> Ok, there are fresh crash files in that location. > Here is the output of the commands you suggested. > > [EMAIL PROTECTED]:/var/crash/atlantis# ls -ltr > total 1684580 > -rw-r--r-- 1 root root 2507840 2008-06-17 13:53 unix.0 > -rw-r--r-- 1 root root 374861824 2008-06-17 13:53 vmcore.0 > -rw-r--r-- 1 root root 2474509 2008-06-17 14:47 unix.1 > -rw-r--r-- 1 root root 337444864 2008-06-17 14:47 vmcore.1 > -rw-r--r-- 1 root root 2474509 2008-06-17 17:27 unix.2 > -rw-r--r-- 1 root root 336556032 2008-06-17 17:27 vmcore.2 > -rw-r--r-- 1 root root 2474509 2008-06-18 14:35 unix.3 > -rw-r--r-- 1 root root 339410944 2008-06-18 14:35 vmcore.3 > -rw-r--r-- 1 root root 2474509 2008-06-18 15:58 unix.4 > -rw-r--r-- 1 root root 322670592 2008-06-18 15:58 vmcore.4 > -rw-r--r-- 1 root root 2 2008-06-18 15:58 bounds
hmm, that opensolaris dom0 is getting frequent panics. What does mdb ::status and '$C' report for the other crash dumps? Do they all crash with a "bad mutex" panic? Does the stack backtrace look similar to the one for vmcore #4? > [EMAIL PROTECTED]:/var/crash/atlantis# echo '$C' | mdb -k 4 > ffffff0003aed710 vpanic() > ffffff0003aed740 mutex_panic+0x73(fffffffffbc0dc28, ffffff0135578130) > ffffff0003aed7b0 mutex_vector_enter+0x452(ffffff0135578130) > ffffff0003aed880 xnb_copy_to_peer+0x53(ffffff0135578000, ffffff013d342140) > ffffff0003aed8b0 xnbo_from_mac+0x1c(ffffff0135578000, 0, ffffff013d342140) > ffffff0003aed930 mac_do_rx+0xb9(ffffff0135e5b638, 0, ffffff013d342140, 0) > ffffff0003aed960 mac_rx+0x1b(ffffff0135e5b638, 0, ffffff013d342140) > ffffff0003aed9b0 vnic_rx+0x59(ffffff0133b18a88, ffffff0133b18a88, > ffffff013d342140) > ffffff0003aeda60 vnic_promisc_rx+0x12e(ffffff0133b17f58, 0, ffffff01355118c0) > ffffff0003aedac0 vnic_classifier_rx+0x43(ffffff0133b17f58, 0, > ffffff01355118c0) > ffffff0003aedb40 mac_do_rx+0xb9(ffffff0135e5cce8, 0, ffffff01355118c0, 0) > ffffff0003aedb70 mac_rx+0x1b(ffffff0135e5cce8, 0, ffffff01355118c0) > ffffff0003aedbc0 e1000g_intr_pciexpress+0x102(ffffff0135805000) Hmm, seems the dom0 nic backend driver did just receive a packet, and tries to forward that to some domU; you said that the dom0 crash happens during domU shutdown, so maybe the nic backend <-> frontend connection is already down... What is printed by: echo 'ffffff0135578000::print xnb_t' | mdb -k 4 This message posted from opensolaris.org _______________________________________________ xen-discuss mailing list [email protected]
