Skip site navigation (1)Skip section navigation (2)
Date:      Sat, 9 May 2015 13:24:51 +0300
From:      Mykola Golub <to.my.trociny@gmail.com>
To:        Daisuke Aoyama <jack@st.rim.or.jp>
Cc:        freebsd-current@freebsd.org
Subject:   Re: hastd fail and panic on MAXPHYS=1m
Message-ID:  <20150509102450.GA35014@gmail.com>
In-Reply-To: <8A0CADA1C80C4310BFFA1828554F473C@ad.peach.ne.jp>
References:  <8A0CADA1C80C4310BFFA1828554F473C@ad.peach.ne.jp>

next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, May 08, 2015 at 06:02:03AM +0900, Daisuke Aoyama wrote:
> Hi all,
> 
> I have problem with MAXPHYS=1m. (I don't know MAXPHYS=1m works on HAST.)
> 
> I put "options MAXPHYS=(1024*1024)" in kernel config.
> Then, update primary node to the kernel and world.
> If the role back to primary on the machine, writing to the hast device cause an error and panic.
> 
> I didn't check carefully, but it seems that geom_gate.ko use MAXPHYS=1m and hastd use 
> MAXPHYS=128k.
> Of course, secondary is MAXPHYS=128k at this time.
> 
> Is it an expected result?

Putting "options MAXPHYS..." in kernel config does not change the
value for userspace (hastd). You might want to try adding to
make.conf:

CFLAGS += -DMAXPHYS=1048576

and rebuild userspace (or just hastd).

Also, running the secondary with smaller MAXPHYS will likely fail
too. If you can't reinstall simultaneously I suggest running only the
updated primary (don't starting the secondary) until the secondary is
updated.

-- 
Mykola Golub



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20150509102450.GA35014>