Date: Mon, 2 Apr 2012 09:55:20 +1000 From: Petro Rossini <petro.rossini@gmail.com> To: freebsd-virtualization@freebsd.org Subject: VirtualBox drama - failing VDI disks, VM not starting headless, Rebuild problems on 8.2 (clang related problem?) and 9-STABLE (libpcre.so.0) Message-ID: <CAExkqpf8MK=zbFo5kfGQg0VfSjjSPKRQ0s%2B07aGKX6bArZmPHQ@mail.gmail.com>
next in thread | raw e-mail | index | archive | help
Hi all, I had some VirtualBox hassle over the weekend. It started with thes messages inside the VM, a Ubuntu 10.04 Server (running Zimbra). 2012-03-30T18:37:11.581591+11:00 mail kernel: [4520695.304357] sd 4:0:0:0: [sdd] Unhandled error code 2012-03-30T18:37:15.201744+11:00 mail kernel: [4520695.304362] sd 4:0:0:0: [sdd] Result: hostbyte=DID_OK driverbyte=DRIVER_TIMEOUT 2012-03-30T18:37:15.201753+11:00 mail kernel: [4520695.304366] sd 4:0:0:0: [sdd] CDB: Write(10): 2a 00 02 1d 44 f8 00 00 08 00 2012-03-30T18:37:15.201756+11:00 mail kernel: [4520695.304457] end_request: I/O error, dev sdd, sector 35472632 2012-03-30T18:37:15.201758+11:00 mail kernel: [4520695.312601] __ratelimit: 9 callbacks suppressed 2012-03-30T18:37:15.201761+11:00 mail kernel: [4520695.312604] Buffer I/O error on device sdd, logical block 4434079 2012-03-30T18:37:15.201763+11:00 mail kernel: [4520695.321978] lost page write due to I/O error on sdd Two hours later it moves to another disk: "sd 3:0:0:0: [sdc] Unhandled error code" The Linux reacts with a read-only remount - practically the box is falling over. The VDI disks are on a ZFS filesystem, and I scrubbed the ZVolume and looked for other possible problems - host-side all seamt to be okay. I restarted the system once - and for a few hours it worked, untiul I see the same problem again. A second reboot is not that successful, I cannot start the VM headless anymore, getting a Result Code: NS_ERROR_FAILURE (0x00004005) I have a second machine with the same setup and continous mirroring of the disks (zfs send/receive) - the same result. Tried an older ZFS snapshot of the VDI disks - still not starting. It was FreeBSD 8.2 and VirtualBox 4.0.4, and some people with similar problems when the VirtualBox kernel module and system where outy of sync. I did some port upgrades recently but left VirtualBox alone. Hmmh.. Anyway, I started upgading to the newest VirtualBox in the ports(4.1.18). Under 8.2 the rebuild does not seem to work anymore, it has some clang related compile errors (apologize that I do not have the detail, it is a dodgy macro with "clang" in it). So I upgraded the base system to 9-STABLE first and was able to build VirtualBox 4.1.18 now. (with a minor hickup, libpcre.so.0 not found, I just symlinked .1 against ist) Still the same error: Result Code: NS_ERROR_FAILURE (0x00004005) So I decided to go home to get access to the X11-GUI (on the road I had ssh only). I thought of rebuilding a new VM with the same disks (because of some forum suggestions the config could be corrupted). To my astonishment the system started! And now I can start it headless again! I am at loss to explain it, maybe someone else may have a clue? Regards Peter
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAExkqpf8MK=zbFo5kfGQg0VfSjjSPKRQ0s%2B07aGKX6bArZmPHQ>