Bug 1210715
| Summary: | migration/rdma: 7.1->7.2: RDMA ERROR: ram blocks mismatch #3! | ||||||
|---|---|---|---|---|---|---|---|
| Product: | Red Hat Enterprise Linux 7 | Reporter: | Dr. David Alan Gilbert <dgilbert> | ||||
| Component: | qemu-kvm-rhev | Assignee: | Dr. David Alan Gilbert <dgilbert> | ||||
| Status: | CLOSED ERRATA | QA Contact: | Virtualization Bugs <virt-bugs> | ||||
| Severity: | unspecified | Docs Contact: | |||||
| Priority: | urgent | ||||||
| Version: | 7.2 | CC: | dgilbert, hhuang, huding, juzhang, mrezanin, qzhang, virt-maint, xfu | ||||
| Target Milestone: | rc | Keywords: | ZStream | ||||
| Target Release: | --- | ||||||
| Hardware: | Unspecified | ||||||
| OS: | Unspecified | ||||||
| Whiteboard: | |||||||
| Fixed In Version: | qemu-kvm-rhev-2.3.0-13.el7 | Doc Type: | Bug Fix | ||||
| Doc Text: | Story Points: | --- | |||||
| Clone Of: | |||||||
| : | 1242032 1248382 (view as bug list) | Environment: | |||||
| Last Closed: | 2015-12-04 16:35:58 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: | 1174920, 1242032, 1248382 | ||||||
| Attachments: |
|
||||||
hmm, the RDMA code seems to be keying off the 'offset' which I think is the offset in RAMBlock address space - that's not in any way intended to stay the same across versions
'mismatch #3' is when qemu_rdma_registration_stop failed to find a matching RAMBlock.
Just confirmed this also affects 7.1 by itself with hotplug, hotplug a device on the source, but have it on the command line on the destination and it fails in a similar way. Fix included in qemu-kvm-rhev-2.3.0-13.el7 source host: qemu-kvm-1.5.3-60.el7.x86_64 Reproduced with qemu-kvm-rhev-2.3.0-8.el7.x86_64 on target host: (qemu) migrate -d x-rdma:192.168.0.21:6666 source_resolve_host RDMA Device opened: kernel name mlx4_0 uverbs device name uverbs0, infiniband_verbs class device path /sys/class/infiniband_verbs/uverbs0, infiniband class device path /sys/class/infiniband/mlx4_0, transport: (1) Infiniband (qemu) (qemu) (qemu) (qemu) (qemu) info migrateRDMA ERROR: ram blocks mismatch #1! Your QEMU command line parameters are probably not identical on both the source and destination. Verify failed with qemu-kvm-rhev-2.3.0-13.el7.x86_64 on target host: (qemu) migrate -d x-rdma:192.168.0.21:6666 source_resolve_host RDMA Device opened: kernel name mlx4_0 uverbs device name uverbs0, infiniband_verbs class device path /sys/class/infiniband_verbs/uverbs0, infiniband class device path /sys/class/infiniband/mlx4_0, transport: (1) Infiniband (qemu) RDMA ERROR: ram blocks mismatch #1! Your QEMU command line parameters are probably not identical on both the source and destination. David, could you have a look? Retest with qemu-kvm-rhev-2.1.2-23.el7_1.6.x86_64 on source host, result are the same. 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://rhn.redhat.com/errata/RHBA-2015-2546.html |
Created attachment 1013118 [details] xml for broken vm Description of problem: This VM migrates fine usign TCP but with RDMA it fails Version-Release number of selected component (if applicable): source: qemu-kvm-rhev-2.1.2-23.el7_1.1.x86_64 dest: qemu-kvm-rhev-2.3.0-0.el7.dgilbert20150410a.x86_64 How reproducible: 100% Steps to Reproduce: 1. See xml attached; 7.1 machine type vm started on 7.1 2. migrate to 7.2 using RDMA - virsh migrate --live --migrateuri rdma://ibpair f20-414 qemu+ssh://ibpair/system --listen-address 192.168.99.13 Actual results: Migration fails, error seen in source logs: RDMA ERROR: ram blocks mismatch #3! Your QEMU command line parameters are probably not identical on both the source and destination. Expected results: A working migration Additional info: