RHEL6.3: nfsd kernel crash - long Veritas vxfs backtrace, kernel crashes and handling the crash we get stack overrun
Issue
- nfsd crashes system with Veritas vxfs in the backtrace, messages in the log include either "scheduling while atomic" or "Thread overran stack, or stack corrupted"
- System crashes in
dequeue_taskon this line:
175--> unsigned long long now = task_rq(t)->clock
- System crashed on nfsd, and loading the vmcore into the 'crash' tool we see "corrupt cpu value"
WARNING: active task ffff887ff2387500 on cpu 50: corrupt cpu value: 3826360320
Environment
- Red Hat Enterprise Linux 6.3
- Veritas modules loaded for vxfs, vxdmp, etc
- version is: 6.0.3"
vxodm(P)(U) vxgms(P)(U) amf(P)(U) vcsmm(P)(U) vxglm(P)(U) vxfen(P)(U) gab(P)(U) llt(P)(U) nfsd autofs4 nfs lockd
fscache nfs_acl auth_rpcgss sunrpc bnx2fc cnic uio fcoe libfcoe libfc dmpjbod(P)(U) dmpap(P)(U) dmpalua(P)(U)
dmpaa(P)(U) vxspec(P)(U) vxio(P)(U) vxdmp(P)(U) pcc_cpufreq bonding ipv6 8021q garp stp llc vxportal(P)(U) fdd(P)(U)
vxfs(P)(U)
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.
