Bug 1327377
Summary: | hotunplug scsi device will cause qemu coredump | ||||||
---|---|---|---|---|---|---|---|
Product: | Red Hat Enterprise Linux 7 | Reporter: | Yang Meng <meyang> | ||||
Component: | qemu-kvm-rhev | Assignee: | Markus Armbruster <armbru> | ||||
Status: | CLOSED DUPLICATE | QA Contact: | Virtualization Bugs <virt-bugs> | ||||
Severity: | unspecified | Docs Contact: | |||||
Priority: | unspecified | ||||||
Version: | 7.3 | CC: | aliang, chayang, huding, juzhang, knoel, meyang, ngu, pezhang, pingl, shuang, virt-maint, xuwei | ||||
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: | 2016-09-09 09:13:33 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: | |||||||
Attachments: |
|
Description
Yang Meng
2016-04-15 02:51:12 UTC
core file location: http://fileshare.englab.nay.redhat.com/pub/section2/images_backup/rhel7/bug1327377/ Could be a duplicate of bug 1318181. I suspect this is duplicate of bug 1341531. We fixed that one in qemu-kvm-rhev-2.6.0-12.el7. Could you please retest this bug with that version? If it appears to be fixed there, also testing the version before would be nice. hi, you can refer to this bug : https://bugzilla.redhat.com/show_bug.cgi?id=1362084 #comment 1 i copy the comment from the author(areis) blockdev-add API is still experimental upstream and not supported by libvirt. But instead of closing this BZ, I'm reassigning to Kevin for further evaluation, while deferring it to 7.4. Yes, blockdev-add is experimental, but the stack backtrace shows a crash in drive_del. It might be reproducible even without blockdev-add. We've recently fixed a crash bug in drive_del (bug 1341531) and a related crash bug in drive_add (bug 1352865). Let's exclude them before we dig deeper. Please retest with qemu-kvm-rhev-2.6.0-16.el7 (contains both fixes). If it works, retest with qemu-kvm-rhev-2.6.0-12.el7 (just the drive_del fix). If it works, retest with qemu-kvm-rhev-2.6.0-11.el7. As usual, provide a stack backtrace when you observe a crash. (In reply to Markus Armbruster from comment #13) > Yes, blockdev-add is experimental, but the stack backtrace shows a crash in > drive_del. It might be reproducible even without blockdev-add. > > We've recently fixed a crash bug in drive_del (bug 1341531) and a related > crash bug in drive_add (bug 1352865). Let's exclude them before we dig > deeper. Please retest with qemu-kvm-rhev-2.6.0-16.el7 (contains both > fixes). If it works, retest with qemu-kvm-rhev-2.6.0-12.el7 (just the > drive_del fix). If it works, retest with qemu-kvm-rhev-2.6.0-11.el7. As > usual, provide a stack backtrace when you observe a crash. the reproduce results is attached as a file: all of the version i tried will also got the crash. the coredump files: http://fileshare.englab.nay.redhat.com/pub/section2/images_backup/rhel7/meyang-blockdev-add/ Created attachment 1198944 [details]
reproduce-log
The crashes observed in comment#14 are almost certainly bug 1362084, as Yang Meng suggested in comment#12. The difference is that the invalid input triggering the crash is '"detect-zeroes": true' for bug 1362084, and "'writeback': false" here. But now I wonder about comment#0... I'm pretty sure qemu-kvm-rhev-2.5.0-4.el7.x86_64 also crashes on "'writeback': false". How could it ever get past that and crash in drive_del? Let's fix bug 1362084 first, then see whether we can still reproduce anything bad here. Yang Meng kindly verified that with bug 1362084 fixed, the incorrect blockdev-add command is rejected cleanly. I can't reproduce the crash reported in comment#0 step 4 __com.redhat_drive_del, because as far as I can tell, all versions of qemu-kvm either crash or reject in step 3 blockdev-add. I tried omitting the invalid part of blockdev-add ('writeback': false), no dice. But perhaps I'm doing something wrong. Yang Meng, could you double-check for me? Use { "execute": "blockdev-add", "arguments": {'options' : {'driver': 'raw', 'id':'drive-disk1', 'discard':'unmap', 'rerror':'stop', 'werror':'stop', 'file': {'driver': 'host_device', 'filename': '/dev/sdb'}, 'cache': { 'direct': true, 'no-flush': false }}} } If this works now, we can close the bug. (In reply to Markus Armbruster from comment #20) > Yang Meng kindly verified that with bug 1362084 fixed, the incorrect > blockdev-add command is rejected cleanly. > > I can't reproduce the crash reported in comment#0 step 4 > __com.redhat_drive_del, because as far as I can tell, all versions of > qemu-kvm either crash or reject in step 3 blockdev-add. > > I tried omitting the invalid part of blockdev-add ('writeback': > false), no dice. But perhaps I'm doing something wrong. > > Yang Meng, could you double-check for me? Use > > { "execute": "blockdev-add", "arguments": {'options' : {'driver': 'raw', > 'id':'drive-disk1', 'discard':'unmap', 'rerror':'stop', 'werror':'stop', > 'file': {'driver': 'host_device', 'filename': '/dev/sdb'}, 'cache': { > 'direct': true, 'no-flush': false }}} } > > If this works now, we can close the bug. hi, using your command : { "execute": "blockdev-add", "arguments": {'options' : {'driver': 'raw', 'id':'drive-disk1', 'discard':'unmap', 'rerror':'stop', 'werror':'stop', 'file': {'driver': 'host_device', 'filename': '/dev/sdb'}, 'cache': { 'direct': true, 'no-flush': false }}} } the hotplug works fine. i just tried your test build in #comment 17 Thanks, Yang Meng! This bug is certainly a duplicate of bug 1362084 and probably also a duplicate of bug 1341531. Closing as duplicate of the former. *** This bug has been marked as a duplicate of bug 1362084 *** |