Bug 2228592 - Incremental export using fs-export generates very slow queries and takes huge time to export
Summary: Incremental export using fs-export generates very slow queries and takes huge...
Keywords:
Status: NEW
Alias: None
Product: Red Hat Satellite
Classification: Red Hat
Component: Pulp
Version: 6.12.4
Hardware: Unspecified
OS: Unspecified
medium
medium
Target Milestone: Unspecified
Assignee: satellite6-bugs
QA Contact: Satellite QE Team
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2023-08-02 19:14 UTC by Joniel Pasqualetto
Modified: 2023-08-17 00:19 UTC (History)
4 users (show)

Fixed In Version:
Doc Type: If docs needed, set a value
Doc Text:
Clone Of:
Environment:
Last Closed:
Target Upstream Version:
Embargoed:


Attachments (Terms of Use)

Description Joniel Pasqualetto 2023-08-02 19:14:24 UTC
Description of problem:

Very big library (complete export of 1.5Tb). When trying to generate an incremental export of the library (all repositories), a very long time is spent on some DB queries. 

Version-Release number of selected component (if applicable):


How reproducible:
Unsure. All the time on customer's environment

Steps to Reproduce:
In theory, that's what customer did:

1. Generate complete export of big library
2. After a while, generate incremental exports using fs-export (syncable method)


Actual results:
Export works, but is very slow. Last test from customer took 24h running. Result had 3.5Gb of data (which is not wrong, just time to generate it is considered too long)

Expected results:

Export to finish faster.

Additional info:

Comment 2 Daniel Alley 2023-08-07 18:58:02 UTC
I don't know that this is one issue, but perhaps a combination of https://github.com/pulp/pulpcore/pull/4161 (which is already tracked by BZ https://bugzilla.redhat.com/show_bug.cgi?id=2226950) and https://github.com/pulp/pulpcore/issues/3969 ?

We have that work in progress currently for other stakeholders anyway, so we can check back in shortly when it is finished and see if the situation is improved.


Note You need to log in before you can comment on or make changes to this bug.