Bug 1402026

Summary: During OverCloud Deploy initial PXE deploy boot is extremely slow when using UCS VIC
Product: Red Hat OpenStack Reporter: John Browning <jobrowni>
Component: ipxeAssignee: RHOS Maint <rhos-maint>
Status: CLOSED NOTABUG QA Contact: Shai Revivo <srevivo>
Severity: medium Docs Contact:
Priority: unspecified    
Version: 9.0 (Mitaka)CC: apevec, bfournie, emarquez, jobrowni, lhh, mburns, slinaber, srevivo
Target Milestone: ---Keywords: ZStream
Target Release: ---Flags: bfournie: needinfo? (jobrowni)
Hardware: x86_64   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: If docs needed, set a value
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2017-09-19 20:31:49 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:

Description John Browning 2016-12-06 15:23:24 UTC
Description of problem:
We are deploying RHOSP9 on UCS B200M4 Blades using UCS VNIC (VIC/VNIC Virtual Interface Card) and during pxe boot, the nodes take a very long time to pull down the ramdisk image.  This has happened at 2 customers now in similar environments.

Version-Release number of selected component (if applicable):
RHOSP9

How reproducible:

Always

Steps to Reproduce:
1. Deploy overcloud on UCS B200 M4 machines
2.
3.

Actual results:
Very slow pxe boot

Expected results:
Fast pxe boot like other hardware

Additional info:
On customers using UCS gear, without using a VNIC/VIC, the machines pxe boot at normal speed.  When using VNIC/VIC on same gear, pxe boot is horribly slow.

Comment 3 Bob Fournier 2017-08-26 00:50:17 UTC
John - any update on this?  Thanks.

Comment 4 Bob Fournier 2017-09-19 20:31:49 UTC
As noted in Comment 2, there is a known issue in Cisco PXE implementation which causes poor deploy performance.  Closing this out as it does appear to be an issue with Ironic PXE implementation.