I have a backup server with almost no I/O activity on which I want to export a ZPOOL to a USB disk. When I do a 'zfs send my_backup_pool| zfs receive my_usb_backup_pool', the process hangs after a few minutes and dmesg reports this error message: mfi1: COMMAND 0xffffff80004fe268 TIMEOUT AFTER 59 SECONDS mfi1: COMMAND 0xffffff80004fbcb0 TIMEOUT AFTER 59 SECONDS mfi1: COMMAND 0xffffff80004fbcb0 TIMEOUT AFTER 89 SECONDS mfi1: COMMAND 0xffffff80004fbcb0 TIMEOUT AFTER 120 SECONDS .. The problem solves as soon as I do a read on the source. Is it possible that 'zfs send' bypasses some low level layers that prevents disks to suspend ? How-To-Repeat: If you do this on a server with no I/O activity: zfs send a_big_zpool | cat - > /dev/null after a few minutes, you should have: mfi1: COMMAND 0xffffff80004fe268 TIMEOUT AFTER 59 SECONDS mfi1: COMMAND 0xffffff80004fbcb0 TIMEOUT AFTER 59 SECONDS mfi1: COMMAND 0xffffff80004fbcb0 TIMEOUT AFTER 89 SECONDS mfi1: COMMAND 0xffffff80004fbcb0 TIMEOUT AFTER 120 SECONDS .. And as soon as you do a 'ls' for example on the zpool, 'zfs send' continues.
Responsible Changed From-To: freebsd-bugs->freebsd-fs Over to maintainer(s).
Responsible Changed From-To: freebsd-fs->freebsd-hardware Pawel says this might be hardware related, assign it to that group and tag the ticket as [mfi].
State Changed From-To: open->feedback What MFI controller is this and what firmware version is it?
As a followup in the originator stead > # mfiutil show firmware > mfi0 Firmware Package Version: 12.10.0-0025 > mfi0 Firmware Images: > Name Version Date Time Status > BIOS 3.18.00_4.09.05.00_0x0416A000 00_0x0416A000 active > APP 2.100.03-1046 Oct 24 2010 04:38:35 active > PCLI 04.04-010:#%00008 May 31 2010 20:21:52 active > CTLR 2.02-0025 Aug 12 2010 14:46:17 active > NVDT 2.07.03-0003 Jul 14 2010 15:53:29 active > BTBL 2.02.00.00-0000 Sep 16 2009 21:37:06 active > BOOT 01.250.04.219 4/28/2009 12:51:38 active > > > # mfiutil show adapter > mfi0 Adapter: > Product Name: PERC H700 Integrated > Serial Number: 0BC0310 > Firmware: 12.10.0-0025 > RAID Levels: JBOD, RAID0, RAID1, RAID5, RAID6, RAID10, RAID50 > Battery Backup: present > NVRAM: 32K > Onboard Memory: 512M > Minimum Stripe: 8k > Maximum Stripe: 1M
Reassign from non-responding assignee. To submitter: does this problem still recur on recent versions of FreeBSD?