Maintenance

Manage services

• Manage services
  

Manage disk

• The actual state of the disk

  
  
  

Resize the disk size

Example: Resize /tmp and /var

• 1️⃣ Create a new partition

cfdisk /dev/sda

• Select Free Space → New → size 20G (or more)(you can do 10G for /tmp + 10G for /var)

• Set type Linux filesystem

• Write changes → Quit

Then refresh partition table:

partprobe /dev/sda

• 2️⃣ Create a PV and extend VG

pvcreate /dev/sda5   # assuming new partition is sda5
vgextend sysvg /dev/sda5
vgdisplay sysvg      # confirm free space is available

• 3️⃣ Extend LVs

• Add 10G to /tmp to be 14G:

lvextend -L +10G /dev/sysvg/lv_tmp

• Add 10G to /var to be 14G:

lvextend -L +9.9G /dev/sysvg/lv_var

• 4️⃣ Resize filesystems

• Check type:

df -T /tmp /var

• For xfs:

xfs_growfs /tmp
xfs_growfs /var

• 5️⃣ Verify

df -h /tmp /var

Clean/Empty the disk

• Sometime we have problem because of the disk saturation and usually the problem in:

  • /tmp => just clean the tmp folder and restart the services with sudo oprestart

  • Logs files

    • /opt/ontoportal/bioportal_web_ui/shared/log

    • /opt/ontoportal/ontologies_api/shared/log

    • /opt/ontoportal/ncbo_cron/shared/logs

    • /var/log/virtuoso-public

    • /var/log/tomcat/FAIR_CACHE/

  • use df -h to see which folder is full

  • use du -ah -d 1 | sort -h inside the folder to see which files are taken space

Processes, RAM, CPU usage

• with htop inside the server

  
  
  

• With virtus (https://virtu.serveurs.inrae.fr/)

  
  
  

Deployment

• Go to /opt/ontoportal/agroportal_deployment/

• do git pull

• use the deploy.sh script

  ontoportal@localhost:/opt/ontoportal/agroportal_deployment$ ./deploy.sh help
  [-] Error: Missing arguments
  Usage: ./deploy.sh <application> <stage>
  
  Applications:
    - ALL
    - bioportal_web_ui
    - UI
    - ontologies_api
    - API
    - ncbo_cron
    - CRON
  
  Stages:
    - production
    - pre_production
  
  Example:
    ./deploy.sh ontologies_api production
    ./deploy.sh API production

• execute ./deploy.sh ALL production or ./deploy.sh ALL pre_production to deploy all apps

• if something wrong after deployment you can do rollback

  • Method 1: using the deploy.sh script: TO-DO

  • Method 2:

    • go to the app folder example bioportal_web_ui

    • Link the current folder with the release before the deployment (capistrano always will let the 5 last releases)

      ontoportal@localhost:/opt/ontoportal/bioportal_web_ui$ ls current releases/  -altr
      lrwxrwxrwx 1 ontoportal ontoportal  56 Mar 16 16:43 current -> /opt/ontoportal/bioportal_web_ui/releases/20260316164234
      
      releases/:
      total 20
      drwxrwxr-x 18 ontoportal ontoportal 4096 Mar  9 15:54 20260309155421
      drwxrwxr-x 18 ontoportal ontoportal 4096 Mar  9 16:15 20260309161504
      drwxrwxr-x 18 ontoportal ontoportal 4096 Mar 10 17:22 20260310172200
      drwxrwxr-x 18 ontoportal ontoportal 4096 Mar 11 17:06 20260311170634
      drwxrwxr-x 18 ontoportal ontoportal 4096 Mar 16 16:42 20260316164234