)]}'
{
  "commit": "97bcceca9f64d74d01c49f6955fad68432c90ff7",
  "tree": "e33489a267c08977d9ec01ad203a45e040dd8803",
  "parents": [
    "f0386ca40ed33bff3721765bb01f22f7b17805e6"
  ],
  "author": {
    "name": "Oleg Bondarev",
    "email": "obondarev@mirantis.com",
    "time": "Wed May 15 11:20:22 2019 +0400"
  },
  "committer": {
    "name": "Oleg Bondarev",
    "email": "obondarev@mirantis.com",
    "time": "Tue May 28 11:22:26 2019 +0000"
  },
  "message": "Bump fs.inotify.max_user_instances on compute nodes for scale\n\nThis problem appeared with one of latest dnsmasq versions.\n\nSince each dnsmasq process consumes one inotify socket, the default\nvalue of fs.inotify.max_user_instances which is 128 lets us scale to\nonly around a 116 neutron subnets (a few other sockets are used by other\nprocesses on the system). Since, we need to provide better defaults,\nthis patch proposes to bump this value to 1024 by default, while giving\nthe user a way to cahnge it. Based on\nhttps://unix.stackexchange.com/a/13757 each inotify watch takes 1KB of\nmemory and we have fs.inotify.max_user_watches set to 8192 by default.\nThis means that even in the worst case we won\u0027t be using more than 8MB\nof memory. Bumping the fs.inotify.max_user_instances value to 4096 is\nsafe because there is fs.inotify.max_user_watches which caps the total\nnumber of files that can be watched by all the inotify instances a user\n\nChange-Id: I5794bd9c815c24302a707a1b88a5c25caa6ee953\nRelated-Prod: PROD-30216\n",
  "tree_diff": [
    {
      "type": "modify",
      "old_id": "c8a0922cd31a0930de691e42b9800cd5728d949d",
      "old_mode": 33188,
      "old_path": "neutron/compute/cluster.yml",
      "new_id": "0766df73d79fdbdc315f58c0bb7c8ac842ae826f",
      "new_mode": 33188,
      "new_path": "neutron/compute/cluster.yml"
    }
  ]
}