеките DEB:
```sh
install -d -m 0700 -o root -g root \
/root/kesl-install /mnt/kesl-iso
mount -o ro,loop "$ISO" /mnt/kesl-iso
find /mnt/kesl-iso -type f \
\( -iname 'kesl_*_amd64.deb' -o -iname '*.deb' \) |
sort
```
Затем:
```sh
DEB='/mnt/kesl-iso/kesl/kesl_12.4.0-1225_amd64.deb'
dpkg-deb -f "$DEB" Package Version Architecture
sha256sum "$DEB"
install -m 0600 -o root -g root \
"$DEB" /root/kesl-install/kesl.deb
umount /mnt/kesl-iso
rmdir /mnt/kesl-iso
```
Ожидаемые свойства: kesl, версия 12.4.x, архитектура amd64.
После этого можно выполнять:
apt-get install /root/kesl-install/kesl.deb
Далее
```sh
install -m 0600 -o root -g root /dev/null /root/kesl-install/autoinstall.ini
cat > /root/kesl-install/autoinstall.ini <<'EOF'
KSVLA_MODE=No
ENDPOINT_AGENT_MODE=No
EULA_AGREED=Yes
PRIVACY_POLICY_AGREED=Yes
USE_KSN=No
GROUP_CLEAN=Yes
LOCALE=ru_RU.UTF-8
INSTALL_LICENSE=None
UPDATER_SOURCE=KLServers
UPDATE_EXECUTE=No
KERNEL_SRCS_INSTALL=No
USE_GUI=No
CONFIGURE_SELINUX=No
DISABLE_PROTECTION=Yes
INTERCEPTOR_MODE=UseFanotify
ENABLE_TRACES_ON_FIRST_STARTUP=No
EOF
chmod 0600 /root/kesl-install/autoinstall.ini
stat -c '%U:%G:%a %n' /root/kesl-install/autoinstall.ini
/opt/kaspersky/kesl/bin/kesl-setup.pl \
--autoinstall=/root/kesl-install/autoinstall.ini
echo "kesl-setup exit=$?"
systemctl --no-pager status kesl
kesl-control --app-info --json
kesl-control --get-task-list
```
```sh
set +o history
unset HISTFILE
read -rsp 'KESL activation code: ' KESL_CODE; echo
kesl-control --add-active-key "$KESL_CODE"
unset KESL_CODE
set -o history
kesl-control -L --query
```
Не стартуйте Update (task 6) и File Threat Protection (task 1), пока не срежете дефолты памяти. На 8 ГБ `ScanMemoryLimit=8192` и `MaxMemory=auto` (до 50% RAM, минимум 2 ГБ) слишком много, тем более пока жив ClamAV.
Ключ из `-L --query` в evidence/репозиторий не копируйте — достаточно статуса `Active subscription`.
Дальше — секция 3 runbook. Кандидат для ВМ2: `ScanMemoryLimit=1024`, `MaxMemory=2048MB`. CPU: `OnDemandCPULimit=15`, как на пилоте ВМ1, пока ClamAV не снят.
**1. Снимок текущих настроек**
```sh
kesl-control --get-app-settings \
--file /root/kesl-install/app-settings.before.ini
install -m 0600 -o root -g root \
/var/opt/kaspersky/kesl/common/kesl.ini \
/root/kesl-install/kesl.ini.before
awk '/^\[General\]/,/^\[/' /var/opt/kaspersky/kesl/common/kesl.ini | head -n 40
```
**2. Лимит CPU (KESL ещё работает)**
```sh
kesl-control --set-app-settings \
UseOnDemandCPULimit=Yes OnDemandCPULimit=15
```
**3. Память — только при остановленном KESL, файл целиком не подменять**
Сначала посмотрите `[General]` из шага 1. Если там уже есть `ScanMemoryLimit` и `MaxMemory`:
```sh
systemctl stop kesl
systemctl is-active kesl
grep -nE '^(ScanMemoryLimit|MaxMemory)=' /var/opt/kaspersky/kesl/common/kesl.ini
sed -i \
-e 's/^ScanMemoryLimit=8192$/ScanMemoryLimit=1024/' \
-e 's/^MaxMemory=auto$/MaxMemory=2048MB/' \
/var/opt/kaspersky/kesl/common/kesl.ini
grep -nE '^(ScanMemoryLimit|MaxMemory)=' /var/opt/kaspersky/kesl/common/kesl.ini
systemctl start kesl
systemctl is-active kesl
kesl-control --get-app-settings | grep -iE 'UseOnDemandCPULimit|OnDemandCPULimit'
```
**Ожидаемо после рестарта**
- `kesl.service` = `active`
- в settings: `ScanMemoryLimit=1024`, `MaxMemory=2048MB`, `UseOnDemandCPULimit=Yes`, `OnDemandCPULimit=15`
- task 1 и 6 по-прежнему `Stopped`
- базы всё ещё не загружены
После этого — загрузка баз:
`kesl-control --start-task 6 -W`
```sh
kesl-control --get-task-state 6
kesl-control --app-info --json
START="$(LC_ALL=C date +'%Y/%b/%d %H:%M:%S;1')"
kesl-control --set-schedule 6 RuleType=Hourly \
"StartTime=${START}" RunMissedStartRules=No RandomInterval=0
kesl-control --get-schedule 6
```
Обязательный следующий шаг — рестарт KESL. В --app-info есть:
Post-update actions: The application module has been updated. Restart the application.
Task 1 до рестарта не запускайте.
```sh
systemctl restart kesl
systemctl is-active kesl
kesl-control --app-info --json
grep -nE '^(ScanMemoryLimit|MaxMemory)=' /var/opt/kaspersky/kesl/common/kesl.ini
kesl-control --get-app-settings | grep -iE 'UseOnDemandCPULimit|OnDemandCPULimit'
kesl-control --get-schedule 6
kesl-control --get-task-state 1
kesl-control --get-task-state 6
```
```sh
kesl-control --get-settings 1 \
--file /root/kesl-install/file-threat.before.ini
kesl-control --set-settings 1 \
--add-exclusion /var/lib/docker/volumes/han-processing_redis-safety-data/_data
kesl-control --set-settings 1 \
--add-exclusion /var/lib/docker/volumes/han-processing_otel-queue/_data
kesl-control --start-task 1
kesl-control --set-settings 1 \
ActionOnThreat=DisinfectDeleteIfNotPossible ScanArchived=No
kesl-control --get-task-state 1
kesl-control --get-settings 1
```
Сразу снимите coexistence, пока OAS прогревается:
```sh
date -Is
free -h
swapon --show
vmstat 1 5
systemctl is-active kesl docker han-processing.service
kesl-control --app-info --json | grep -E 'File Threat|databases loaded|license'
/usr/local/sbin/han-vm2-compose ps
docker stats --no-stream
```
Установка брокера
```sh
set -Eeuo pipefail
BASE=/opt/han-chat/services/deployment/kesl
for file in \
kesl_scan_broker.py \
han-kesl-scan-broker.service \
han-kesl-scan-broker.socket \
han-kesl-scan-broker.tmpfiles.conf
do
test -f "$BASE/$file"
test ! -L "$BASE/$file"
done
if getent group han-message-safety >/dev/null; then
test "$(getent group han-message-safety | cut -d: -f3)" = 10001
else
! getent group 10001 >/dev/null
groupadd --system --gid 10001 han-message-safety
fi
install -d -m 0755 -o root -g root /usr/local/libexec
install -m 0750 -o root -g root \
"$BASE/kesl_scan_broker.py" \
/usr/local/libexec/han-kesl-scan-broker
install -m 0644 -o root -g root \
"$BASE/han-kesl-scan-broker.service" \
/etc/systemd/system/han-kesl-scan-broker.service
install -m 0644 -o root -g root \
"$BASE/han-kesl-scan-broker.socket" \
/etc/systemd/system/han-kesl-scan-broker.socket
install -m 0644 -o root -g root \
"$BASE/han-kesl-scan-broker.tmpfiles.conf" \
/etc/tmpfiles.d/han-kesl-scan-broker.conf
systemd-tmpfiles --create \
/etc/tmpfiles.d/han-kesl-scan-broker.conf
systemctl daemon-reload
systemd-analyze verify \
/etc/systemd/system/han-kesl-scan-broker.service \
/etc/systemd/system/han-kesl-scan-broker.socket
```
Если systemd-analyze verify завершился без ошибок, запуск:
```sh
systemctl enable --now han-kesl-scan-broker.socket
systemctl is-active kesl han-kesl-scan-broker.socket
stat -c '%U:%G:%a %n' /run/han-kesl /run/han-kesl/scan.sock
/usr/local/libexec/han-kesl-scan-broker --probe
```
Ожидаемые права:
/run/han-kesl root:han-message-safety:750
/run/han-kesl/scan.sock root:han-message-safety:660
Создаём папку:
```sh
install -d -m 0700 -o root -g han-message-safety \
/var/lib/han-kesl-scan \
/var/lib/han-kesl-scan/staging
```
Проверка и запуск socket
```
systemd-analyze verify \
/etc/systemd/system/han-kesl-scan-broker.service \
/etc/systemd/system/han-kesl-scan-broker.socket
systemctl enable --now han-kesl-scan-broker.socket
systemctl is-active kesl han-kesl-scan-broker.socket
stat -c '%U:%G:%a %n' /run/han-kesl /run/han-kesl/scan.sock
/usr/local/libexec/han-kesl-scan-broker --probe
journalctl --since '-10 min' --no-pager -u han-kesl-scan-broker.service
```
Обновить .env
editor /opt/han-chat/services/.env
MESSAGE_SAFETY_IMAGE=cr.selcloud.ru/han-images/han-message-safety@sha256:225808bd111be03112ab6693b5a10e3d97ae8cdbe5370a5d0aa5a8e2278a3044
Удалить
CLAMAV_IMAGE
MESSAGE_SAFETY_CLAMAV_HOST
MESSAGE_SAFETY_CLAMAV_PORT
Добавить
MESSAGE_SAFETY_ANTIVIRUS_SOCKET=/run/han-kesl/scan.sock
Активировать config schema 2
NEXT_VERSION='2'
/usr/local/sbin/han-vm2-compose --profile ops run --rm \
--entrypoint message-safety-config message-safety-migrate \
validate /app/app/artifacts/seed-config.yaml
/usr/local/sbin/han-vm2-compose --profile ops run --rm \
--entrypoint message-safety-config message-safety-migrate \
create /app/app/artifacts/seed-config.yaml \
--version "$NEXT_VERSION" --actor '<OPERATOR>'
/usr/local/sbin/han-vm2-compose --profile ops run --rm \
--entrypoint message-safety-config message-safety-migrate \
activate --version "$NEXT_VERSION" --approved-by '<APPROVER>'
Вернуть затертые IP
editor /opt/han-chat/services/nginx/allowlists/bitrix-webhook-allowlist.conf
allow 195.208.184.200;
allow 89.208.230.2;
allow 195.208.184.34;
allow 89.208.228.219;
deny all;
editor /opt/han-chat/services/nginx/allowlists/private-caller-allowlist.conf
allow 10.0.0.11;
deny all;
Переключить только Message Safety
cd /opt/han-chat/services/
deployment/preflight.sh
/usr/local/sbin/han-vm2-compose config --quiet
/usr/local/sbin/han-vm2-compose up -d --no-deps --force-recreate \
message-safety-api message-safety-worker
/usr/local/sbin/han-vm2-compose ps
/usr/local/sbin/han-vm2-compose ps --format "table {{.Service}}\t{{.Status}}\t{{.Ports}}"
Тестирование:
Секция 7 — не один скрипт, а **пять независимых gates**. EICAR только после письменного Security; образец не класть в git, chat и evidence — брать с [eicar.org](https://www.eicar.org). Пункты 45 и cutover **нельзя** закрыть текущим образом Message Safety (он ещё на ClamAV): сначала 1–3 на хосте.
Все команды на `devhanservice` — под **root**. `han-processing.service` не restart/stop.
---
### 0. Перед тестами
```sh
kesl-control --get-task-state 1
kesl-control --app-info --json | grep -E 'File Threat|databases loaded|license'
systemctl is-active kesl han-kesl-scan-broker.socket
test -S /run/han-kesl/scan.sock
stat -c '%U:%G:%a %n' /run/han-kesl/scan.sock
/usr/local/libexec/han-kesl-scan-broker --probe
```
Нужно: task 1 `Started`, базы `Yes`, socket `root:han-message-safety:660`, probe `"status":"ready"`. Если sock нет — `systemctl restart han-kesl-scan-broker.socket`, не `.service`.
---
### 1. Host OAS (АВЗ.1)
Каталог **не** в docker volume, `/tmp`, release, staging:
```sh
install -d -m 0700 -o root -g root /root/kesl-eicar-test
date -Is
kesl-control --get-task-state 1
```
Положите официальный `eicar.com` в `/root/kesl-eicar-test/eicar.com`:
curl -fsSL -o /root/kesl-eicar-test/eicar.com \
'https://secure.eicar.org/eicar.com'
chmod 0600 /root/kesl-eicar-test/eicar.com
```sh
cat /root/kesl-eicar-test/eicar.com >/dev/null
ls -l /root/kesl-eicar-test
kesl-control -E --query -n 20 --reverse
```
Ожидаемо: объект удалён/в Backup, в журнале `EICAR-Test-File` / `ThreatDetected`. Если OAS на запись молчит (как на пилоте ВМ1) — явный ODS:
```sh
kesl-control --scan-file /root/kesl-eicar-test/eicar.com
kesl-control -E --query -n 20 --reverse
```
Уборка (не весь Backup):
```sh
kesl-control -B --query --reverse -n 20
kesl-control -B --mass-remove --query "DetectName == 'EICAR-Test-File'"
rm -rf /root/kesl-eicar-test
```
---
### 2. Broker: clean / EICAR / KESL down
install -d -m 0700 -o root -g root /root/kesl-install
cat > /root/kesl-install/broker-client.py <<'PY'
#!/usr/bin/env python3
import asyncio, json, struct, sys
SOCK = "/run/han-kesl/scan.sock"
async def call(op: str, body: bytes = b"") -> dict:
r, w = await asyncio.open_unix_connection(SOCK)
header = {"op": op}
if op == "scan":
header["size"] = len(body)
raw = json.dumps(header, separators=(",", ":")).encode()
w.write(struct.pack(">I", len(raw)) + raw + body)
await w.drain()
n = struct.unpack(">I", await r.readexactly(4))[0]
resp = json.loads(await r.readexactly(n))
w.close()
await w.wait_closed()
return resp
async def main() -> None:
op = sys.argv[1]
body = sys.stdin.buffer.read() if op == "scan" else b""
print(json.dumps(await call(op, body), ensure_ascii=True, sort_keys=True))
if __name__ == "__main__":
if sys.argv[1:] not in (["status"], ["scan"]):
raise SystemExit("usage: broker-client.py status|scan")
asyncio.run(main())
PY
chmod 0700 /root/kesl-install/broker-client.py
Проверка позитивного решения
python3 /root/kesl-install/broker-client.py status
printf 'HAN kesl broker clean canary\n' | python3 /root/kesl-install/broker-client.py scan
Проверка негативного решения
curl -fsSL 'https://secure.eicar.org/eicar.com' |
python3 /root/kesl-install/broker-client.py scan
Проверка недоступного сканнера
systemctl stop kesl
systemctl is-active kesl || true
python3 /root/kesl-install/broker-client.py status
printf 'error canary\n' |
python3 /root/kesl-install/broker-client.py scan
Оба запроса должны вернуть scanner_unavailable.
Затем восстановить:
systemctl start kesl
systemctl is-active kesl
kesl-control --start-task 1
kesl-control --get-task-state 1
systemctl restart han-kesl-scan-broker.socket
/usr/local/libexec/han-kesl-scan-broker --probe
---
### 3. 5 слотов и ~2 файла/с
На 4 ГБ DEV это нагрузка: смотрите `free -h` / `vmstat`. Откат OAS: `kesl-control --stop-task 1`.
```sh
python3 - <<'PY'
import asyncio, json, struct, time, statistics
SOCK = "/run/han-kesl/scan.sock"
BODY = b"HAN kesl throughput canary\n"
async def scan():
r, w = await asyncio.open_unix_connection(SOCK)
h = json.dumps({"op": "scan", "size": len(BODY)}, separators=(",", ":")).encode()
t0 = time.perf_counter()
w.write(struct.pack(">I", len(h)) + h + BODY)
await w.drain()
n = struct.unpack(">I", await r.readexactly(4))[0]
resp = json.loads(await r.readexactly(n))
dt = time.perf_counter() - t0
w.close(); await w.wait_closed()
return dt, resp
async def main():
burst = await asyncio.gather(*[scan() for _ in range(5)])
print("burst5", [(round(dt, 3), r.get("verdict") or r.get("error")) for dt, r in burst])
times = []
for _ in range(20):
dt, r = await scan()
assert r.get("verdict") == "clean", r
times.append(dt)
await asyncio.sleep(0.5) # ~2 file/s
times.sort()
p95 = times[int(0.95 * (len(times) - 1))]
print(f"n={len(times)} p50={times[len(times)//2]:.3f}s p95={p95:.3f}s max={times[-1]:.3f}s")
asyncio.run(main())
PY
```
Gate: все `clean`, p95 ≤ 60 с. Queue age p95 ≤ 5 с — из метрик Safety (после пункта 4), не из этого скрипта.
---
### 4. Message Safety 200 / 403 / 503
Сейчас контейнеры ещё на ClamAV. Пункт 4 — **после** recreate `message-safety-api`/`worker` на образ с bind `/run/han-kesl` (digest из окна 1.0.1). Clamd пока можно не снимать.
Проверка, что API видит KESL:
```sh
/usr/local/sbin/han-vm2-compose exec -T message-safety-api \
python -c 'import json,urllib.request; print(json.dumps(json.load(urllib.request.urlopen("http://127.0.0.1:8080/health/ready")), indent=2))'
```
`antivirus=ok`, `files=ready` (не mock).
- **clean `200`** — text canary с ВМ1, как в `RUNBOOK.ru.md` (`/internal/safety/v2/messages/check`, `content_kind=text`).
- **malware `403` + sticky** — file через **реальный** versioned object в quarantine S3 (выдуманный key нельзя; EICAR в bucket только если Security разрешил). Повтор того же `message_id`+body — тот же `403`.
- **scanner error `503`** — `systemctl stop kesl`, file-check, ожидание `503` / files unavailable; затем start kesl + restart **socket** + `--start-task 1`.
Пока worker ходит в clamd, это **не** gate KESL.
---
### 5. Smoke, firewall, 24 ч
```sh
iptables -S HAN-CHAT-VM2
ufw status verbose
/usr/local/sbin/han-vm2-compose ps
free -h
swapon --show
kesl-control --get-task-state 6
kesl-control --get-schedule 6
```
С ВМ1: `openssl s_client` + `curl` `/internal/safety/status` из `RUNBOOK.ru.md`. Сравните `HAN-CHAT-VM2` с baseline. 24 ч: без OOM/restart/5xx, hourly Update сам зелёный.
curl --fail --silent --show-error \
--cacert /etc/han/ca/vm2-internal-ca.crt \
https://processing.internal:8443/internal/safety/status
https://prodhanservice.ihan.ru:8443/internal/safety/status
---
### Cutover (только после 1–5)
Новая **монотонная** config version (не реактивировать старую):
```sh
cd /opt/han-chat/services/
NEXT_VERSION='3'
/usr/local/sbin/han-vm2-compose --profile ops run --rm \
--entrypoint message-safety-config message-safety-migrate \
create /app/app/artifacts/seed-config.yaml \
--version "$NEXT_VERSION" --actor '<OPERATOR>'
/usr/local/sbin/han-vm2-compose --profile ops run --rm \
--entrypoint message-safety-config message-safety-migrate \
activate --version "$NEXT_VERSION" --approved-by '<APPROVER>'
/usr/local/sbin/han-vm2-compose up -d --no-deps --force-recreate \
message-safety-api message-safety-worker
```
В `.env` / compose не должно остаться `clamd`/`freshclam`. Затем:
```sh
/usr/local/sbin/han-vm2-compose up -d --remove-orphans
/usr/local/sbin/han-vm2-compose ps
docker ps -a --format '{{.Names}}' | grep -E 'clamd|freshclam' || echo 'no clamd/freshclam'
docker volume ls | grep -i clam || true
```
/usr/local/sbin/han-vm2-compose exec -T nginx nginx -t -c /etc/nginx/nginx.conf
NGINX_ID=$(/usr/local/sbin/han-vm2-compose ps --status running --quiet nginx)
docker kill --signal HUP "$NGINX_ID"
/usr/local/sbin/han-vm2-compose exec -T nginx getent hosts message-safety-api
/usr/local/sbin/han-vm2-compose exec -T message-safety-api \
python -c 'import json,urllib.request; print(json.dumps(json.load(urllib.request.urlopen("http://127.0.0.1:8080/health/ready")), indent=2))'
Старый signature-CDN egress закрывается отдельно в firewall/SG — не через `compose down -v`.
На этом DEV-хосте (3.8 ГиБ, swap 391 МиБ) пункт 3 и file-EICAR через Safety лучше не совмещать с чем-то ещё. Если нужно, могу выписать только gate 1+2 компактным блоком под копирование в `260907_1.md`.