Bug 98462
| Summary: | bonding TLB load sharing fails under heavy UDP Tx stress | ||
|---|---|---|---|
| Product: | Red Hat Enterprise Linux 3 | Reporter: | Need Real Name <shmulik.hen> |
| Component: | kernel | Assignee: | Jeff Garzik <jgarzik> |
| Status: | CLOSED RAWHIDE | QA Contact: | Brian Brock <bbrock> |
| Severity: | medium | Docs Contact: | |
| Priority: | medium | ||
| Version: | 3.0 | CC: | peterm, riel, tao |
| Target Milestone: | --- | ||
| Target Release: | --- | ||
| Hardware: | i686 | ||
| OS: | Linux | ||
| URL: | http://sourceforge.net/projects/bonding/ | ||
| Whiteboard: | |||
| Fixed In Version: | 2.4.21-1.1931.2.349.2.2.ent | Doc Type: | Bug Fix |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | 2003-08-03 13:54:50 UTC | Type: | --- |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | Category: | --- | |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
| Embargoed: | |||
ISSUE TRACKER 25886 opened as sev 1 Jeff, does Taroon already have the patch for this or is it still in your queue ? Appears to be fix implemented in RHEL 3 B1 candidate kernel (version 2.4.21- 1.1931.2.349.2.2.ent). |
Description of problem: When running very heavy UDP Tx stress traffic with 10/100 adapters, load- sharing collapses to only one slave after a few seconds. Caused by a unsigned/signed cast error in the TLB code. Version-Release number of selected component (if applicable): kernel-2.4.20-1.1931.2.231.2.11.ent How reproducible: Configure a bond team with only 10/100 adapters and run very heavy UDP Tx stress traffic to many clients. Monitor Tx/Rx activity of the slaves. Steps to Reproduce: 1. insmod bonding mode=5 2. ifconfig bond0 <ip-addr> 3. ifenslave bond0 eth0 eth1 eth2 4. start stress application (e.g iperf, netperf, etc.) Actual results: After a few seconds only one slave takes part of load sharing while others stay idle. Traffic may pass from slave to slave at 10 sec. intervals (re-balance timeout). Expected results: All slaves continuously take part of the load sharing. Additional info: A bug fix patch was sent by me on June 26th to bond-devel, linux-net and linux- netdev lists. It was already accepted by Jeff Garzik into his net-drivers-2.4 BK tree.