Solaris 10 T2 series servers without FW 7.4.4.f or later can hang in shutdown due to a stuck fmd thread

Asset ID:	1-72-1637803.1
Update Date:	2014-03-24
Keywords:

Solution Type Problem Resolution Sure

Solution 1637803.1 : Solaris 10 T2 series servers without FW 7.4.4.f or later can hang in shutdown due to a stuck fmd thread

Applies to:

Sun Netra T5220 Server - Version All Versions and later
Solaris SPARC Operating System - Version 10 3/05 and later
Sun SPARC Enterprise T5220 Server - Version All Versions and later
Sun SPARC Enterprise T5120 Server - Version All Versions and later
Sun SPARC Enterprise T5140 Server - Version All Versions and later
Oracle Solaris on SPARC (64-bit)

Symptoms

Solaris 10 T2 servers without FW 7.4.4.f or later can hang during shutdown due to a stuck fmd thread.

From a forced crash dump, we can see that a stuck fmd thread which is waiting in ds_pro_ioctl() is holding off the shutdown.

> ::ptree
00000000018886c0 sched
     0000060010aa6c28 fsflush
     0000060010aa7848 pageout
     0000060010aa8468 init
          000006001ec0e470 automountd
               000003005568a478 automountd
          000003007404cc68 rpcbind
          0000030055674c40 in.mpathd
          0000030055675860 kcfd
          0000060010aa6008 syseventd
          00000300608f90a8 devfsadm
          000006001ec0d850 svc.configd
          000006001ec0f090 svc.startd
               0000030075a88058 fmd
                    0000030075d78068 fmd <<<<<

> 30075d78068::ps -t
S    PID   PPID   PGID    SID    UID      FLAGS             ADDR NAME
R    624    595    624    624      0 0x42000000 0000030075d78068 fmd
        T     0x300605a8b60 <TS_SLEEP>   <<<<<<
        T     0x30073af2740 <TS_SLEEP>

> 0x300605a8b60::findstack
stack pointer for thread 300605a8b60: 2a10205af51
[ 000002a10205af51 cv_wait_sig+0x114() ]
000002a10205b001 ds_pri_ioctl+0xa0()
000002a10205b0e1 fop_ioctl+0x20()
000002a10205b191 ioctl+0x184()
000002a10205b2e1 syscall_trap32+0xcc()

This issue affects all T2 servers without FW 7.4.4.f or later.

Cause

Bug 15705246 - SUNBT7031127 Cluster node hung during shutdown as fmd thread stuck in ds_pri_ope

Solution

This issue is fixed by System Firmware 7.4.4.f or later in the following patches

147307-07 or later for SPARC Enterprise T5120+T5220
147308-05 or later for Sun Blade T6320+T6320-G2 Server Module
147309-06 or later for Netra T5220
147310-07 or later for SPARC Enterprise T5140+T5240
147311-05 or later for SPARC Enterprise T5440
147312-06 or later for Sun Blade T6340 Server Module
147313-06 or later for Netra T5440
147314-06 or later for Netra T6340

References

<BUG:15705246> - SUNBT7031127 CLUSTER NODE HUNG DURING SHUTDOWN AS FMD THREAD STUCK IN DS_PRI_OPE

Attachments

This solution has no attachment