Bug 697289

Summary: libvirtd crash after minutes/hours of responding to loads of virsh domstat
Product: [Community] Virtualization Tools Reporter: masoud
Component: libvirtAssignee: Libvirt Maintainers <libvirt-maint>
Status: CLOSED DEFERRED QA Contact:
Severity: urgent Docs Contact:
Priority: unspecified    
Version: unspecifiedCC: crobinso, xen-maint
Target Milestone: ---   
Target Release: ---   
Hardware: x86_64   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2016-03-23 12:35:40 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Attachments:
Description Flags
gdb generated log of issue none

Description masoud 2011-04-17 14:12:38 UTC
Created attachment 492703 [details]
gdb generated log of issue

Description of problem:
libvirt crash when you run loads of "virsh domstat vm_name" 

Version-Release number of selected component (if applicable):
0.9.0 (and all older versions)

How reproducible:
It's easily reproducible with fedora core 14 and libvirt-0.9.0 

Steps to Reproduce:
1.Install a fedora core 14 and configure libvirt-0.9.0  with netcf
2.Run some vms, and let some other vms be off
3.Put vms' nics on different bridge and isolated network
4.save this script to a file and name it putload:
#!/bin/bash
vms=( vm1 vm2 vm3 vm4 vm5 )
for i in `seq 10000`;
do
        sleep 1
        for vm in "${vms[@]}";
        do
                virsh domstate $vm
        done
done
5. run putload script more than 5 times at the same time like this
./putload & sleep 1 && ./putload & sleep 1 && ./putload & sleep 1 && ./putload & sleep 1 && ./putload & sleep 1 && ./putload 

6. some times after some minutes, and sometime after some hours libvirtd crash
NOTICE: be patient some times it takes more than 12 hours to crash.

  
Actual results:
libvirtd crash

Expected results:
libvirtd should be stable and not crash at all

Additional info:
I think it's related to netcf and fedora, because I haven't had such problem with debian and ubuntu as they don't use netcf. I attached gdb.txt to this bug report.
It's really urgent because you can't run any monitoring and HA solution with virsh and libvirtd with such a bug, as you need to also watch libvirtd daemon.

Comment 1 Cole Robinson 2016-03-23 12:35:40 UTC
Sorry this never received a response. I don't know about that particular crash but I assume it has been fixed. Closing as DEFERRED, but if anyone can still reproduce with newer libvirt please reopen with recent gdb output