[OpenIndiana-discuss] ZFS hangs - causes host to panic

[Stephan Budach]

Hi, 


I am currently putting my new NVME servers through their paces and I already experinced two panics on one of those hosts. 
After taking "forever" writing the crash dump, I found this in the syslog after reboot: 



Jan 16 13:25:29 nfsvmpool09 savecore: [ID 570001 auth.error] reboot after panic: I/O to pool 'nvmeTank02' appears to be hung. 
Jan 16 13:25:29 nfsvmpool09 savecore: [ID 771660 auth.error] Panic crashdump pending on dump device but dumpadm -n in effect; run savecore(1M) manually to extract. Image UUID 995846d5-8c94-4f68-bada-e05ae5e4cb25(fault-management initiated). 


I ran mdb against the crash dump, but I am still a dummy, reading those information: 



root at nfsvmpool09:/var/crash/nfsvmpool09# mdb unix.0 vmcore.0 
Loading modules: [ unix genunix specfs dtrace mac cpu.generic uppc apix scsi_vhci zfs sata sd ip hook neti sockfs arp usba fctl stmf stmf_sbd mm lofs i40e idm cpc crypto fcip fcp random ufs logindmux nsmb ptm smbsrv nfs sppp ipc ] 
> $C 
ffffd000f5dd79d0 vpanic() 
ffffd000f5dd7a20 vdev_deadman+0x10b(ffffd0320fb69980) 
ffffd000f5dd7a70 vdev_deadman+0x4a(ffffd0333b018940) 
ffffd000f5dd7ac0 vdev_deadman+0x4a(ffffd03228f796c0) 
ffffd000f5dd7af0 spa_deadman+0xad(ffffd03229543000) 
ffffd000f5dd7b90 cyclic_softint+0xfd(ffffd031eac4db00, 0) 
ffffd000f5dd7ba0 cbe_low_level+0x14() 
ffffd000f5dd7bf0 av_dispatch_softvect+0x78(2) 
ffffd000f5dd7c20 apix_dispatch_softint+0x35(0, 0) 
ffffd000f5da1990 switch_sp_and_call+0x13() 
ffffd000f5da19e0 apix_do_softint+0x6c(ffffd000f5da1a50) 
ffffd000f5da1a40 apix_do_interrupt+0x362(ffffd000f5da1a50, 2) 
ffffd000f5da1a50 _interrupt+0xba() 
ffffd000f5da1bc0 acpi_cpu_cstate+0x11b(ffffd031e98a43e0) 
ffffd000f5da1bf0 cpu_acpi_idle+0x8d() 
ffffd000f5da1c00 cpu_idle_adaptive+0x13() 
ffffd000f5da1c20 idle+0xa7() 
ffffd000f5da1c30 thread_start+8() 
> 


Can anybody make something useful of that? 


Thanks, 
Stephan