[Pulp-list] what removes my sync schedules? *workaround found*

[Andreas Piesk]

On 23.02.2013 14:15, Andreas Piesk wrote:
>>
>> what's going on? i'm completely in the dark, no log messages, nothing. does anyone have an idea what
>> might be the reason for this erratic behaviour?
>>

OK, i think i have an idea. the task are not loaded at startup because the task snapshots cannot be
de-serialized.


async.initialize()
async._load_persisted_tasks()
TaskSnapshot(snapshot).to_task()
pickle.loads(task_class) <- hangs

i had two task snapshots and one of them was the culprit:

$ mongo pulp_database
MongoDB shell version: 2.2.3
connecting to: pulp_database
> db.task_snapshots.find()
{ "_id" : "2456012e-8f5f-4628-a36c-ff971b3e029d", "synchronizer_class" :
"cpulp.server.api.synchronizers\nYumSynchronizer\np0\n.", "weight" : 2, "class_name" : null, "hooks"
:
"(dp0\\nS'dequeue'\\np1\\n(lp2\\ncpulp.server.api.repo_sync\\npost_sync\\np3\\naccopy_reg\\n_reconstructor\\np4\\n(cpulp.server.auth.authorization\\nRevokePermissionsForTask\\np5\\nc__builtin__\\nobject\\np6\\nNtp7\\nRp8\\n(dp9\\nS'user_name'\\np10\\nVSYSTEM\\np11\\nsbasS'enqueue'\\np12\\n(lp13\\ng4\\n(cpulp.server.auth.authorization\\nGrantPermissionsForTask\\np14\\ng6\\nNtp15\\nRp16\\n(dp17\\ng10\\ng11\\nsbas.",
"start_time" : "N.", "args" : "(lp0\\nS'centos5-i386-os'\\np1\\na.", "callable" :
"cpulp.server.api.repo_sync\\n_sync\\np0\\n.", "result" : "N.", "timeout_delta" : "N.", "id" :
"5110dce8-76a1-11e2-b158-001676c998d5", "_ns" : "task_snapshots", "method_name" : "_sync",
"exception" : "N.", "repo_id" : "centos5-i386-os", "job_id" : null, "cancel_attempts" : 0,
"schedule_threshold" : "cdatetime\\ntimedelta\\np0\\n(I0\\nI300\\nI0\\ntp1\\nRp2\\n.", "traceback" :
"N.", "task_class" : "cpulp.server.api.repo_sync_task\\nRepoSyncTask\\np0\\n.", "finish_time" :
"N.", "state" : "waiting", "failure_threshold" : null, "_progress_callback" :
"cpulp.server.api.synchronizers\\nyum_rhn_progress_callback\\np0\\n.", "kwargs" :
"(dp0\\nVskip\\np1\\nccopy_reg\\n_reconstructor\\np2\\n(cbson.son\\nSON\\np3\\nc__builtin__\\ndict\\np4\\n(dp5\\nVerrata\\np6\\nI1\\nstp7\\nRp8\\n(dp9\\nS'_SON__keys'\\np10\\n(lp11\\ng6\\nasbsVthreads\\np12\\nI1\\ns.",
"progress" : null, "consecutive_failures" : 0 }
{ "_id" : "f9e13bcb-673d-4dea-ba1f-0dfa827d6896", "synchronizer_class" :
"cpulp.server.api.synchronizers\nYumSynchronizer\np0\n.", "weight" : 2, "class_name" : null, "hooks"
:
"(dp0\\nS'dequeue'\\np1\\n(lp2\\ncpulp.server.api.repo_sync\\npost_sync\\np3\\naccopy_reg\\n_reconstructor\\np4\\n(cpulp.server.auth.authorization\\nRevokePermissionsForTask\\np5\\nc__builtin__\\nobject\\np6\\nNtp7\\nRp8\\n(dp9\\nS'user_name'\\np10\\nVSYSTEM\\np11\\nsbasS'enqueue'\\np12\\n(lp13\\ng4\\n(cpulp.server.auth.authorization\\nGrantPermissionsForTask\\np14\\ng6\\nNtp15\\nRp16\\n(dp17\\ng10\\ng11\\nsbas.",
"start_time" : "N.", "args" : "(lp0\\nS'centos5-x86_64-os'\\np1\\na.", "callable" :
"cpulp.server.api.repo_sync\\n_sync\\np0\\n.", "result" : "N.", "timeout_delta" : "N.", "id" :
"50d414dc-76a1-11e2-afca-001676c998d5", "_ns" : "task_snapshots", "method_name" : "_sync",
"exception" : "N.", "repo_id" : "centos5-x86_64-os", "job_id" : null, "cancel_attempts" : 0,
"schedule_threshold" : "cdatetime\\ntimedelta\\np0\\n(I0\\nI300\\nI0\\ntp1\\nRp2\\n.", "traceback" :
"N.", "task_class" : "cpulp.server.api.repo_sync_task\\nRepoSyncTask\\np0\\n.", "finish_time" :
"N.", "state" : "waiting", "failure_threshold" : null, "_progress_callback" :
"cpulp.server.api.synchronizers\\nyum_rhn_progress_callback\\np0\\n.", "kwargs" :
"(dp0\\nVskip\\np1\\nccopy_reg\\n_reconstructor\\np2\\n(cbson.son\\nSON\\np3\\nc__builtin__\\ndict\\np4\\n(dp5\\nVerrata\\np6\\nI1\\nstp7\\nRp8\\n(dp9\\nS'_SON__keys'\\np10\\n(lp11\\ng6\\nasbsVthreads\\np12\\nI1\\ns.",
"progress" : null, "consecutive_failures" : 0 }
>


After deleting all task snapshots all tasks are loaded at startup. Problem is not solved but at
least a have a work around.

regards,
-ap