linux-kernel - Re: Have the velociraptors in a test system now, checkout the errors.

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <493A5E62.1020508@msgid.tls.msk.ru>
Date:	Sat, 06 Dec 2008 14:13:38 +0300
From:	Michael Tokarev <mjt@....msk.ru>
To:	Justin Piszcz <jpiszcz@...idpixels.com>
CC:	linux-kernel@...r.kernel.org, linux-raid@...r.kernel.org,
	xfs@....sgi.com, smartmontools-support@...ts.sourceforge.net
Subject: Re: Have the velociraptors in a test system now, checkout the errors.

Justin Piszcz wrote:
> Point of thread: Two problems, mentioned in detail below, NCQ in Linux
[]

> Two-disks failed out of the RAID5 and I currentlty cannot even 'see' one
> of the drives with smartctl, will reboot the host and check sde again.

Not to say it's your case (i remember you mentioned a new powerful PSU
in your last emails), but anyway.

We had numerous, countless disk failures like this in the past - with
seagate scsi drives (not sas, not sata but ol'good scsi - 9Gb and 36Gb
barracuda ones).  As that - a drive suddenly disappears from the bus,
without any indication it was/is here, only power-cycle cures the prob.

One such case were related to a broken (it seems) batch of those 9Gb
drives (it was back in 2000 or so).  The frequency of such failures
fluctuated a lot, and did not depend on system load - it was possible
to see disk disappearance after a few mins after boot without any load,
or it may run for several weeks under a good load.  The failing drive
was always the same, replace it and voila, it works again.  There was
about 10..20 such drives we had, some are still here somewhere (not in
use).

And another case was with 36gb 10krpm barracudas, at about 2004 or so.
And also with 18gb 15Krpm maxtors.  Some of them.

This case looked really mysterious to me.  Until I found (after many
many times experimenting with all that) that the cause is under-powered
PSU.  For example, when there were 2 disks running on the system, no
hdd stopped, but with 4 disks rinning, one were quite likely to stop
(always the same, other disks were working still).  When this new
problem started appearing and I had not yet understand the cause,
we also tried to replace the "failing" drives, and it helped somewhat, --
i.e., there were high chances that the replacement disk will actually
work better.  But some non-zero chance existed that it will not work
the same or even worse way the "failing" drive failed.

It come to good surprize to me that the problem was the PSU.  It was
350W (quite descent in 2002 when the test system was bought), but
obviously not enough for the load with all the 15krpm drives...
(and later on the system become instable too, and now I know why -
also lack of proper power, now for chipset/cpu).

The prob with 9gb drives were real (not due to the PSU), but
Seagate never acknowleged it.

Just... another "funny" scenario which happened for real.
And my probs obviously were NOT related to NCQ (TCQ really) -
TCQ worked on all those drives just fine, much better and
with much better effect than all thouse modern NCQ-aware
drives.....  Ohwell.

/mjt
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/