Note: This bug is displayed in read-only format because
the product is no longer active in Red Hat Bugzilla.
Red Hat Satellite engineering is moving the tracking of its product development work on Satellite to Red Hat Jira (issues.redhat.com). If you're a Red Hat customer, please continue to file support cases via the Red Hat customer portal. If you're not, please head to the "Satellite project" in Red Hat Jira and file new tickets here. Individual Bugzilla bugs will be migrated starting at the end of May. If you cannot log in to RH Jira, please consult article #7032570. That failing, please send an e-mail to the RH Jira admins at rh-issues@redhat.com to troubleshoot your issue as a user management inquiry. The email creates a ServiceNow ticket with Red Hat. Individual Bugzilla bugs that are migrated will be moved to status "CLOSED", resolution "MIGRATED", and set with "MigratedToJIRA" in "Keywords". The link to the successor Jira issue will be found under "Links", have a little "two-footprint" icon next to it, and direct you to the "Satellite project" in Red Hat Jira (issue links are of type "https://issues.redhat.com/browse/SAT-XXXX", where "X" is a digit). This same link will be available in a blue banner at the top of the page informing you that that bug has been migrated.
Description of problem:
When katello:clean_backend_objects rake script is running on a system with >1500 candlepin consumers at a time a consumer updates its facts, the rake script can wrongly detect a system is missing in candlepin, like:
Host 16080 hostname.example.com ff0a6edf-311c-4924-a7b9-b72707931c7b is partially missing subscription information. Un-registering
The reason is katello within ::Katello::Resources::Candlepin::Consumer::all_uuids call queries candlepin in pages:
2023-03-23 07:46:34,701 [thread=https-jsse-nio-127.0.0.1-23443-exec-9] [req=849f90be-9f62-40e9-9541-341e3055916b, org=, csid=] INFO org.candlepin.servlet.filter.logging.LoggingFilter - Request: verb=GET, uri=/candlepin/consumers/?owner=MyOrganization&include=uuid&per_page=1500&page=1
2023-03-23 07:48:50,123 [thread=https-jsse-nio-127.0.0.1-23443-exec-75] [req=a5d01c04-bf19-45e2-a42a-9b23493598c3, org=, csid=] INFO org.candlepin.servlet.filter.logging.LoggingFilter - Request: verb=GET, uri=/candlepin/consumers/?owner=MyOrganization&include=uuid&per_page=1500&page=2
2023-03-23 07:50:21,081 [thread=https-jsse-nio-127.0.0.1-23443-exec-26] [req=40f37098-fbed-40a6-972e-f70510000ced, org=, csid=391be221-c85d-4224-953b-d7df9143650b] INFO org.candlepin.servlet.filter.logging.LoggingFilter - Request: verb=PUT, uri=/candlepin/consumers/c8a22c7e-d959-42ef-b6a2-1eecbffd2459
2023-03-23 07:50:41,851 [thread=https-jsse-nio-127.0.0.1-23443-exec-20] [req=944317ac-69f3-45c5-9467-92b78734b276, org=, csid=e68d1403-cc68-4c79-a6a3-b8bd329fe9da] INFO org.candlepin.servlet.filter.logging.LoggingFilter - Request: verb=PUT, uri=/candlepin/consumers/8484b642-fbb1-43ff-800a-26e792c37dfc
2023-03-23 07:51:11,228 [thread=https-jsse-nio-127.0.0.1-23443-exec-63] [req=6b6a49be-eaff-4a16-9626-2ecc58bfd97e, org=, csid=] INFO org.candlepin.servlet.filter.logging.LoggingFilter - Request: verb=GET, uri=/candlepin/consumers/?owner=MyOrganization&include=uuid&per_page=1500&page=3
The problem occurs when a client updates consumer facts *between* the "get me next 1.5k consumers" requests - like the PUT requests above. This shuffles ordering of the consumers in a response, causing some UUID is skipped and some is present twice in the overall all_uuids call (we can easily demonstrate this).
Those skipped consumers are then wrongly marked as "partially missing subscription information".
Version-Release number of selected component (if applicable):
Sat 6.12.2 / any older as well
How reproducible:
very reliably
Steps to Reproduce:
1. Have a Satellite with >1.5k Content Hosts
2. invoke "foreman-rake katello:clean_backend_objects"
3. while it is running, run "subscription-manager facts --update" on a few hosts
4. check the rake script output
Actual results:
4. with some (high) probability, the rake script will wrongly detect a missing candlepin consumer
Expected results:
4. no such false alarms
Additional info:
We can easily demonstrate the "PUTing consumer facts shuffles consumers ordeing" on >10 hosts system as well. Just enable candlepin API from cmdline per https://access.redhat.com/solutions/2955931 , and replace "get me clients in 1.5k batches" requests by ".. in batches of, say, 7"
for i in $(seq 1 4); do curl -sk -H "Content-Type:application/json" -u admin:admin "https://localhost:23443/candlepin/consumers/?owner=YOUR_ORGANIZATION&include=uuid&per_page=7&page=${i}" | json_reformat; sleep 1; echo "uuid"; done | grep uuid
Normally, it will print something like:
"uuid": "08513930-d0f7-4882-8129-77673dc721c2"
"uuid": "cbc30fb8-9939-467d-9273-052eaf791eb4"
"uuid": "7ba62210-3c30-4a4c-ae0a-fb31df3bc33d"
"uuid": "f83e11cc-c0c3-479d-83ae-573fad2090bc"
"uuid": "9c8bbc55-ac77-4e2c-853d-a2a6fc24862a"
"uuid": "f72296a5-0b56-48b3-8712-f2ff4813e778"
"uuid": "762a8c10-2793-489e-a53f-a3ffb2e3acb0"
uuid
"uuid": "aac2882d-0bf6-4336-9d41-b30c94c452de"
"uuid": "60787875-b416-4aaf-bf1e-bd365639f4a8"
"uuid": "adb8a530-03a9-4c90-95b6-845b742adeb4"
"uuid": "f8bffcba-cd62-4a0f-b86d-f02a2f27e14f"
"uuid": "b8787ab1-58aa-4a14-b1f6-68eae436954e"
"uuid": "8f7d4e4e-6b7f-4498-9019-30dad9b8a62a"
"uuid": "f8677b63-d1b3-4495-a86e-00d415995bfd"
uuid
"uuid": "62d3e041-00c2-4adf-8135-055815954203"
"uuid": "f72296a5-0b56-48b3-8712-f2ff4813e778"
"uuid": "0e7893a7-fc9c-4b5b-97d7-b44c8b120071"
"uuid": "8f7d4e4e-6b7f-4498-9019-30dad9b8a62a"
"uuid": "663af0ad-4206-4294-a947-83fd5cfa612d"
"uuid": "adb8a530-03a9-4c90-95b6-845b742adeb4"
"uuid": "f8677b63-d1b3-4495-a86e-00d415995bfd"
uuid
"uuid": "7ba62210-3c30-4a4c-ae0a-fb31df3bc33d"
"uuid": "b8787ab1-58aa-4a14-b1f6-68eae436954e"
"uuid": "aac2882d-0bf6-4336-9d41-b30c94c452de"
"uuid": "82480bc1-3737-429c-8947-d1de745e9e27"
"uuid": "762a8c10-2793-489e-a53f-a3ffb2e3acb0"
uuid
Every time the same sequence. Now, *during* executing the script (play with the sleep time there), run "subscription-manager facts --update" on a client. The output will be e.g.:
"uuid": "08513930-d0f7-4882-8129-77673dc721c2"
"uuid": "cbc30fb8-9939-467d-9273-052eaf791eb4"
"uuid": "7ba62210-3c30-4a4c-ae0a-fb31df3bc33d"
"uuid": "f83e11cc-c0c3-479d-83ae-573fad2090bc"
"uuid": "9c8bbc55-ac77-4e2c-853d-a2a6fc24862a"
"uuid": "f72296a5-0b56-48b3-8712-f2ff4813e778"
"uuid": "762a8c10-2793-489e-a53f-a3ffb2e3acb0"
uuid
"uuid": "aac2882d-0bf6-4336-9d41-b30c94c452de"
"uuid": "60787875-b416-4aaf-bf1e-bd365639f4a8"
"uuid": "adb8a530-03a9-4c90-95b6-845b742adeb4"
"uuid": "f8bffcba-cd62-4a0f-b86d-f02a2f27e14f"
"uuid": "b8787ab1-58aa-4a14-b1f6-68eae436954e"
"uuid": "8f7d4e4e-6b7f-4498-9019-30dad9b8a62a"
"uuid": "f8677b63-d1b3-4495-a86e-00d415995bfd"
uuid
"uuid": "f83e11cc-c0c3-479d-83ae-573fad2090bc"
"uuid": "f72296a5-0b56-48b3-8712-f2ff4813e778"
"uuid": "0e7893a7-fc9c-4b5b-97d7-b44c8b120071"
"uuid": "8f7d4e4e-6b7f-4498-9019-30dad9b8a62a"
"uuid": "663af0ad-4206-4294-a947-83fd5cfa612d"
"uuid": "adb8a530-03a9-4c90-95b6-845b742adeb4"
"uuid": "f8677b63-d1b3-4495-a86e-00d415995bfd"
uuid
"uuid": "7ba62210-3c30-4a4c-ae0a-fb31df3bc33d"
"uuid": "b8787ab1-58aa-4a14-b1f6-68eae436954e"
"uuid": "aac2882d-0bf6-4336-9d41-b30c94c452de"
"uuid": "82480bc1-3737-429c-8947-d1de745e9e27"
"uuid": "762a8c10-2793-489e-a53f-a3ffb2e3acb0"
uuid
See the 62d3e041-.. is replaced by another f83e11cc-.., so the rake script won't get UUID 62d3e041 any more, during this run.
Hi Michael and Jeremy,
rather double-check confirmation: The code fix in ':sort_by => "uuid"' will help also when virt-who is updating it mapping between the paginated "get me consumers" requests, am I right?
(we had a very unlucky customer where 6.12->6.13 upgrade run automatic clean_backend_objects script with COMMIT=true, and at exactly that "bad" time, virt-who sent its mapping - this ended up in 600 Hosts unregistered during the upgrade)
Also, when considering impacts of the sorting; shall not it be rather per `created` timestamp (ascending ordering)? What if a new Host is registered just during the clean_backend_objects execution - its new uuid can break the linearity/ordering of "get me consumers per pages" responses..
> shall not it be rather per `created` timestamp (ascending ordering)? What if a new Host is registered just during the clean_backend_objects execution - its new uuid can break the linearity/ordering of "get me consumers per pages" responses..
Hey Pavel
I thought about this as well. If we still have problems we can change it again to sort by created, but the reason we didn't do this from the start is that the "all_uuids" method we altered is currently restricted to only the "uuid" field and does not include the "created" field in the response.
This change should help when more consumers are added during the rake task, no matter if they're added by virt-who mapping or another method.
Hi Pavel/Jeremy,
The issue I discovered was because candlepin was duplicating ids, instead of getting a unique list. It happened with an organization with 6,065 host ids. Candlepin was returning 6,065 host uuids, but they were not all unique, it was getting duplicate hosts in its output. So really candlepin had roughly 5,000 unique ids, while katello had 6,065 unique ids. Which means they had a different count, so hosts got removed.
I don't think there is any linearity/ordering issue. The issue was just about a unique list of hosts ids. So if one is added, and is out of order, shouldn't matter (unless katello doesn't have that id in its list). I could have missed something, but that was my understanding of the code.
Kind regards,
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.
For information on the advisory (Important: Satellite 6.15.0 release), and where to find the updated
files, follow the link below.
If the solution does not work for you, open a new bug report.
https://access.redhat.com/errata/RHSA-2024:2010