[Compass-ci] Re: [PATCH compass-ci] fix: sched: cluster job needs more fields

About this ,i'll write a todo.

Thanks, RenWen

On Fri, Oct 23, 2020 at 10:30:44AM +0800, Zhang Yu wrote:

On Thu, Oct 22, 2020 at 10:15:23PM +0800, Ren Wen wrote:

...

[Why] Nodes need to communicate with each other when testing cluster jobs. Ips can bind to on macs after setting 'direct_mac' and 'direct_ips' for cluster jobs

[HowTo]

1. Need a cluster config file in $LKP_SRC/cluster,

for example:

assumption: <assumption01>
switch: switch01
ip0: 3

---

vm-2p8g--renwen:
  roles: [ server ]
  macs: [ "ec:f4:bb:cb:7b:92" ]

vm-2p4g--renwen:
  roles: [ client ]
  macs: [ "ec:f4:bb:cb:54:90" ]

nr_node=2 cs1 switch=sw1 ip0=0 ($ip_prefix.0-.1) cs2 switch=sw1 ip0=2 ($ip_prefix.2-.3)

This means cs1 or cs2 both have ip0, and this ip0 means the physical machine's last mask scope? if switch same they are use same ip_prefix.

...

"switch": only need if 2+ clusters share the same switch "ip0": only need if shared switch, and != 0

The ip0 field is mandatory not only if shared switch, it's the cluster machine's ip scope.

...

"assumption": the info about this cluster config

How to write this "assumption"? may be can give a exaple.

...

2. Scheduler get config from config file and set 'direct_macs' and

'direct_ips' for jobs.

3. When running a cluster job, bind direct_ips to direct_ macs

in test environment.

4. Test client connect with daemon client to test job.

                                   ^ server

Thanks, Zhangyu

...

Signed-off-by: Ren Wen 15991987063@163.com

src/lib/sched.cr | 19 ++++++++++++++++--- 1 file changed, 16 insertions(+), 3 deletions(-)

diff --git a/src/lib/sched.cr b/src/lib/sched.cr index 070a16b..8922874 100644 --- a/src/lib/sched.cr +++ b/src/lib/sched.cr @@ -144,7 +144,7 @@ class Sched def get_cluster_config(cluster_file, lkp_initrd_user, os_arch) lkp_src = Jobfile::Operate.prepare_lkp_tests(lkp_initrd_user, os_arch) cluster_file_path = Path.new(lkp_src, "cluster", cluster_file)

• return YAML.parse(File.read(cluster_file_path)).as_h

• return YAML.parse_all(File.read(cluster_file_path)) end

  def get_commit_date(job)

@@ -198,8 +198,13 @@ class Sched # collect all job ids job_ids = [] of String

• basic_config = cluster_config[0]
• net_id = basic_config["net_id"]? || "192.168.222"
• ip0 = basic_config["ip0"].as_i
• # switch = basic_config["switch"].to_s # useless now
• # steps for each host

• cluster_config.each do |host, config|

• cluster_config[-1].as_h.each do |host, config| tbox_group = host.to_s job_id = add_task(tbox_group, lab)

@@ -219,7 +224,15 @@ class Sched job["testbox"] = tbox_group job.update_tbox_group(tbox_group) job["node_roles"] = config["roles"].as_a.join(" ")

•  job["node_macs"] = config["macs"].as_a.join(" ")
  

•  direct_macs = config["macs"].as_a
  

•  direct_ips = [] of String
  

•  direct_macs.size.times do
  

•    raise "Host id is greater than 254, host_id: #{ip0}" if ip0 > 254
  

•    direct_ips << "#{net_id}.#{ip0}"
  

•    ip0 += 1
  

•  end
  

•  job["direct_macs"] = direct_macs.join(" ")
  

•  job["direct_ips"] = direct_ips.join(" ")
  
  response = add_job(job, job_id)
  message = (response["error"]? ? response["error"]["root_cause"] : "")
  

-- 2.23.0