Bug 1287129

Summary: Manifest import gets really slow for increasing number of entitlements in one pool
Product: Red Hat Satellite Reporter: Pavel Moravec <pmoravec>
Component: CandlepinAssignee: Barnaby Court <bcourt>
Status: CLOSED ERRATA QA Contact: Lukas Pramuk <lpramuk>
Severity: high Docs Contact:
Priority: high    
Version: 6.1.3CC: bbuckingham, lpramuk, oshtaier, sthirugn
Target Milestone: UnspecifiedKeywords: Triaged
Target Release: Unused   
Hardware: x86_64   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2016-07-27 09:21:11 UTC Type: Bug
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:

Description Pavel Moravec 2015-12-01 14:51:35 UTC 
Description of problem:
Increasing number of entitlements in a single pool in manifest causes import time of the manifest to grow very rapidly. I.e. for manifest with 15k subscriptions, it took several minutes, but for manifest with 30k subscriptions, it didnt finish within 45 minutes(sic!).

Checking candlepin logs, the cause is:

2015-12-01 07:40:44,026 [req=6868c8d9-121c-496b-a603-41c4c8847ae5, org=customer] INFO  org.candlepin.controller.CandlepinPoolManager - Revoked entitlement: 2c9281204f5ff656014ffcf7fbc46d38

takes more and more time each. Having just partial logs (all to be attached), it ended after 45 minutes of importing in revoking one entitlement in 2 seconds(!).

Since the time of processing one entitlement grows substantially over time, it is assumed some Map/Hash-like structure used in candlepin is very inefficient on insert or update when growing over some limit.

Note also a valid workaround is to split the entitlements into more pools.


Version-Release number of selected component (if applicable):
seen on 6.0.8, supposed to be elsewhere


How reproducible:
100% (expected, just seen at customer)


Steps to Reproduce:
1. Create a manifest with 2x15k entitlements (15k entitlements in 2 pools each).
2. Import the manifest to Satellite
3. Create a manifest with 30k entitlements in single pool
4. Try to import it to the Satellite (after increasing rest_client_timeout in /etc/foreman/plugins/katello.yaml)


Actual results:
2. succeeds, 4. doesnt wthin 45 minutes


Expected results:
both to succeed within similar time


Additional info:
Comment 5 Barnaby Court 2016-02-26 19:36:38 UTC 
This should be fixed in Sat 6.1.7
Comment 7 Lukas Pramuk 2016-04-26 11:12:44 UTC 
VERIFIED.

@Sat6.2.0-Beta-Snap9.2 

# time hammer subscription upload --organization "Testing Organization" --file newest_problematic_manifest.zip
[..............................................................................................................] [100%]


>>> real	0m15.797s
...

# hammer subscription list --organization "Testing Organization"
-----|----------|---------|---------|----------|----------|----------|----|---------|----------|---------
NAME | CONTRACT | ACCOUNT | SUPPORT | QUANTITY | CONSUMED | END DATE | ID | PRODUCT | QUANTITY | ATTACHED
-----|----------|---------|---------|----------|----------|----------|----|---------|----------|---------

>>> none subs imported yet, lets refresh manifest

# time hammer subscription refresh-manifest --organization "Testing Organization"
[..............................................................................................................] [100%]


>>> real	0m23.887s
...

# hammer subscription list --organization "Testing Organization"--------------------------------------------------------|----------|---------|---------|----------|----------|------------------------------|----|---------------------------------------------------------|----------|---------
NAME                                                    | CONTRACT | ACCOUNT  | SUPPORT | QUANTITY | CONSUMED | END DATE                     | ID | PRODUCT                                                 | QUANTITY | ATTACHED
--------------------------------------------------------|----------|----------|---------|----------|----------|------------------------------|----|---------------------------------------------------------|----------|---------
Red Hat Infrastructure Foundation Layer - for Providers | <edited> | <edited> | Premium | 100      | 0        | 2017-03-01T04:59:59.000+0000 | 34 | Red Hat Infrastructure Foundation Layer - for Providers | 100      | 0       
Red Hat Infrastructure Foundation Layer - for Providers | <edited> | <edited> | Premium | 12000    | 0        | 2017-03-01T04:59:59.000+0000 | 32 | Red Hat Infrastructure Foundation Layer - for Providers | 12000    | 0       
Red Hat Infrastructure Foundation Layer - for Providers | <edited> | <edited> | Premium | 5000     | 0        | 2017-03-01T04:59:59.000+0000 | 30 | Red Hat Infrastructure Foundation Layer - for Providers | 5000     | 0       
Red Hat Infrastructure Foundation Layer - for Providers | <edited> | <edited> | Premium | 10000    | 0        | 2017-03-01T04:59:59.000+0000 | 28 | Red Hat Infrastructure Foundation Layer - for Providers | 10000    | 0       
Red Hat Infrastructure Foundation Layer - for Providers | <edited> | <edited> | Premium | 3800     | 0        | 2017-03-01T04:59:59.000+0000 | 26 | Red Hat Infrastructure Foundation Layer - for Providers | 3800     | 0       
--------------------------------------------------------|----------|---------|---------|----------|----------|------------------------------|----|---------------------------------------------------------|----------|---------

>>> subs (of problematic manifest) are imported successfully within a minute 
>>> however for big subs count it seems it's a two step process now: after import (16sec) it needs a refresh (24sec) - lets have another bz for that
Comment 9 errata-xmlrpc 2016-07-27 09:21:11 UTC 
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, 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/RHBA-2016:1501