Merge lp:~qzhang/lava-dispatcher/fix-874594 into lp:lava-dispatcher

[Spring Zhang]

1. fix 874594
2. Add default timeout to in_test_shell() to 10s
3. Delete LavaAndroidClient.in_test_shell for its duplication
4. Add timeout protection for more pexpect.expect()
5. Adjust general LavaClient timeout to 20mins
6. add boot_linaro_image action if not in test image

[Yongqin Liu (liuyq0307)]

Thanks, all seems ok except the "6. add boot_linaro_image action if not in test image" one.
I guess we execute "client.in_test_shell()" at a assumption that already boot_linaro_image,
if we do this in the test action, do we need to check "in_test_shell" again?

And for test action of android, I think it should be client.boot_linaro_android_image.

review: Needs Fixing

[lp:~qzhang/lava-dispatcher/fix-874594]

157. By Spring Zhang on 2011-10-25

fix boot_linaro_android_image in android test cases

[Spring Zhang (qzhang)]

> Thanks, all seems ok except the "6. add boot_linaro_image action if not in
> test image" one.
> I guess we execute "client.in_test_shell()" at a assumption that already
> boot_linaro_image,
> if we do this in the test action, do we need to check "in_test_shell" again?
I'd like to make sure again, assuming if some test cases run ordinally, one test case may make system no response, before the next case, it's better to make sure the system still alive, or do the action to boot to test image.

>
> And for test action of android, I think it should be
> client.boot_linaro_android_image.
Good catch, fixed, thanks.

review: Needs Resubmitting

[Spring Zhang (qzhang)]

I have finished the test in normal image, Yongqin, if you are available it's better to have a try on Android image.

[Yongqin Liu (liuyq0307)]

I think it's better do this after the merge https://code.launchpad.net/~mwhudson/lava-dispatcher/eliminate-old-android-tests/+merge/80522. mwhudson submitted. because that merge will delete some files we modified here. after that we only need to review and commit the useful files.

[lp:~qzhang/lava-dispatcher/fix-874594]

158. By Spring Zhang on 2011-10-27

merge with mainline

[Spring Zhang (qzhang)]

tested with mainline merged and pass

review: Needs Resubmitting

[Yongqin Liu (liuyq0307)]

looks ok except the problem on page http://pastebin.ubuntu.com/720486/.
after that fixing, the test with android can run and submit result.

But not test the timeout things.

review: Needs Fixing

[lp:~qzhang/lava-dispatcher/fix-874594]

159. By Spring Zhang on 2011-10-27

fix grammar error

160. By Spring Zhang on 2011-10-27

Merge Yongqin's fix to Android part

[Spring Zhang (qzhang)]

Thanks for the android test

[Yongqin Liu (liuyq0307)]

forgot to change to approve

review: Approve

=== modified file 'lava_dispatcher/__init__.py'
--- lava_dispatcher/__init__.py	2011-10-27 02:32:02 +0000
+++ lava_dispatcher/__init__.py	2011-10-27 09:14:23 +0000
@@ -25,6 +25,7 @@
 from uuid import uuid1
 import base64
 import pexpect
+import logging
 
 from lava_dispatcher.actions import get_all_cmds
 from lava_dispatcher.config import get_config, get_device_config
@@ -87,6 +88,7 @@
                     status = 'pass'
                 finally:
                     err_msg = ""
+                    logging.info("Action %s finished." % cmd['command'])
                     if status == 'fail':
                         err_msg = "Lava failed at action %s with error: %s\n" %\
                                   (cmd['command'], err)

=== modified file 'lava_dispatcher/actions/android_deploy.py'
--- lava_dispatcher/actions/android_deploy.py	2011-10-20 18:24:55 +0000
+++ lava_dispatcher/actions/android_deploy.py	2011-10-27 09:14:23 +0000
@@ -38,7 +38,7 @@
         logging.info("  boot: %s" % boot)
         logging.info("  system: %s" % system)
         logging.info("  data: %s" % data)
-        logging.info("Booting master image")
+        logging.info("Boot master image")
         client.boot_master_image()
 
         logging.info("Waiting for network to come up...")

=== modified file 'lava_dispatcher/actions/boot_control.py'
--- lava_dispatcher/actions/boot_control.py	2011-10-26 02:59:28 +0000
+++ lava_dispatcher/actions/boot_control.py	2011-10-27 09:14:23 +0000
@@ -34,8 +34,8 @@
         client.proc.sendline("")
         try:
             client.boot_linaro_android_image()
-        except:
-            logging.exception("boot_linaro_android_image failed")
+        except Exception as e:
+            logging.exception("boot_linaro_android_image failed: %s" % e)
             raise CriticalError("Failed to boot test image.")
 
 class cmd_boot_linaro_image(BaseAction):
@@ -61,5 +61,5 @@
     """
     def run(self):
         client = self.client
-        logging.info("Boot Master image")
+        logging.info("Boot master image")
         client.boot_master_image()

=== modified file 'lava_dispatcher/actions/lava-test.py'
--- lava_dispatcher/actions/lava-test.py	2011-10-26 02:59:28 +0000
+++ lava_dispatcher/actions/lava-test.py	2011-10-27 09:14:23 +0000
@@ -91,7 +91,10 @@
         logging.info("Executing lava_test_run %s command" % test_name)
         #Make sure in test image now
         client = self.client
-        client.in_test_shell()
+        try:
+            client.in_test_shell()
+        except:
+            client.boot_linaro_image()
         client.run_cmd_tester('mkdir -p %s' % self.context.lava_result_dir)
         client.export_display()
         bundle_name = test_name + "-" + datetime.now().strftime("%H%M%S")

=== modified file 'lava_dispatcher/client.py'
--- lava_dispatcher/client.py	2011-10-27 01:09:17 +0000
+++ lava_dispatcher/client.py	2011-10-27 09:14:23 +0000
@@ -99,15 +99,18 @@
             timeout=timeout)
         if id == 1:
             raise OperationFailed
+        logging.info("System is in master image now")
 
-    def in_test_shell(self):
+    def in_test_shell(self, timeout=10):
         """
         Check that we are in a shell on the test image
         """
         self.proc.sendline("")
-        match_id = self.proc.expect([self.tester_str, pexpect.TIMEOUT])
+        match_id = self.proc.expect([self.tester_str, pexpect.TIMEOUT],
+                    timeout=timeout)
         if match_id == 1:
             raise OperationFailed
+        logging.info("System is in test image now")
 
     def boot_master_image(self):
         """
@@ -122,20 +125,20 @@
             self.proc.hard_reboot()
             self.in_master_shell(300)
         self.proc.sendline('export PS1="$PS1 [rc=$(echo \$?)]: "')
-        self.proc.expect(self.master_str)
+        self.proc.expect(self.master_str, timeout=10)
 
     def boot_linaro_image(self):
         """
         Reboot the system to the test image
         """
         self.proc._boot(self.boot_cmds)
-        self.in_test_shell()
+        self.in_test_shell(300)
         # set PS1 to include return value of last command
         # Details: system PS1 is set in /etc/bash.bashrc and user PS1 is set in
         # /root/.bashrc, it is
         # "${debian_chroot:+($debian_chroot)}\u@\h:\w\$ "
         self.proc.sendline('export PS1="$PS1 [rc=$(echo \$?)]: "')
-        self.proc.expect(self.tester_str)
+        self.proc.expect(self.tester_str, timeout=10)
 
     def run_shell_command(self, cmd, response=None, timeout=-1):
         self.empty_pexpect_buffer()
@@ -228,7 +231,7 @@
 
     def boot_linaro_android_image(self):
         """Reboot the system to the test android image."""
-        self._boot(string_to_list(self.config.get('boot_cmds_android')))
+        self.proc._boot(string_to_list(self.config.get('boot_cmds_android')))
         self.in_test_shell()
         self.proc.sendline("export PS1=\"root@linaro: \"")
 

=== modified file 'lava_dispatcher/connection.py'
--- lava_dispatcher/connection.py	2011-10-26 03:24:05 +0000
+++ lava_dispatcher/connection.py	2011-10-27 09:14:23 +0000
@@ -66,6 +66,7 @@
     def soft_reboot(self):
         self.proc.sendline("reboot")
         # set soft reboot timeout 120s, or do a hard reset
+        logging.info("Rebooting the system")
         id = self.proc.expect(
             ['Will now restart', pexpect.TIMEOUT], timeout=120)
         if id != 0:
@@ -79,12 +80,13 @@
 
     def _make_connection(self, sio):
         cmd = "conmux-console %s" % self.device_option("hostname")
-        proc = pexpect.spawn(cmd, timeout=3600, logfile=sio)
+        proc = pexpect.spawn(cmd, timeout=1200, logfile=sio)
         #serial can be slow, races do funny things if you don't increase delay
         proc.delaybeforesend=1
         return proc
 
     def hard_reboot(self):
+        logging.info("Perform hard reset on the system")
         self.proc.send("~$")
         self.proc.sendline("hardreset")
         # XXX Workaround for snowball