)]}'
{"/PATCHSET_LEVEL":[{"author":{"_account_id":2475,"name":"Antoine Musso","email":"hashar@free.fr","username":"hashar"},"change_message_id":"a70477e4055e549e020060d77c1e4291050df2b4","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":1,"id":"2f7787ac_992e51b3","updated":"2024-12-18 16:29:29.000000000","message":"Clark pointed the compacting happens when the database are disconnecting. I think that introduces a perceivable delay when we shutdown our Gerrit, but I have never really investigated.\n\nI did restart Wikimedia Gerrit this morning as a follow up of your incident yesterday: `git_file_diff.h2.db` went from 12G to 500MB and `gerrit_file_diff.h2.db` from 3.3G to 683M which is what you are looking for and we have discussed yesterday.\n\n\nSide note, the `systemctl restart gerrit` took 90 seconds to complete:\n```\nDec 18 12:36:38 gerrit1003 systemd[1]: Stopping Gerrit code review tool...\nDec 18 12:37:08 gerrit1003 systemd[1]: Stopped Gerrit code review tool.\n```\nI highly suspect that is systemd terminating it cause we have `TimeoutStopUSec\u003d1min 30s`. I have no idea why it is taking so long to stop, it might be caused by compaction so that is something to keep in mind if your Gerrit suddenly takes age to terminate.","commit_id":"ddca4838692a55df76d0ee2eb64a89bbe83b207a"},{"author":{"_account_id":12898,"name":"Tony Breeds","email":"tony@bakeyournoodle.com","username":"tonyb"},"change_message_id":"71231f4a405a140d943f5f36b88c1f572621ad41","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":2,"id":"1147ad7b_41099a86","updated":"2024-12-19 01:11:51.000000000","message":"Looks good to me.\n\nHolding off on +W to make sure we do this in a controlled way with appropriate warnings and eyeballs","commit_id":"5b5f94d7649f7075912c2343cd0d23e8bd011a82"}],"playbooks/roles/gerrit/templates/docker-compose.yaml.j2":[{"author":{"_account_id":4146,"name":"Clark Boylan","email":"cboylan@sapwetik.org","username":"cboylan"},"change_message_id":"4fd406a8d6367fcd19f048c04052ac8dd14c93bb","unresolved":true,"context_lines":[{"line_number":26,"context_line":"    network_mode: host"},{"line_number":27,"context_line":"    user: gerrit"},{"line_number":28,"context_line":"    stop_signal: SIGHUP"},{"line_number":29,"context_line":"    stop_grace_period: 5m"},{"line_number":30,"context_line":"    volumes:"},{"line_number":31,"context_line":"{% for volume in gerrit_container_volumes %}"},{"line_number":32,"context_line":"      - {{ volume }}"}],"source_content_type":"text/x-jinja2","patch_set":1,"id":"9807737f_a6734cfe","line":29,"updated":"2024-12-18 16:25:37.000000000","message":"Do we think the 5 minute grace period is long enough if each h2 db can spend up to 15 seconds compacting? I suspect it is still enough as Gerrit shuts down very quickly and doesn\u0027t come close to using that 5 minutes prior to this change.","commit_id":"ddca4838692a55df76d0ee2eb64a89bbe83b207a"},{"author":{"_account_id":12898,"name":"Tony Breeds","email":"tony@bakeyournoodle.com","username":"tonyb"},"change_message_id":"71231f4a405a140d943f5f36b88c1f572621ad41","unresolved":true,"context_lines":[{"line_number":26,"context_line":"    network_mode: host"},{"line_number":27,"context_line":"    user: gerrit"},{"line_number":28,"context_line":"    stop_signal: SIGHUP"},{"line_number":29,"context_line":"    stop_grace_period: 5m"},{"line_number":30,"context_line":"    volumes:"},{"line_number":31,"context_line":"{% for volume in gerrit_container_volumes %}"},{"line_number":32,"context_line":"      - {{ volume }}"}],"source_content_type":"text/x-jinja2","patch_set":1,"id":"119b9920_233a7446","line":29,"in_reply_to":"9807737f_a6734cfe","updated":"2024-12-19 01:11:51.000000000","message":"Maybe we can keep an eye on how long the shutdown takes with the longer compacttime.  I suspect that 5 mins is still plenty of wiggle room.","commit_id":"ddca4838692a55df76d0ee2eb64a89bbe83b207a"}]}