On Sat, Jan 16, 2021 at 03:03:02PM +0800, Wei Jihui wrote:

[why] user can choose when to ssh, use ssh.yaml is ssh when task normal finish use ssh-on-fail.yaml is ssh when task is failed use ssh-always is ssh always whether task is failed or not.

Using ssh|ssh-on-fail|ssh-always may be suitable.

[usage] submit -i ssh.yaml job.yaml

-i ssh?

Thanks, RenWen

Signed-off-by: Wei Jihui weijihuiall@163.com

jobs/ssh-always.yaml | 4 ++++ jobs/ssh-on-fail.yaml | 12 ++++++++++++ jobs/{fault_reproduction.yaml => ssh.yaml} | 4 ---- 3 files changed, 16 insertions(+), 4 deletions(-) create mode 100644 jobs/ssh-always.yaml create mode 100644 jobs/ssh-on-fail.yaml rename jobs/{fault_reproduction.yaml => ssh.yaml} (80%)

diff --git a/jobs/ssh-always.yaml b/jobs/ssh-always.yaml new file mode 100644 index 00000000..0166fe12 --- /dev/null +++ b/jobs/ssh-always.yaml @@ -0,0 +1,4 @@ +# ssh to testbox and sleep always

+<< : jobs/ssh.yaml +<< : jobs/ssh-on-fail.yaml diff --git a/jobs/ssh-on-fail.yaml b/jobs/ssh-on-fail.yaml new file mode 100644 index 00000000..5f28d001 --- /dev/null +++ b/jobs/ssh-on-fail.yaml @@ -0,0 +1,12 @@ +ssh_pub_key:

• <%=
• begin

• File.read("#{ENV['HOME']}/.ssh/id_rsa.pub").chomp
  

• rescue

• nil
  

• end
• %>

+on_fail:

• sshd:
• sleep: 6h

diff --git a/jobs/fault_reproduction.yaml b/jobs/ssh.yaml similarity index 80% rename from jobs/fault_reproduction.yaml rename to jobs/ssh.yaml index 46a1e042..11e5b175 100644 --- a/jobs/fault_reproduction.yaml +++ b/jobs/ssh.yaml @@ -8,7 +8,3 @@ ssh_pub_key: %> sshd: sleep: 6h

-on_fail:

•    sshd:
  

•    sleep: 6h
  

-- 2.23.0

On Sat, Jan 16, 2021 at 03:03:02PM +0800, Wei Jihui wrote:

[why] user can choose when to ssh, use ssh.yaml is ssh when task normal finish use ssh-on-fail.yaml is ssh when task is failed use ssh-always is ssh always whether task is failed or not.

[usage] submit -i ssh.yaml job.yaml

Signed-off-by: Wei Jihui weijihuiall@163.com

jobs/ssh-always.yaml | 4 ++++ jobs/ssh-on-fail.yaml | 12 ++++++++++++ jobs/{fault_reproduction.yaml => ssh.yaml} | 4 ---- 3 files changed, 16 insertions(+), 4 deletions(-) create mode 100644 jobs/ssh-always.yaml create mode 100644 jobs/ssh-on-fail.yaml rename jobs/{fault_reproduction.yaml => ssh.yaml} (80%)

diff --git a/jobs/ssh-always.yaml b/jobs/ssh-always.yaml new file mode 100644 index 00000000..0166fe12 --- /dev/null +++ b/jobs/ssh-always.yaml @@ -0,0 +1,4 @@ +# ssh to testbox and sleep always

"ssh-always" 会被人无解为别的意思，而不是"sleep always". 而且其实也不是sleep always，是sleep on fail.

+<< : jobs/ssh.yaml +<< : jobs/ssh-on-fail.yaml

两个sshd，第二个会失败吧？