On Thu, Jan 15, 2026 at 04:12:01PM -0700, Bavajadas de Benadam wrote: > System is a Dell PowerEdge R630, 32 CPUs, 64G mem. Tested with > -stable. Reproduced with a kernel compiled with > > DEBUG VFSLCKDEBUG > WITNESS > WITNESS_LOCKTRACE > WITNESS_WATCH > MP_LOCKDEBUG > > Output from `dmesg` is at the end of this message. > > > ## Description > > Over ssh I ran signify on a 7.8 package directory that I had > recently rsync'd. Specifically, I ran > > `signify -C -p "/etc/signify/openbsd-78-pkg.pub" -x SHA256.sig` > > from a package directory at > `/var/www/pub/OpenBSD/7.8/packages/amd64`. After approximately two > minutes, signify in top showed a gradual decline from 90%+ cpu > activity to an idle state, where it stayed indefinitely, showing the > process as waiting on `getblk`.
You've got the disks attached to the PERC (mfii0) in this system configured as physical disks, which fails horribly as soon as anything goes wrong because mfii cannot time out or abort commands. > mfii0 at pci3 dev 0 function 0 "Symbios Logic MegaRAID SAS3108" rev 0x02: > msi > mfii0: "PERC H730 Mini", firmware 25.5.5.0005, 1024MB cache > scsibus1 at mfii0: 8 targets > scsibus2 at mfii0: 256 targets > sym0 at scsibus2 targ 2 lun 0: <SEAGATE, ST9900805SS, CS09> > naa.5000c500541df8b7 > sd0 at scsibus0 targ 0 lun 0: <SEAGATE, ST9900805SS, CS09> > naa.5000c500541df8b7 > sd0: 858483MB, 512 bytes/sector, 1758174768 sectors > sym1 at scsibus2 targ 3 lun 0: <SEAGATE, ST9900805SS, CS09> > naa.5000c5005faef177 > sd1 at scsibus0 targ 1 lun 0: <SEAGATE, ST9900805SS, CS09> > naa.5000c5005faef177 > sd1: 858483MB, 512 bytes/sector, 1758174768 sectors if you changed this to a RAID1 volume you'd probably have a better time.
