ctrd: support to restart containerd if it is dead

[fuweid]

Signed-off-by: Wei Fu fuweid89@gmail.com

Ⅰ. Describe what this PR did

Monitor the containerd process and auto restart it if the containerd is dead.

Ⅱ. Does this pull request fix one issue?

NONE

Ⅲ. Why don't you add test cases (unit test/integration test)? (你真的觉得不需要加测试吗？)

no need

Ⅳ. Describe how to verify it

start pouchd and pkill containerd. The containerd should be restarted.

Ⅴ. Special notes for reviews

[codecov]

Codecov Report

Merging #2829 into master will decrease coverage by 0.01%.
The diff coverage is 36.36%.

@@            Coverage Diff             @@
##           master    #2829      +/-   ##
==========================================
- Coverage   69.18%   69.17%   -0.02%     
==========================================
  Files         279      279              
  Lines       17555    17607      +52     
==========================================
+ Hits        12146    12179      +33     
- Misses       4041     4053      +12     
- Partials     1368     1375       +7

Flag	Coverage Δ
#criv1alpha2_test	39.32% <30.9%> (+0.02%)	⬆️
#integration_test_0	36.38% <36.36%> (+0.04%)	⬆️
#integration_test_1	35.18% <30.9%> (+0.03%)	⬆️
#integration_test_2	36.36% <30.9%> (-0.02%)	⬇️
#integration_test_3	35.3% <30.9%> (+0.06%)	⬆️
#node_e2e_test	35.08% <30.9%> (+0.09%)	⬆️
#unittest	28.7% <ø> (ø)	⬆️

Impacted Files	Coverage Δ
ctrd/supervisord/daemon.go	50.67% <36.36%> (+2.74%)	⬆️
pkg/streams/utils.go	82.14% <0%> (-7.15%)	⬇️
cri/v1alpha2/cri.go	69.31% <0%> (+0.5%)	⬆️
ctrd/container.go	54.64% <0%> (+0.75%)	⬆️

[HusterWan]

Already reviewed, LGTM