Seasearch - error loading snapshot epoch

Hi,

I´m using seafile-pro-server_13.0.29_x86-64_Ubuntu.tar.gz inside a docker container - identical to the official seafile docker container - without any additional services inside the container.

But I can’t get seasearch to work - re-indexing spills out myriads of errors in the seasearch log:

[ERROR] error loading snapshot epoch  resource temporarily unavailable
[ERROR] error loading snapshot epoch error opening segment 669: error loading segment no such file or directory

Running seasearch with log level debug gives:

[2026-11-14 10:59:22] [DEBUG] open writer index shard file_8c second shard 0, current ref count: 1
[2026-11-14 10:59:22] [DEBUG] close writer index shard file_8c second shard 0, current ref count: 0
[2026-11-14 10:59:22] [DEBUG] open writer index shard repo_status_file/yU second shard 0, current ref count: 1
[2026-11-14 10:59:22] [DEBUG] open writer index shard repo_status_file/yU second shard 0, current ref count: 2
[2026-11-14 10:59:22] [DEBUG] close writer index shard file_Va second shard 0, current ref count: 0
[2026-11-14 10:59:22] [DEBUG] open writer index shard file_Va second shard 0, current ref count: 1
[2026-11-14 10:59:22] [DEBUG] close writer index shard file_Va second shard 0, current ref count: 0
[2026-11-14 10:59:22] [ERROR] error loading snapshot epoch: 5: resource temporarily unavailable
[2026-11-14 10:59:22] [ERROR] error loading snapshot epoch: 5: resource temporarily unavailable
[2026-11-14 10:59:22] [ERROR] error loading snapshot epoch: 5: resource temporarily unavailable

The seasearch container / docker-compose.yml is from the official seasearch documentation - communcation seafevents - seasearch seems ok.

  seasearch:  
    image: seafileltd/seasearch:1.0-latest
    container_name: seasearch
    restart: unless-stopped
    user: 1000:1000
    ulimits:
      nofile:
        soft: 200000
        hard: 200000
    volumes:  
      - /srv/seasearch/data:/opt/seasearch-data
      - /srv/seasearch/log:/opt/seasearch-log
      - /etc/timezone:/etc/timezone:ro 
      - /etc/localtime:/etc/localtime:ro
      - /usr/share/zoneinfo:/usr/share/zoneinfo:ro
    environment:
      - TIME_ZONE=${TIME_ZONE}
      - SS_DATA_PATH=/opt/seasearch-data
      - SS_FIRST_ADMIN_USER=...
      - SS_FIRST_ADMIN_PASSWORD=...
      # SS_MAX_OBJ_CACHE_SIZE=10GB
      - SS_LOG_TO_STDOUT=false
      - SS_LOG_DIR=/opt/seasearch-log
      - SS_LOG_LEVEL=info
      - SEAFILE_LOG_TO_STDOUT=false
    expose:
      - 4080

I’ve tried:

  • clear and rebuild the index (pro/pro.py search --clear / update) multiple times
  • increases nofiles ulimit in seafile / seasearch containers to 200k (ulimit -n shows 200k inside the containers)
  • change seasearch filesystem from ZFS to ext4 and back - no difference

does not help! - as soon I start pro/pro.py search --update on cleared seasearch directories, seasearch starts spilling out myriads of these errors

Parallel the seafevents log is flooded with:

seasearch error: {"error":"id not found"}

What could be wrong? - could it be a problem with size/number of files in our seafile libs (30TB)?

Regards,
Gerald

update-1:
switching seafile search to the elasticsearch container makes the re-index process working again - without any errors in the seafevents log.

Thanks for reporting this. We will investigate it in the next week.

Hi,

Thanks for the detailed report. The snapshot errors in the SeaSearch log do not necessarily mean that a request failed, so we’ll need more information to determine what is causing the id not found errors in seafevents.

Could you share the complete SeaSearch and seafevents logs covering the same re-indexing attempt, including the SeaSearch access log (seasearch-access.log)? The access log should help us correlate the requests with their response status codes.