upgrade ceos to version 4.25.5.1M, fix the issue of arista endless lopp (#4228)

Description of PR
Summary:
Fixes # (issue)

Type of change
 Bug fix
 Testbed and Framework(new/improvement)

Approach
What is the motivation for this PR?
cEOS containers' mgmt-ip occasionally unavaillable, upgrade cEOS image version will fix that
cEOS terminal behavior slightly different from vEOS, enhance arista.py to avoid endless loop

How did you do it?
upgrade cEOS image version
break loop if ssh connect returns merely an 'exit', as for other commands who relies on the return e.g. 'show interfaces po1 | json', 'exit' won't appear in input_buffer, so the change won't influence it.

How did you verify/test it?
Run platform_tests/test_advanced_reboot.py test on physical testbed and it didn't stuck again.
Tried run command 'show interfaces po1 | json', it's not influenced.

Co-authored-by: Jianquan Ye <[email protected]>

Author: yejianquan

ansible/group_vars/all/ceos.yml

@@ -1,4 +1,4 @@
-ceos_image_filename: cEOS64-lab-4.23.2F.tar.xz
-ceos_image_orig: ceosimage:4.23.2F
-ceos_image: ceosimage:4.23.2F-1
+ceos_image_filename: cEOS64-lab-4.25.5.1M.tar
+ceos_image_orig: ceosimage:4.25.5.1M
+ceos_image: ceosimage:4.25.5.1M-1
 skip_ceos_image_downloading: false

ansible/roles/test/files/ptftests/arista.py

@@ -91,8 +91,15 @@ def do_cmd(self, cmd, prompt = None):
             self.shell.send(cmd + '\n')
 
         input_buffer = ''
+        loop_times = 0
         while re.search(prompt, input_buffer) is None:
             input_buffer += self.shell.recv(16384)
+            loop_times += 1
+            # cEOS will not return a arista_prompt if you send lots of 'exit' to close the ssh connect(vEOS do will),
+            # then an endless loop emerges,
+            # so if input_buffer is merely an 'exit', we can break the loop immediately
+            if loop_times > 10 and input_buffer.replace('\n', '').replace('\r', '').strip().lower() == 'exit':
+                break
 
         return input_buffer