Bug 1723879

Summary: Performance Problems with Podman on systems with IO load
Product: Red Hat Enterprise Linux 8 Reporter: Matthew Heon <mheon>
Component: podmanAssignee: Brent Baude <bbaude>
Status: CLOSED ERRATA QA Contact: atomic-bugs <atomic-bugs>
Severity: high Docs Contact:
Priority: urgent    
Version: 8.1CC: dciabrin, ddarrah, dornelas, dwalsh, jligon, jnovy, lsm5, mheon, pthomas, ypu
Target Milestone: rcKeywords: ZStream
Target Release: 8.1   
Hardware: Unspecified   
OS: Unspecified   
Whiteboard:
Fixed In Version: podman-1.4.2-2 Doc Type: If docs needed, set a value
Doc Text:
Story Points: ---
Clone Of:
: 1724522 (view as bug list) Environment:
Last Closed: 2019-11-05 21:02:47 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:
Bug Depends On:    
Bug Blocks: 1724522, 1734574    

Description Matthew Heon 2019-06-25 15:34:20 UTC
Description of problem:

Podman is extremely slow on systems under heavy I/O load. 

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

Podman 1.4.2 (as build for RHEL 8.1 and Fedora - reproduces on both)

How reproducible:

100%, though severity varies based on the system it is run on

Steps to Reproduce:
1. In terminal 1, run a few containers (`podman run -t -i -d alpine top` works, but the exact contents don't matter)
2. In terminal 2: `dd if=/dev/zero of=<file on same FS as /var/lib/containers> bs=1M`
3. Back in terminal 1: `time podman info` (All other Podman commands except version also affected)
4. Stop the `dd` in terminal 2 and delete the file

Actual results:

The `podman info` in command 3 takes approximately 90 seconds to complete

Expected results:

A more reasonable execution time - a few seconds at most

Additional info:

Problem is caused by several disk writes that are holding execution waiting for the disk. While some writes are unavoidable, we can make Podman's fastpast avoid them in almost all cases.

Comment 1 Matthew Heon 2019-06-25 15:35:15 UTC
We have patches upstream that reduce execution time from 90 seconds down to around 4 on our test system.

Comment 5 Derrick Ornelas 2019-07-01 16:10:50 UTC
*** Bug 1718230 has been marked as a duplicate of this bug. ***

Comment 9 Joy Pu 2019-09-27 15:53:33 UTC
Test with podman-1.4.2-5.module+el8.1.0+4240+893c1ab8.x86_64, and time the podman info for several times, normally it will take less than 1 second to get the output. But sometimes it takes 6 or 7 seconds. All of these are fit a few seconds. So set this to verified.

Comment 11 errata-xmlrpc 2019-11-05 21:02:47 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/RHSA-2019:3403