Re: kern/56978: nvme hangs under very heavy loads

[Paul Goyette <paul%whooppee.com@localhost>]

The following reply was made to PR kern/56978; it has been noted by GNATS.

From: Paul Goyette <paul%whooppee.com@localhost>
To: gnats-bugs%netbsd.org@localhost
Cc: 
Subject: Re: kern/56978: nvme hangs under very heavy loads
Date: Sun, 28 Aug 2022 14:41:56 -0700 (PDT)

 On Sun, 28 Aug 2022, Michael van Elst wrote:
 
 > paul%whooppee.com@localhost (Paul Goyette) writes:
 >
 > > Another nvme-killer job seems to be building either lang/llvm or
 > > lang/rust from pkgsrc, with MAKE_JOBS=12.
 >
 > Two of the NVME devices are behind a PCIe switch. Is it possible to
 > test without the switch ?
 
 That would be rather difficult.  The ASUS 99X motherboard has only
 one nvme slot.  The other two nvme area mounted on a dual-nvme-to-
 pcie adapter.  All current failure scenarios are running with the
 high-load directed to one of the nvme-behind-a-switch.
 
 It might be possible to switch around the three nvme, and test with
 differing nvme in the direct-attach slot, but it would be very
 intrusive (especially if I break something, or lose one of the
 hold-down screw things).  Also would require some careful labelling
 of the nvme so all partitions get mounted regardless of the slot
 assignments.  (I know how to do that, it's just a PITA.)
 
 
 
 +--------------------+--------------------------+----------------------+
 | Paul Goyette       | PGP Key fingerprint:     | E-mail addresses:    |
 | (Retired)          | FA29 0E3B 35AF E8AE 6651 | paul%whooppee.com@localhost    |
 | Software Developer | 0786 F758 55DE 53BA 7731 | pgoyette%netbsd.org@localhost  |
 | & Network Engineer |                          | pgoyette99%gmail.com@localhost |
 +--------------------+--------------------------+----------------------+