Merge pull request #7 from thalesmg/shortids

sync latest changes to master

.gitignore

@@ -0,0 +1 @@
+venv/

fetch-logs.yaml

@@ -0,0 +1,20 @@
+---
+
+- name: define hosts
+  hosts: localhost
+  gather_facts: no
+  roles:
+    - define_inventory
+
+- name: fetch logs
+  hosts: emqx
+  become: yes
+  become_user: root
+  tasks:
+    - name: syslog
+      fetch:
+        src: "/var/log/syslog"
+        dest: "./tmp/data/{{ emqx_script_result_file }}/{{ inventory_hostname_short }}/syslog"
+        flat: yes
+
+...

files/tuning-emqx.sh

@@ -13,9 +13,9 @@ set -eu
   # ulimit -n 2097152
 
   # Increase number of incoming connections backlog:
-  sysctl -w net.core.somaxconn=32768
-  sysctl -w net.ipv4.tcp_max_syn_backlog=16384
-  sysctl -w net.core.netdev_max_backlog=16384
+  # sysctl -w net.core.somaxconn=32768
+  # sysctl -w net.ipv4.tcp_max_syn_backlog=16384
+  # sysctl -w net.core.netdev_max_backlog=16384
 
   # local port range
   # sysctl -w net.ipv4.ip_local_port_range='1000 65535'
@@ -28,32 +28,33 @@ set -eu
   # sysctl -w net.core.wmem_default=262144
   # sysctl -w net.core.rmem_max=16777216
   # sysctl -w net.core.wmem_max=16777216
-  sysctl -w net.core.optmem_max=16777216
+  # sysctl -w net.core.optmem_max=16777216
 
   # sysctl -w net.ipv4.tcp_mem='16777216 16777216 16777216'
-  sysctl -w net.ipv4.tcp_rmem='1024 4096 16777216'
-  sysctl -w net.ipv4.tcp_wmem='1024 4096 16777216'
+  # sysctl -w net.ipv4.tcp_rmem='1024 4096 16777216'
+  # sysctl -w net.ipv4.tcp_wmem='1024 4096 16777216'
 
   # TMG FIXME: Outdated or module not loaded; ignore
+  # ... but trying again after bumping into a limit.
   # TCP connection tracking:
-  # sysctl -w net.nf_conntrack_max=1000000
+  sysctl -w net.nf_conntrack_max=10000000
   # sysctl -w net.netfilter.nf_conntrack_max=1000000
   # sysctl -w net.netfilter.nf_conntrack_tcp_timeout_time_wait=30
 
   # TIME-WAIT Bucket Pool, Recycling and Reuse:
-  sysctl -w net.ipv4.tcp_max_tw_buckets=1048576
+  # sysctl -w net.ipv4.tcp_max_tw_buckets=1048576
 
   # Enabling following option is not recommended. It could cause connection reset under NAT
   # sysctl -w net.ipv4.tcp_tw_recycle=1
   # sysctl -w net.ipv4.tcp_tw_reuse=1
 
   # Timeout for FIN-WAIT-2 Sockets:
-  sysctl -w net.ipv4.tcp_fin_timeout=15
+  # sysctl -w net.ipv4.tcp_fin_timeout=15
 
   # William's config
-  sysctl -w net.core.rmem_default=262144000
-  sysctl -w net.core.wmem_default=262144000
-  sysctl -w net.core.rmem_max=262144000
-  sysctl -w net.core.wmem_max=262144000
-  sysctl -w net.ipv4.tcp_mem="378150000  504200000  756300000"
+  # sysctl -w net.core.rmem_default=262144000
+  # sysctl -w net.core.wmem_default=262144000
+  # sysctl -w net.core.rmem_max=262144000
+  # sysctl -w net.core.wmem_max=262144000
+  # sysctl -w net.ipv4.tcp_mem="378150000  504200000  756300000"
 }

full-restart.yaml

@@ -0,0 +1,31 @@
+---
+
+- name: define hosts
+  hosts: localhost
+  gather_facts: no
+  roles:
+    - define_inventory
+
+- name: stop emqx and clean logs
+  hosts: emqx
+  become: yes
+  become_user: root
+  roles:
+    - stop_emqx
+    - clean_logs
+
+- name: start cores
+  hosts: cores
+  become: yes
+  become_user: root
+  roles:
+    - start_emqx
+
+- name: start replicants
+  hosts: replicants
+  become: yes
+  become_user: root
+  roles:
+    - start_emqx
+
+...

make-emqx-inv

@@ -1,12 +1,25 @@
 #!/usr/bin/env bb
 ;; -*- mode: clojure; -*-
 
-(let [[bastion-ip num-emqx] *command-line-args*
-      emqxs (map #(str "emqx-" % ".int.thalesmg") (range (Integer/parseInt num-emqx)))]
-  (doseq [emqx emqxs]
-    (println (str "ssh"
-                  " -o StrictHostKeyChecking=no"
-                  " -J ec2-user@"
-                  bastion-ip
-                  " ubuntu@"
-                  emqx))))
+(let [[bastion-ip num-emqx num-cores] *command-line-args*
+      emqxs (map #(str "emqx-" % ".int.thalesmg") (range (Integer/parseInt num-emqx)))
+      make-line (fn [emqx]
+                  (str "ssh"
+                       " -o StrictHostKeyChecking=no"
+                       " -J ec2-user@"
+                       bastion-ip
+                       " ubuntu@"
+                       emqx))
+      cores (->> emqxs
+                 (take (Integer/parseInt num-cores))
+                 (map make-line)
+                 (str/join "\n"))
+      replicants (->> emqxs
+                      (drop (Integer/parseInt num-cores))
+                      (map make-line)
+                      (str/join "\n"))
+      all (str/join "\n" [cores replicants])]
+  (spit "par-hosts.txt" all)
+  (spit "cores.txt" cores)
+  (spit "replicants.txt" replicants)
+  (println all))

roles/clean_logs/tasks/main.yml

@@ -0,0 +1,12 @@
+- name: clear mnesia dir
+  shell: "rm -rf /var/lib/emqx/mnesia/*"
+
+- name: clear emqx logs
+  shell: "rm -rf /var/log/emqx/*"
+
+- name: clear syslog
+  shell: "echo > /var/log/syslog"
+
+- name: clear mem ets dumps
+  shell: "rm /tmp/*_mem-ets-dump.txt"
+  ignore_errors: yes

roles/collect_logs/tasks/main.yml

@@ -0,0 +1,54 @@
+---
+
+- name: produce node_dump
+  shell: "/usr/lib/emqx/bin/node_dump"
+  register: node_dump_invoke
+  ignore_errors: yes
+- name: find node dump location
+  when: not node_dump_invoke.failed
+  set_fact:
+    node_dump_path: "{{ node_dump_invoke.stdout | regex_search(regexp, '\\1') }}"
+  vars:
+    regexp: "Created a node dump (.+)"
+- name: fetch node dump
+  when: not node_dump_invoke.failed and node_dump_path is defined and node_dump_path
+  fetch:
+    src: "{{ node_dump_path[0] }}"
+    dest: "./tmp/data/{{ emqx_script_result_file }}/{{ inventory_hostname_short }}-node_dump.tar.gz"
+    flat: yes
+- name: crashdump?
+  stat:
+    path: "/var/log/emqx/erl_crash.dump"
+  register: crashdump_stat
+- name: syslog?
+  stat:
+    path: "/var/log/syslog"
+  register: syslog_stat
+- name: fetch crashdump
+  when: crashdump_stat.stat.exists
+  fetch:
+    src: "{{ crashdump_stat.stat.path }}"
+    dest: "./tmp/data/{{ emqx_script_result_file }}/{{ inventory_hostname_short }}-crashdump"
+    flat: yes
+- name: fetch syslog
+  when: syslog_stat.stat.exists
+  fetch:
+    src: "{{ syslog_stat.stat.path }}"
+    dest: "./tmp/data/{{ emqx_script_result_file }}/{{ inventory_hostname_short }}-syslog"
+    flat: yes
+- name: find mem ets dumps
+  shell: "ls /tmp/*_mem-ets-dump.txt 2>/dev/null || true"
+  register: mem_ets_dumps
+- name: tar mem ets dumps
+  when: mem_ets_dumps.stdout_lines
+  shell: |
+    cd /tmp
+    tar -cjf {{ emqx_script_result_file }}.mem-ets-dump.tar.bz2 *_mem-ets-dump.txt
+- name: fetch mem ets dumps
+  when: mem_ets_dumps.stdout_lines
+  fetch:
+    src: "/tmp/{{ emqx_script_result_file }}.mem-ets-dump.tar.bz2"
+    dest: "./tmp/data/{{ emqx_script_result_file }}/{{ inventory_hostname_short }}-mem-ets-dump.tar.bz2"
+    flat: yes
+
+...

roles/move_results/tasks/main.yml

@@ -0,0 +1,19 @@
+---
+
+- name: results?
+  stat:
+    path: "./tmp/data/{{ emqx_script_result_file }}"
+  register: results_stat
+- block:
+    - name: copy to destination
+      copy:
+        remote_src: yes
+        src: "{{ results_stat.stat.path }}"
+        dest: "{{ emqx_script_result_dest }}/"
+    - name: remove original
+      file:
+        path: "{{ results_stat.stat.path }}"
+        state: absent
+  when: results_stat.stat.exists and results_stat.stat.isdir and emqx_script_result_dest is defined and emqx_script_result_dest
+
+...

roles/start_emqx/tasks/main.yml

@@ -0,0 +1,2 @@
+- name: start emqx
+  shell: emqx start

roles/stop_emqx/tasks/main.yml

@@ -0,0 +1,19 @@
+- name: stop systemd emqx
+  systemd:
+    name: emqx
+    state: stopped
+    daemon_reload: yes
+
+- name: stop emqx
+  block:
+    - name: stop emqx
+      shell: env EMQX_WAIT_FOR_STOP=20 emqx stop
+      register: stop_result
+      retries: 1
+      delay: 3
+      until: '"badrpc,timeout" not in stop_result.stderr'
+      failed_when: '"badrpc,timeout" in stop_result.stderr'
+  rescue:
+    - name: REALLY stop emqx
+      shell: "pkill -9 beam.smp"
+      ignore_errors: yes

run-test.yml

@@ -19,6 +19,7 @@
         - emqx_bastion_server
         - emqx_test_script_file
         - emqtt_bench_interval
+        - emqtt_bench_pub_interval
         - emqtt_bench_number_of_connections
         - emqtt_bench_session_expiry_interval
         - script_timezone
@@ -61,8 +62,8 @@
         timeout: "{{ script_timeout_s | default('300') }}"
         num_procs: "{{ emqx_test_procs | default(1) | int }}"
         start_n: "{{ (loadgen_num * (emqx_test_procs | default(1) | int)) | int }}"
-    - debug:
-        var: test_runner
+    # - debug:
+    #     var: test_runner
     - name: give the scripts some time...
       pause:
         seconds: "{{ script_timeout_s | default('300') }}"
@@ -73,23 +74,37 @@
       shell: "ps -ef | grep bench | grep -v grep | awk '{print $2}' | xargs kill -9"
       # shell: "pkill beam"
       ignore_errors: yes
-    - name: fetch output
-      tags: [collect]
-      fetch:
-        src: "/tmp/{{ emqx_script_result_file }}"
-        dest: "./tmp/data/{{ emqx_script_result_file }}/{{ inventory_hostname_short }}"
-        flat: yes
+    # - name: fetch output
+    #   tags: [collect]
+    #   fetch:
+    #     src: "/tmp/{{ emqx_script_result_file }}"
+    #     dest: "./tmp/data/{{ emqx_script_result_file }}/{{ inventory_hostname_short }}"
+    #     flat: yes
+
+- name: collect logs
+  hosts: emqx
+  become: yes
+  become_user: root
+  tags: [collect_logs]
+  roles:
+    - collect_logs
 
-- name: test
+- name: move results
   hosts: localhost
-  tags: [collect]
-  tasks:
-    - name: compress
-      archive:
-        path:
-          - './tmp/data/{{ emqx_script_result_file }}/*'
-        dest: "./tmp/data/{{ emqx_script_result_file }}.tar.bz2"
-        remove: yes
-        format: bz2
+  tags: [collect_logs]
+  roles:
+    - move_results
+
+# - name: test
+#   hosts: localhost
+#   tags: [collect]
+#   tasks:
+#     - name: compress
+#       archive:
+#         path:
+#           - './tmp/data/{{ emqx_script_result_file }}/*'
+#         dest: "./tmp/data/{{ emqx_script_result_file }}.tar.bz2"
+#         remove: yes
+#         format: bz2
 
 ...

scripts/cc.py

@@ -93,11 +93,17 @@ def fetch_mem_ets_dump(c : Connection, outdir : Path, prefix : str):
         for med in c.run("ls /tmp/", hide=True).stdout.splitlines()
         if med.endswith("_mem-ets-dump.txt")
     ]
-    for dump in dumps:
-        infile = f"/tmp/{dump}"
-        outfile = f"{prefix}.{c.host}.{dump}"
+    if dumps:
+        outfile = f"{prefix}.{c.host}.mem-ets-dump.tar.bz2"
+        tar_dump = f"/tmp/{outfile}"
+        c.run(f"cd /tmp && tar -cjf {tar_dump} *_mem-ets-dump.txt")
         outfilepath = outdir.joinpath(outfile)
-        c.get(infile, local=str(outfilepath))
+        c.get(tar_dump, local=str(outfilepath))
+    # for dump in dumps:
+    #     infile = f"/tmp/{dump}"
+    #     outfile = f"{prefix}.{c.host}.{dump}"
+    #     outfilepath = outdir.joinpath(outfile)
+    #     c.get(infile, local=str(outfilepath))
 
 
 def fetch_logs(args):
@@ -108,8 +114,12 @@ def fetch_logs(args):
     prefix = args.prefix
 
     for c in inventory_emqx(num_emqx, bastion_ip, cluster_name):
+        print(f"dumping {c}")
         fetch_syslog(c, outdir, prefix)
-        fetch_node_dump(c, outdir, prefix)
+        try:
+            fetch_node_dump(c, outdir, prefix)
+        except Exception as e:
+            print(e)
         fetch_crashdump(c, outdir, prefix)
         fetch_mem_ets_dump(c, outdir, prefix)