Description of problem: The customer is facing an issue when pushing information related to the huge virtualization environment *267 hypervisors and 16723 guests found* Version-Release number of selected component (if applicable): 6.3.x How reproducible: 100% Steps to Reproduce: 1. Using JSON customer file 2. Configure fake virt-who conf 3. virt-who -od Actual results: --- 2018-08-31 20:44:11,229 [virtwho.destination_-4756648715778885625 ERROR] MainProcess(3194):Thread-3 @virt.py:run:388 - Thread 'destination_-4756648715778885625' fails with exception: Traceback (most recent call last): File "/usr/lib/python2.7/site-packages/virtwho/virt/virt.py", line 379, in run self._run() File "/usr/lib/python2.7/site-packages/virtwho/virt/virt.py", line 335, in _run self._send_data(data_to_send) File "/usr/lib/python2.7/site-packages/virtwho/virt/virt.py", line 665, in _send_data self.dest.check_report_state(batch_host_guest_report) File "/usr/lib/python2.7/site-packages/virtwho/manager/subscriptionmanager/subscriptionmanager.py", line 353, in check_report_state for fail in result_data.get('failedUpdate', []): AttributeError: 'unicode' object has no attribute 'get' 2018-08-31 20:44:11,229 [virtwho.destination_-4756648715778885625 INFO] MainProcess(3194):Thread-3 @virt.py:_send_data:553 - Error report received --- Expected results: No errors Additional info:
Waldirio, we looked into this, it seems our internal handling of in-clause batching generates SQL requests which are too large. As a workaround, we successfully tried setting "db.config.in.operator.block.size = 1000" in candlepin.conf (default is 32767), and then candlepin will automatically break the resulting queries up into much smaller queries. We still need to do some further investigation in order to determine if there is a more optimal value, but this setting should suffice as a work-around in the meanwhile. For reference, we run with a limit of 15000 when running against mariadb; we'll do some testing and change the postgre default as appropriate (I suspect either 1000 or 15000 will become the new default).
Given that the workaround suggested in comment 3 works, marking as dupe. *** This bug has been marked as a duplicate of bug 1599752 ***