| Summary: | strange kvm log | ||
|---|---|---|---|
| Product: | Red Hat Enterprise Linux 6 | Reporter: | Levente Farkas <lfarkas> |
| Component: | qemu-kvm | Assignee: | Amit Shah <amit.shah> |
| Status: | CLOSED NOTABUG | QA Contact: | Virtualization Bugs <virt-bugs> |
| Severity: | unspecified | Docs Contact: | |
| Priority: | unspecified | ||
| Version: | 6.1 | CC: | aarcange, amit.shah, juzhang, mkenneth, mtosatti, riel, virt-maint |
| Target Milestone: | rc | ||
| Target Release: | --- | ||
| Hardware: | Unspecified | ||
| OS: | Unspecified | ||
| Whiteboard: | |||
| Fixed In Version: | Doc Type: | Bug Fix | |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | 2011-06-23 09:54:52 UTC | Type: | --- |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | Category: | --- | |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
|
Description
Levente Farkas
2011-05-26 20:27:12 UTC
As far as I can make out, the host is out of memory here so qemu kvm can't open a socket. It says guest but qemu-kvm does not run on guest - I think it's a mistake and the log triggers on host. Can you confirm please? Is the host really out of memory? Did you overcommit the host memory aggressively? this log is on the host. and no we do not overcommit the memory. there're 6 guest with 512MB and one with 2G while the host has 8GB ram. Looks like the host is receiving a (large) network packet and wants to allocate two contiguous pages of memory with the GFP_ATOMIC allocation flag, which is an allocation that cannot wait for memory to be defragmented. Levente, does the warning message you get result in any problem for your KVM guests at all? Judging from the backtrace, the likely result is just the loss of a network packet, which will no doubt be retransmitted and delivered later. the truth is that i don''t know:-( i've just found this message in the log just before we reboot the host since we've got problem with our pxe setup (but turn out that was completely different from this, and even from this server). i just report this since imho it's not a 'normal' behavior... Since the backtrace looks harmless (dropped packet) and you did not notice any malfunction, lets assume for now that nothing serious happened. Having occasional packet loss is pretty much expected on modern computers, since networking is incredibly fast and memory cannot always be freed in time for the next packets. Agree with Rik. Closing for now. If problem persists, let us know. |