- Compass-ci - mailweb.openeuler.org

[PATCH v1 compass-ci] based on queue field add the job into specify queue
by Xiao Shenwei 27 Oct '20

27 Oct '20

use queue replace tbox_group Signed-off-by: Xiao Shenwei <xiaoshenwei96(a)163.com> --- src/lib/sched.cr | 21 ++++++++++++--------- 1 file changed, 12 insertions(+), 9 deletions(-) diff --git a/src/lib/sched.cr b/src/lib/sched.cr index 692180f..d4929cd 100644 --- a/src/lib/sched.cr +++ b/src/lib/sched.cr @@ -30,12 +30,15 @@ class Sched mac.gsub(":", "-") end - def set_host_mac(mac : String, hostname : String) + def set_host_mac(mac : String, hostname : String, queues : String) + puts "queues: #{queues}" @redis.hash_set("sched/mac2host", normalize_mac(mac), hostname) + @redis.hash_set("sched/mac2queues", normalize_mac(mac), queues) end def del_host_mac(mac : String) @redis.hash_del("sched/mac2host", normalize_mac(mac)) + @redis.hash_del("sched/mac2queues", normalize_mac(mac)) end # return: @@ -251,24 +254,24 @@ class Sched # success: [{"job_id" => job_id, "message" => "", job_state => "submit"}] # failure: [{"job_id" => "0", "message" => err_msg, job_state => "submit"}] def submit_single_job(job) - tbox_group = job.tbox_group + queue = job.queue return [{ "job_id" => "0", - "message" => "get tbox group failed", + "message" => "get queue failed", "job_state" => "submit", - }] unless tbox_group + }] unless queue # only single job will has "idle job" and "execute rate limiter" if job["idle_job"].empty? - tbox_group += "#{job.get_uuid_tag}" + queue += "#{job.get_uuid_tag}" else - tbox_group = "#{tbox_group}/idle" + queue = "#{queue}/idle" end - job_id = add_task(tbox_group, job.lab) + job_id = add_task(queue, job.lab) return [{ "job_id" => "0", - "message" => "add task queue sched/#{tbox_group} failed", + "message" => "add task queue sched/#{queue} failed", "job_state" => "submit", }] unless job_id @@ -372,7 +375,7 @@ class Sched hostname = @redis.hash_get("sched/mac2host", normalize_mac(api_param)) if hostname.nil? # auto name new/unknown machine hostname = "sut-#{api_param}" - set_host_mac(api_param, hostname) + set_host_mac(api_param, hostname, "") # auto submit a job to collect the host information # grub hostname is link with ":", like "00:01:02:03:04:05" -- 2.23.0

2 2

[PATCH compass-ci] sched: mount default config dir for idle job
by Cao Xueliang 27 Oct '20

27 Oct '20

[why] There no default config yaml in scheduler container, when submit idle job use the submit script. [how] Mount default config dir in scheduler container. Signed-off-by: Cao Xueliang <caoxl78320(a)163.com> --- container/scheduler/start | 3 +++ src/scheduler/jobfile_operate.cr | 10 +++------- 2 files changed, 6 insertions(+), 7 deletions(-) diff --git a/container/scheduler/start b/container/scheduler/start index d61616a..6a2a4b0 100755 --- a/container/scheduler/start +++ b/container/scheduler/start @@ -45,6 +45,8 @@ SCHED_HOST = defaults['SCHED_HOST'] || '172.17.0.1' FLUENTD_HOST = defaults['FLUENTD_HOST'] || '172.17.0.1' FLUENTD_PORT = defaults['FLUENTD_PORT'] || '24224' FLUENTD_SERVER_PORT = defaults['FLUENTD_SERVER_PORT'] || '24224' +DEFAULT_CONFIG_DIR = '/etc/compass-ci/defaults' + docker_rm "s001-alpine-#{SCHED_PORT}" bind_list = [] @@ -70,6 +72,7 @@ cmd = %W[ -e FLUENTD_SERVER_PORT=#{FLUENTD_SERVER_PORT} -v #{ENV['LKP_SRC']}:#{DEFAULT_LKP} -v #{ENV['CCI_SRC']}:#{DEFAULT_CCI} + -v #{DEFAULT_CONFIG_DIR}:#{DEFAULT_CONFIG_DIR}:ro -v #{CCI_REPOS}/lab-#{LAB}:#{CCI_REPOS}/lab-#{LAB} -v /srv/cci/scheduler/alpine:/srv/cci/scheduler -v /etc/localtime:/etc/localtime:ro diff --git a/src/scheduler/jobfile_operate.cr b/src/scheduler/jobfile_operate.cr index e0e298f..d3c26c5 100644 --- a/src/scheduler/jobfile_operate.cr +++ b/src/scheduler/jobfile_operate.cr @@ -202,14 +202,10 @@ module Jobfile::Operate end def self.auto_submit_job(job_file, override_parameter, other_parameters = nil) - job_fields = ["SCHED_HOST=#{ENV["SCHED_HOST"]}", - "SCHED_PORT=#{ENV["SCHED_PORT"]}", - "GIT_SERVER=#{ENV["GIT_SERVER"]}", - "lab=#{ENV["lab"]}"] - - job_fields.concat(other_parameters) if other_parameters cmd = "#{ENV["LKP_SRC"]}/sbin/submit " - cmd += job_fields.join(" ") + if other_parameters + cmd += other_parameters.join(" ") + end cmd += " -s '#{override_parameter}' #{job_file}" puts `#{cmd}` end -- 2.23.0

1 0

[PATCH v3 compass-ci] add queue field into job.yaml
by Xiao Shenwei 27 Oct '20

27 Oct '20

[why] after one job submit, it will be add specified queue before: use tbox_group to discriminate now : use queue field to discriminate so, we should add this field into job.yaml it's come from user's specify(job.yaml) or dafault Signed-off-by: Xiao Shenwei <xiaoshenwei96(a)163.com> --- src/lib/job.cr | 12 ++++++++++++ 1 file changed, 12 insertions(+) diff --git a/src/lib/job.cr b/src/lib/job.cr index 8f5dd36..0376f9f 100644 --- a/src/lib/job.cr +++ b/src/lib/job.cr @@ -73,6 +73,7 @@ class Job tbox_group testbox lab + queue initrd_pkg initrd_deps initrds_uri @@ -139,6 +140,7 @@ class Job set_kernel_params() set_lkp_server() set_sshr_port() + set_queue() end private def append_init_field @@ -199,6 +201,16 @@ class Job self["result_service"] = "raw_upload" end + private def set_queue + return if @hash.has_key?("queue") + + # set default value + self["queue"] = tbox_group + if tbox_group.to_s.starts_with?(/(vm|dc).*/) + self["queue"] = "#{tbox_group}.#{arch}" + end + end + # if not assign tbox_group, set it to a match result from testbox # ?if job special testbox, should we just set tbox_group=textbox private def update_tbox_group_from_testbox -- 2.23.0

3 4

[PATCH compass-ci] fix: taskqueue api: use same client for 1 server
by Tong Qunfeng 27 Oct '20

27 Oct '20

[why] interaction to taskqueue will be frequently, so keep same client as possible. - btw: caller, that do not need keep-alive with taskqueue, should set header["Connection"]="close", or active close the connection. Signed-off-by: Tong Qunfeng <taxcom(a)tom.com> --- src/lib/taskqueue_api.cr | 28 +++++++++++++++++----------- 1 file changed, 17 insertions(+), 11 deletions(-) diff --git a/src/lib/taskqueue_api.cr b/src/lib/taskqueue_api.cr index 5a83a15..a8bbf16 100644 --- a/src/lib/taskqueue_api.cr +++ b/src/lib/taskqueue_api.cr @@ -4,25 +4,33 @@ require "json" require "http/client" +def loop_till_connectable + loop do + begin + return yield + rescue ex + puts "#{ex}, retrying in 5s..." + sleep(5) + end + end +end + class TaskQueueAPI def initialize - @port = ENV.has_key?("TASKQUEUE_PORT") ? ENV["TASKQUEUE_PORT"].to_i32 : 3060 - @host = ENV.has_key?("TASKQUEUE_HOST") ? ENV["TASKQUEUE_HOST"] : "172.17.0.1" + port = ENV.has_key?("TASKQUEUE_PORT") ? ENV["TASKQUEUE_PORT"].to_i32 : 3060 + host = ENV.has_key?("TASKQUEUE_HOST") ? ENV["TASKQUEUE_HOST"] : "172.17.0.1" + @client = HTTP::Client.new(host, port: port) end def add_task(service_queue_path : String, task : JSON::Any) params = HTTP::Params.encode({"queue" => service_queue_path}) - client = HTTP::Client.new(@host, port: @port) - response = client.post("/add?" + params, body: task.to_json) - client.close + response = loop_till_connectable { @client.post("/add?" + params, body: task.to_json) } arrange_response(response) end def query_keys(service_key_with_wild_char : String) params = HTTP::Params.encode({"queue" => service_key_with_wild_char}) - client = HTTP::Client.new(@host, port: @port) - response = client.get("/keys?" + params) - client.close + response = loop_till_connectable { @client.get("/keys?" + params) } arrange_response(response) end @@ -42,9 +50,7 @@ class TaskQueueAPI end private def response_put_api(cmd : String, params : String) - client = HTTP::Client.new(@host, port: @port) - response = client.put("/#{cmd}?" + params) - client.close + response = loop_till_connectable { @client.put("/#{cmd}?" + params) } arrange_response(response) end -- 2.23.0

1 0

[PATCH compass-ci] fix: scheduler: close connection after API call
by Tong Qunfeng 27 Oct '20

27 Oct '20

[why] most client call to scheduler do not need keep-alive. when API call is finished, close the connection. [how] set response header["Connection"]="close" Signed-off-by: Tong Qunfeng <taxcom(a)tom.com> --- src/scheduler/scheduler.cr | 12 ++++++++++++ 1 file changed, 12 insertions(+) diff --git a/src/scheduler/scheduler.cr b/src/scheduler/scheduler.cr index 5befcf9..371e4ed 100644 --- a/src/scheduler/scheduler.cr +++ b/src/scheduler/scheduler.cr @@ -39,6 +39,7 @@ module Scheduler # echo alive get "/" do |env| + env.response.headers["Connection"] = "close" debug_message(env, "Env= {\n#{`export`}}") "LKP Alive! The time is #{Time.local}, version = #{VERSION}" end @@ -49,6 +50,7 @@ module Scheduler # /boot.xxx/host/${hostname} # /boot.yyy/mac/${mac} get "/boot.:boot_type/:parameter/:value" do |env| + env.response.headers["Connection"] = "close" response = sched.find_job_boot(env) job_id = response[/tmpfs\/(.*)\/job\.cgz/, 1]? @@ -59,6 +61,7 @@ module Scheduler # /~lkp/cgi-bin/gpxelinux.cgi?hostname=:hostname&mac=:mac&last_kernel=:last_kernel get "/~lkp/cgi-bin/gpxelinux.cgi" do |env| + env.response.headers["Connection"] = "close" response = sched.find_next_job_boot(env) job_id = response[/tmpfs\/(.*)\/job\.cgz/, 1]? @@ -71,6 +74,7 @@ module Scheduler # - echo job_id to caller # -- job_id = "0" ? means failed post "/submit_job" do |env| + env.response.headers["Connection"] = "close" job_messages = sched.submit_job(env) job_messages.each do |job_message| @@ -82,6 +86,7 @@ module Scheduler # file download server get "/job_initrd_tmpfs/:job_id/:job_package" do |env| + env.response.headers["Connection"] = "close" job_id = env.params.url["job_id"] job_package = env.params.url["job_package"] file_path = ::File.join [Kemal.config.public_folder, job_id, job_package] @@ -98,6 +103,7 @@ module Scheduler # # curl -X PUT "http://localhost:3000/set_host_mac?hostname=wfg&mac=00-01-02-03-04-05" put "/set_host_mac" do |env| + env.response.headers["Connection"] = "close" if (client_hostname = env.params.query["hostname"]?) && (client_mac = env.params.query["mac"]?) sched.set_host_mac(client_mac, client_hostname) @@ -109,6 +115,7 @@ module Scheduler # curl -X PUT "http://localhost:3000/del_host_mac?mac=00-01-02-03-04-05" put "/del_host_mac" do |env| + env.response.headers["Connection"] = "close" if client_mac = env.params.query["mac"]? sched.del_host_mac(client_mac) @@ -124,6 +131,7 @@ module Scheduler # ?job_file=/lkp/scheduled/job.yaml&job_state=post_run&job_id=10 # ?job_file=/lkp/scheduled/job.yaml&loadavg=0.28 0.82 0.49 1/105 3389&start_time=1587725398&end_time=1587725698&job_id=10 get "/~lkp/cgi-bin/lkp-jobfile-append-var" do |env| + env.response.headers["Connection"] = "close" puts sched.update_job_parameter(env) "Done" @@ -147,6 +155,7 @@ module Scheduler # response: get "server ip" from cluster state, # return "server=<server ip>". get "/~lkp/cgi-bin/lkp-cluster-sync" do |env| + env.response.headers["Connection"] = "close" response = sched.request_cluster_state(env) debug_message(env, response) @@ -158,6 +167,7 @@ module Scheduler # /~lkp/cgi-bin/lkp-post-run?job_file=/lkp/scheduled/job.yaml&job_id=40 # curl "http://localhost:3000/~lkp/cgi-bin/lkp-post-run?job_file=/lkp/scheduled/job…" get "/~lkp/cgi-bin/lkp-post-run" do |env| + env.response.headers["Connection"] = "close" # get job_id from request job_id = env.params.query["job_id"]? if job_id @@ -168,12 +178,14 @@ module Scheduler end get "/~lkp/cgi-bin/lkp-wtmp" do |env| + env.response.headers["Connection"] = "close" puts sched.update_tbox_wtmp(env) "Done" end get "/~lkp/cgi-bin/report_ssh_port" do |env| + env.response.headers["Connection"] = "close" testbox = env.params.query["tbox_name"] ssh_port = env.params.query["ssh_port"].to_s job_id = env.params.query["job_id"].to_s -- 2.23.0

1 0

[PATCH v2 compass-ci] /lib/es_query.rb: add range query
by Zhang Yuhang 27 Oct '20

27 Oct '20

[why] Some queries need to limit the range of fields. Signed-off-by: Zhang Yuhang <zhangyuhang25(a)huawei.com> --- lib/es_query.rb | 15 +++++++++++++++ 1 file changed, 15 insertions(+) diff --git a/lib/es_query.rb b/lib/es_query.rb index aaaadc2..037213d 100644 --- a/lib/es_query.rb +++ b/lib/es_query.rb @@ -36,12 +36,27 @@ class ESQuery end end +# Range Query Example: +# range = { +# start_time: { +# gte: '2020-09-10 01:50:00', +# lte: '2020-09-10 01:53:00' +# }, +# end_time: { +# gt: '2020-09-10 01:52:00', +# lt: '2020-09-10 01:54:00' +# } +# } +# items['range'] = range +# build_mutli_field_subquery_body(items) def build_mutli_field_subquery_body(items) query_fields = [] items.each do |key, value| if value.is_a?(Array) inner_query = build_multi_field_or_query_body(key, value) query_fields.push({ bool: { should: inner_query } }) + elsif key.to_s == 'range' + query_fields.concat(value.map { |k, v| { range: { k => v } } }) else query_fields.push({ term: { key => value } }) end -- 2.23.0

2 2

[PATCH lab-z9 3/3] rename 3 files for memory change to 512G
by Zhang Dewan 27 Oct '20

27 Oct '20

--- ...200-2280-2s48p-256g--a102 => taishan200-2280-2s48p-512g--a102} | 0 ...200-2280-2s48p-256g--a105 => taishan200-2280-2s48p-512g--a105} | 0 ...an200-2280-2s48p-256g--a95 => taishan200-2280-2s48p-512g--a95} | 0 3 files changed, 0 insertions(+), 0 deletions(-) rename hosts/{taishan200-2280-2s48p-256g--a102 => taishan200-2280-2s48p-512g--a102} (100%) rename hosts/{taishan200-2280-2s48p-256g--a105 => taishan200-2280-2s48p-512g--a105} (100%) rename hosts/{taishan200-2280-2s48p-256g--a95 => taishan200-2280-2s48p-512g--a95} (100%) diff --git a/hosts/taishan200-2280-2s48p-256g--a102 b/hosts/taishan200-2280-2s48p-512g--a102 similarity index 100% rename from hosts/taishan200-2280-2s48p-256g--a102 rename to hosts/taishan200-2280-2s48p-512g--a102 diff --git a/hosts/taishan200-2280-2s48p-256g--a105 b/hosts/taishan200-2280-2s48p-512g--a105 similarity index 100% rename from hosts/taishan200-2280-2s48p-256g--a105 rename to hosts/taishan200-2280-2s48p-512g--a105 diff --git a/hosts/taishan200-2280-2s48p-256g--a95 b/hosts/taishan200-2280-2s48p-512g--a95 similarity index 100% rename from hosts/taishan200-2280-2s48p-256g--a95 rename to hosts/taishan200-2280-2s48p-512g--a95 -- 2.23.0

1 0

[PATCH lab-z9 2/3] add new host file
by Zhang Dewan 27 Oct '20

27 Oct '20

--- hosts/m | 43 ++------------------------ hosts/taishan200-2280-2s48p-256g--a102 | 1 - hosts/taishan200-2280-2s48p-256g--a103 | 24 ++++++++++++++ hosts/taishan200-2280-2s48p-256g--a104 | 23 ++++++++++++++ hosts/taishan200-2280-2s48p-256g--a105 | 24 ++++++++++++++ hosts/taishan200-2280-2s48p-256g--a106 | 24 ++++++++++++++ hosts/taishan200-2280-2s48p-256g--a107 | 24 ++++++++++++++ hosts/taishan200-2280-2s48p-256g--a16 | 20 ++++++++++++ hosts/taishan200-2280-2s48p-256g--a35 | 21 +++++++++++++ hosts/taishan200-2280-2s48p-256g--a63 | 19 ++++++++++++ hosts/taishan200-2280-2s48p-256g--a87 | 20 ++++++++++++ hosts/taishan200-2280-2s48p-256g--a95 | 1 - 12 files changed, 202 insertions(+), 42 deletions(-) create mode 100644 hosts/taishan200-2280-2s48p-256g--a103 create mode 100644 hosts/taishan200-2280-2s48p-256g--a104 create mode 100644 hosts/taishan200-2280-2s48p-256g--a105 create mode 100644 hosts/taishan200-2280-2s48p-256g--a106 create mode 100644 hosts/taishan200-2280-2s48p-256g--a107 create mode 100644 hosts/taishan200-2280-2s48p-256g--a16 create mode 100644 hosts/taishan200-2280-2s48p-256g--a35 create mode 100644 hosts/taishan200-2280-2s48p-256g--a63 create mode 100644 hosts/taishan200-2280-2s48p-256g--a87 diff --git a/hosts/m b/hosts/m index b73dfb7..73ba010 100644 --- a/hosts/m +++ b/hosts/m @@ -1,40 +1,3 @@ -taishan200-2280-2s48p-256g--a11 a11 -taishan200-2280-2s48p-256g--a13 a13 -taishan200-2280-2s48p-256g--a14 a14 -taishan200-2280-2s48p-256g--a15 a15 -taishan200-2280-2s48p-256g--a17 a17 -taishan200-2280-2s48p-256g--a18 a18 -taishan200-2280-2s48p-256g--a19 a19 -taishan200-2280-2s48p-256g--a21 a21 -taishan200-2280-2s48p-256g--a22 a22 -taishan200-2280-2s48p-256g--a24 a24 -taishan200-2280-2s48p-256g--a25 a25 -taishan200-2280-2s48p-256g--a26 a26 -taishan200-2280-2s48p-256g--a27 a27 -taishan200-2280-2s48p-256g--a28 a28 -taishan200-2280-2s48p-256g--a29 a29 -taishan200-2280-2s48p-256g--a30 a30 -taishan200-2280-2s48p-256g--a31 a31 -taishan200-2280-2s48p-256g--a32 a32 -taishan200-2280-2s48p-256g--a33 a33 -taishan200-2280-2s48p-256g--a34 a34 -taishan200-2280-2s48p-256g--a52 a52 -taishan200-2280-2s48p-256g--a53 a53 -taishan200-2280-2s48p-256g--a60 a60 -taishan200-2280-2s48p-256g--a61 a61 -taishan200-2280-2s48p-256g--a62 a62 -taishan200-2280-2s48p-256g--a66 a66 -taishan200-2280-2s48p-256g--a67 a67 -taishan200-2280-2s48p-256g--a68 a68 -taishan200-2280-2s48p-256g--a69 a69 -taishan200-2280-2s48p-256g--a70 a70 -taishan200-2280-2s48p-256g--a71 a71 -taishan200-2280-2s64p-256g--a10 a10 -taishan200-2280-2s64p-256g--a3 a3 -taishan200-2280-2s64p-256g--a36 a36 -taishan200-2280-2s64p-256g--a4 a4 -taishan200-2280-2s64p-256g--a40 a40 -taishan200-2280-2s64p-256g--a41 a41 -taishan200-2280-2s64p-256g--a42 a42 -taishan200-2280-2s64p-256g--a8 a8 -taishan200-2280-2s64p-256g--a9 a9 +taishan200-2280-2s48p-256g--a16 a16 +taishan200-2280-2s48p-256g--a35 a35 +taishan200-2280-2s48p-256g--a87 a87 diff --git a/hosts/taishan200-2280-2s48p-256g--a102 b/hosts/taishan200-2280-2s48p-256g--a102 index e1d3f10..267a962 100644 --- a/hosts/taishan200-2280-2s48p-256g--a102 +++ b/hosts/taishan200-2280-2s48p-256g--a102 @@ -1,4 +1,3 @@ -arch: aarch64 nr_node: 4 nr_cpu: 96 memory: 512G diff --git a/hosts/taishan200-2280-2s48p-256g--a103 b/hosts/taishan200-2280-2s48p-256g--a103 new file mode 100644 index 0000000..4afef3c --- /dev/null +++ b/hosts/taishan200-2280-2s48p-256g--a103 @@ -0,0 +1,24 @@ +nr_node: 4 +nr_cpu: 96 +memory: 256G +nr_hdd_partitions: 4 +hdd_partitions: + - /dev/disk/by-id/scsi-350000399b88bdc69 + - /dev/disk/by-id/scsi-350000399b88bdd85 + - /dev/disk/by-id/scsi-350000399b88bdc4d + - /dev/disk/by-id/scsi-350000399b88bdc7d +mac_addr: + - 44:67:47:c9:d8:b8 + - 44:67:47:c9:d8:b9 + - 44:67:47:c9:d8:ba + - 44:67:47:c9:d8:bb + - 44:67:47:c9:d8:bc + - 44:67:47:c9:d8:bd + - 44:67:47:c9:d8:be + - 44:67:47:c9:d8:bf + - cc:05:77:ee:b5:4f + - cc:05:77:ee:b5:50 + - cc:05:77:ee:b5:51 + - cc:05:77:ee:b5:52 +model_name: Kunpeng-920 +ipmi_ip: 9.3.10.8 diff --git a/hosts/taishan200-2280-2s48p-256g--a104 b/hosts/taishan200-2280-2s48p-256g--a104 new file mode 100644 index 0000000..3c5b807 --- /dev/null +++ b/hosts/taishan200-2280-2s48p-256g--a104 @@ -0,0 +1,23 @@ +nr_node: 4 +nr_cpu: 96 +memory: 256G +nr_hdd_partitions: 4 +hdd_partitions: + - /dev/disk/by-id/scsi-350000399b8892431 + - /dev/disk/by-id/scsi-350000399b8892f81 + - /dev/disk/by-id/scsi-350000399b8891c49 + - /dev/disk/by-id/scsi-350000399b88930d1 +mac_addr: + - 44:67:47:d7:6d:f0 + - 44:67:47:d7:6d:f1 + - 44:67:47:d7:6d:f2 + - 44:67:47:d7:6d:f3 + - 44:67:47:d7:6d:f4 + - 44:67:47:d7:6d:f5 + - 44:67:47:d7:6d:f6 + - 44:67:47:d7:6d:f7 + - cc:05:77:ee:b9:6f + - cc:05:77:ee:b9:70 + - cc:05:77:ee:b9:71 + - cc:05:77:ee:b9:72 +model_name: Kunpeng-920 diff --git a/hosts/taishan200-2280-2s48p-256g--a105 b/hosts/taishan200-2280-2s48p-256g--a105 new file mode 100644 index 0000000..3c28d2d --- /dev/null +++ b/hosts/taishan200-2280-2s48p-256g--a105 @@ -0,0 +1,24 @@ +nr_node: 4 +nr_cpu: 96 +memory: 512G +nr_hdd_partitions: 4 +hdd_partitions: + - /dev/disk/by-id/scsi-350000399b891945d + - /dev/disk/by-id/scsi-350000399b8919135 + - /dev/disk/by-id/scsi-350000399b8919699 + - /dev/disk/by-id/scsi-350000399b89190d5 +mac_addr: + - 44:67:47:d7:76:c4 + - 44:67:47:d7:76:c5 + - 44:67:47:d7:76:c6 + - 44:67:47:d7:76:c7 + - 44:67:47:d7:76:c8 + - 44:67:47:d7:76:c9 + - 44:67:47:d7:76:ca + - 44:67:47:d7:76:cb + - cc:05:77:ee:b2:8f + - cc:05:77:ee:b2:90 + - cc:05:77:ee:b2:91 + - cc:05:77:ee:b2:92 +model_name: Kunpeng-920 +ipmi_ip: 9.3.10.10 diff --git a/hosts/taishan200-2280-2s48p-256g--a106 b/hosts/taishan200-2280-2s48p-256g--a106 new file mode 100644 index 0000000..a7b44a0 --- /dev/null +++ b/hosts/taishan200-2280-2s48p-256g--a106 @@ -0,0 +1,24 @@ +nr_node: 4 +nr_cpu: 96 +memory: 256G +nr_hdd_partitions: 4 +hdd_partitions: + - /dev/disk/by-id/scsi-350000399b883bf99 + - /dev/disk/by-id/scsi-350000399b883ba89 + - /dev/disk/by-id/scsi-350000399b883b94d + - /dev/disk/by-id/scsi-350000399b883b981 +mac_addr: + - 44:67:47:86:1c:30 + - 44:67:47:86:1c:31 + - 44:67:47:86:1c:32 + - 44:67:47:86:1c:33 + - 44:67:47:86:1c:34 + - 44:67:47:86:1c:35 + - 44:67:47:86:1c:36 + - 44:67:47:86:1c:37 + - cc:05:77:ee:b7:df + - cc:05:77:ee:b7:e0 + - cc:05:77:ee:b7:e1 + - cc:05:77:ee:b7:e2 +model_name: Kunpeng-920 +ipmi_ip: 9.3.10.11 diff --git a/hosts/taishan200-2280-2s48p-256g--a107 b/hosts/taishan200-2280-2s48p-256g--a107 new file mode 100644 index 0000000..9d20e57 --- /dev/null +++ b/hosts/taishan200-2280-2s48p-256g--a107 @@ -0,0 +1,24 @@ +nr_node: 4 +nr_cpu: 96 +memory: 256G +nr_hdd_partitions: 4 +hdd_partitions: + - /dev/disk/by-id/scsi-350000399b8917c39 + - /dev/disk/by-id/scsi-350000399b8918005 + - /dev/disk/by-id/scsi-350000399b8900a4d + - /dev/disk/by-id/scsi-350000399b8900799 +mac_addr: + - 84:46:fe:73:a2:71 + - 84:46:fe:73:a2:72 + - 84:46:fe:73:a2:73 + - 84:46:fe:73:a2:74 + - 84:46:fe:73:a2:75 + - 84:46:fe:73:a2:76 + - 84:46:fe:73:a2:77 + - 84:46:fe:73:a2:78 + - 44:a1:91:35:1a:70 + - 44:a1:91:35:1a:71 + - 44:a1:91:35:1a:72 + - 44:a1:91:35:1a:73 +model_name: Kunpeng-920 +ipmi_ip: 9.3.10.12 diff --git a/hosts/taishan200-2280-2s48p-256g--a16 b/hosts/taishan200-2280-2s48p-256g--a16 new file mode 100644 index 0000000..62cb124 --- /dev/null +++ b/hosts/taishan200-2280-2s48p-256g--a16 @@ -0,0 +1,20 @@ +nr_node: 4 +nr_cpu: 96 +memory: 256G +nr_hdd_partitions: 4 +hdd_partitions: + - /dev/disk/by-id/scsi-350000399b88bf271 + - /dev/disk/by-id/scsi-350000399b88bf1a9 + - /dev/disk/by-id/scsi-350000399b88bf0d5 + - /dev/disk/by-id/scsi-350000399b88bf385 +mac_addr: + - 84:46:fe:3d:a2:17 + - 84:46:fe:3d:a2:18 + - 84:46:fe:3d:a2:19 + - 84:46:fe:3d:a2:1a + - 84:46:fe:3d:a2:1b + - 84:46:fe:3d:a2:1c + - 84:46:fe:3d:a2:1d + - 84:46:fe:3d:a2:1e +model_name: Kunpeng-920 +ipmi_ip: 9.3.5.6 diff --git a/hosts/taishan200-2280-2s48p-256g--a35 b/hosts/taishan200-2280-2s48p-256g--a35 new file mode 100644 index 0000000..ec1e524 --- /dev/null +++ b/hosts/taishan200-2280-2s48p-256g--a35 @@ -0,0 +1,21 @@ +nr_node: 4 +nr_cpu: 96 +memory: 256G +nr_hdd_partitions: 3 +hdd_partitions: + - /dev/disk/by-id/scsi-350000399886a71a5 + - /dev/disk/by-id/scsi-350000399886b66dd-part1 + - /dev/disk/by-id/scsi-350000399886b66dd-part2 + - /dev/disk/by-id/scsi-350000399886b66dd-part3 + - /dev/disk/by-id/scsi-362c97b1c6eedb00026e527a2b8b998f2 +mac_addr: + - 44:67:47:c9:ec:ec + - 44:67:47:c9:ec:ed + - 44:67:47:c9:ec:ee + - 44:67:47:c9:ec:ef + - 44:67:47:c9:ec:f0 + - 44:67:47:c9:ec:f1 + - 44:67:47:c9:ec:f2 + - 44:67:47:c9:ec:f3 +model_name: Kunpeng-920 +ipmi_ip: 9.3.9.12 diff --git a/hosts/taishan200-2280-2s48p-256g--a63 b/hosts/taishan200-2280-2s48p-256g--a63 new file mode 100644 index 0000000..6d4b636 --- /dev/null +++ b/hosts/taishan200-2280-2s48p-256g--a63 @@ -0,0 +1,19 @@ +nr_node: 4 +nr_cpu: 96 +memory: 256G +nr_hdd_partitions: 4 +hdd_partitions: + - /dev/disk/by-id/scsi-350000399886b0431 + - /dev/disk/by-id/scsi-350000399886adc71 + - /dev/disk/by-id/scsi-350000399886b66f5 + - /dev/disk/by-id/scsi-350000399886b6b75 +mac_addr: + - 44:67:47:c9:e5:44 + - 44:67:47:c9:e5:45 + - 44:67:47:c9:e5:46 + - 44:67:47:c9:e5:47 + - 44:67:47:c9:e5:48 + - 44:67:47:c9:e5:49 + - 44:67:47:c9:e5:4a + - 44:67:47:c9:e5:4b +model_name: Kunpeng-920 diff --git a/hosts/taishan200-2280-2s48p-256g--a87 b/hosts/taishan200-2280-2s48p-256g--a87 new file mode 100644 index 0000000..b87417b --- /dev/null +++ b/hosts/taishan200-2280-2s48p-256g--a87 @@ -0,0 +1,20 @@ +nr_node: 4 +nr_cpu: 96 +memory: 256G +nr_hdd_partitions: 4 +hdd_partitions: + - /dev/disk/by-id/scsi-350000399886a6b55 + - /dev/disk/by-id/scsi-350000399886ad4a5 + - /dev/disk/by-id/scsi-350000399886a7139 + - /dev/disk/by-id/scsi-350000399886acb21 +mac_addr: + - 44:67:47:c9:eb:66 + - 44:67:47:c9:eb:67 + - 44:67:47:c9:eb:68 + - 44:67:47:c9:eb:69 + - 44:67:47:c9:eb:6a + - 44:67:47:c9:eb:6b + - 44:67:47:c9:eb:6c + - 44:67:47:c9:eb:6d +model_name: Kunpeng-920 +ipmi_ip: 9.3.8.4 diff --git a/hosts/taishan200-2280-2s48p-256g--a95 b/hosts/taishan200-2280-2s48p-256g--a95 index 213d10f..3fd8124 100644 --- a/hosts/taishan200-2280-2s48p-256g--a95 +++ b/hosts/taishan200-2280-2s48p-256g--a95 @@ -1,4 +1,3 @@ -arch: aarch64 nr_node: 4 nr_cpu: 96 memory: 512G -- 2.23.0

1 0

[PATCH compass-ci] feat: scheduler: comsume exists queue
by Tong Qunfeng 27 Oct '20

27 Oct '20

[why] when there has no job for a testbox, scheduler will consume 4-5 times testbox*1 tbox_group*1 tbox*1 tbox_group/idle*(1|2) [how] query taskqueue the queue-name first. if the queue-name not exist, then there should be no job for this queue. submit idle job at background, let the consume not wait for submit idle job finished. Signed-off-by: Tong Qunfeng <taxcom(a)tom.com> --- src/lib/sched.cr | 52 +++++++++++++++++++++++++++++++++++++++++++----- 1 file changed, 47 insertions(+), 5 deletions(-) diff --git a/src/lib/sched.cr b/src/lib/sched.cr index 7215df5..8fbe9aa 100644 --- a/src/lib/sched.cr +++ b/src/lib/sched.cr @@ -409,19 +409,26 @@ class Sched tbox_group = JobHelper.match_tbox_group(testbox) tbox = tbox_group.partition("--")[0] - boxes = [testbox, tbox_group, tbox] + queue_list = query_consumable_keys(tbox) + + boxes = ["sched/" + testbox, + "sched/" + tbox_group, + "sched/" + tbox, + "sched/" + tbox_group + "/idle"] boxes.each do |box| + next if queue_list.select(box).size == 0 count.times do - job = prepare_job("sched/#{box}", testbox) + job = prepare_job(box, testbox) return job if job sleep(1) unless count == 1 end end - # when find no job at "sched/#{tbox_group}" - # try to get from "sched/#{tbox_group}/idle" - return get_idle_job(tbox_group, testbox) + # when find no job, auto submit idle job at background + spawn { auto_submit_idle_job(tbox_group) } + + return nil end private def prepare_job(queue_name, testbox) @@ -571,4 +578,39 @@ class Sched return %({"job_id": "#{job_id}", "job_state": "complete"}) end + + private def query_consumable_keys(shortest_queue_name) + keys = [] of String + search = "sched/" + shortest_queue_name + "*" + response = @task_queue.query_keys(search) + + return keys unless response[0] == 200 + + key_list = JSON.parse(response[1].to_json).as_a + + # add consumable keys + key_list.each do |key| + queue_name = consumable_key?("#{key}") + keys << queue_name if queue_name + end + + return keys + end + + private def consumable_key?(key_name) + if key_name =~ /(.*)\/(.*)$/ + case $2 + when "in_process" + return nil + when "ready" + return $1 + when "idle" + return key_name + else + return key_name + end + end + + return nil + end end -- 2.23.0

4 6

[PATCH compass-ci] fix: container: avoid bind to all ip
by Tong Qunfeng 27 Oct '20

27 Oct '20

[why] default docker "-p port:port" will bind to all ips. - "-p 0.0.0.0:port:port" will also bind to all ips. - if exists ipv6 ip, it will bind to ipv6 address too. - if bind to ipv6 address, there will be tcp6 socket in TIME_WAIT no need bind to all ips. Signed-off-by: Tong Qunfeng <taxcom(a)tom.com> --- container/scheduler/start | 9 +++++++-- container/taskqueue/start | 9 ++++++++- 2 files changed, 15 insertions(+), 3 deletions(-) diff --git a/container/scheduler/start b/container/scheduler/start index 305916e..d61616a 100755 --- a/container/scheduler/start +++ b/container/scheduler/start @@ -47,14 +47,19 @@ FLUENTD_PORT = defaults['FLUENTD_PORT'] || '24224' FLUENTD_SERVER_PORT = defaults['FLUENTD_SERVER_PORT'] || '24224' docker_rm "s001-alpine-#{SCHED_PORT}" +bind_list = [] +[SCHED_HOST, '172.17.0.1'].uniq.each do |ip| + bind_list << '-p' + bind_list << "#{ip}:#{SCHED_PORT}:#{SCHED_PORT}" +end + cmd = %W[ docker run --restart=always --name s001-alpine-#{SCHED_PORT} -d -u 1090:1090 - -p #{SCHED_PORT}:#{SCHED_PORT} -] + env + %W[ +] + bind_list + env + %W[ -e CCI_REPOS=#{CCI_REPOS} -e LKP_SRC=#{DEFAULT_LKP} -e CCI_SRC=#{DEFAULT_CCI} diff --git a/container/taskqueue/start b/container/taskqueue/start index 4760b7d..bdd5e7c 100755 --- a/container/taskqueue/start +++ b/container/taskqueue/start @@ -16,19 +16,26 @@ defaults = relevant_defaults(names) env = docker_env(defaults) service_port = defaults['TASKQUEUE_PORT'] || '3060' +service_host = defaults['TASKQUEUE_HOST'] || '172.17.0.1' service_name = 'taskqueue' service_image = "#{service_name}-alpine:v0.001-#{service_port}" docker_rm "#{service_name}-#{service_port}" +bind_list = [] +[service_host, '172.17.0.1'].uniq.each do |ip| + bind_list << '-p' + bind_list << "#{ip}:#{service_port}:#{service_port}" +end + cmd = %W[ docker run --name #{service_name}-#{service_port} --restart=always -d -u 1090:1090 - -p #{service_port}:#{service_port} ] +cmd += bind_list cmd += env cmd += %W[ -v /etc/localtime:/etc/localtime:ro -- 2.23.0

1 0