Re: Dell Perc 6 disk geometry problem with RAID5 (both 6.3 final and 7.0 RC1)
- From: Tom Judge <tom@xxxxxxxxxxxx>
- Date: Sun, 20 Jan 2008 18:31:53 +0000
Erik Trulsson wrote:
On Sun, Jan 20, 2008 at 04:48:56PM +0000, David Wood wrote:[ambrisko@ and scottl@ added to CCs]
Hi there,
In message <9c1614e00801200608k49195944mf241b7b0aa6a48@xxxxxxxxxxxxxx>, Aldas Nabazas <freebsd.stable1@xxxxxxxxx> writesWe bought a new Dell PowerEdge 2950III with Perc 6/i and have the disk
geometry problem using 6.3 final or 7.0 RC1. Seems that we are not alone at
least one guy has similar problem reported earlier:
http://unix.derkeiler.com/Mailing-Lists/FreeBSD/questions/2008-01/msg00506.html
This post is related to using BSDLabel and FDISK MBR partition tables with a disk over 2Tb. This is a known limitation of this type of partition table where the max size is 2Tb. If you wish to make a larger partition you must use GPT or the raw device with no partition table (Not recommended!!!!).
I was reading the mailing list and found that some of the people are happily
using this hardware with the latest FreeBSD:
http://lists.freebsd.org/pipermail/freebsd-stable/2008-January/039675.html
I just wonder what the status of mfi driver? Maybe it's not fully tested or
there will be some important fixes before 7.0 final? We are going to try
different RAID combinations but it definitely not working using 6x146GB as
RAID5.
I do not know if the mfi(4) driver has any problems with large disks, but it
is however well known that fdisk(8) and bsdlabel(8) (the tools normally used
to partition disks) have problems with volumes larger than 2TB.
If you want to use volumes larger than 2TB then gpt(8) is the recommended
way to partition the disks. It is however doubtful if the BIOS in your
system will allow you to boot from a gpt(8) parttioned volume which is
best solved with having a separate - smaller - boot volume where the OS
itself is installed.
I have 5 PERC6/i based systems awaiting deployment atm, might be able to do some very limited testing on one of them if time permits. However I have some very large arrays running on PERC5/e's (6TB Raid 50 - 5 disk spans - 15 * 500Gb disks spread over 2 md1000 shelves) (the systems run RELENG_6_2) and have not seem any issues with them. They are a single gpt partition with UFS2 on them..... As far as I have seen in the last ~8 months there have been no issues with the mfi driver and large arrays. However I cannot say at this stage if the same can be said for the PERC6/i.....
This is extremely disappointing to read, as I was relatively close to buying a Poweredge 2950 III with PERC 6/i. However, it's no good to me if mfi(4) has issues with large virtual disks; the tentative disk configuration is 2x146GB as RAID 1 and 4x750GB (or 1TB) as RAID 5.
Looking at CVSweb:
http://www.freebsd.org/cgi/cvsweb.cgi/src/sys/dev/mfi/
There is one related PCI ID's change to do with DELL sub vendor id's I think that is only on HEAD. Any chance of getting this MFC'd?
Tom
the updates for the 1078 chip which powers the PERC 6 series were contributed by LSI, so you would have hoped things were right. There is a disclaimer on the code, but you would also hope that someone in the know is testing it, especially as my impression has always been that the FreeBSD community is favourably disposed towards LSI storage controllers and that LSI and their vendors try to help the FreeBSD developers.
Maybe someone could share the RAID combinations they successfully areI haven't been keeping a very close eye on the problem as I don't currently have any hardware - but is the issue simply one of virtual disk size - there's a cut-off size after which things don't work properly?
running on?
You could try pulling disks from your server (or removing them from the virtual disk) one by one until things start to work. To save a lot of pain you could create a virtual disk containing one disk as RAID 0 (a single disk) and install the OS on it, leaving the other five disks to play with.
I do hope that someone is in a position to investigate this quickly - even if it's too late to get the fix in 7.0-RELEASE now. There's nothing that I can see as relevant that's waiting for MFC, anyway.
Of course, it's worth checking whether you've got the latest firmware on the PERC 6/i - the latest version from Dell appears to be 6.0.1-0080.
If this doesn't get fixed soon, I'm either going to have to go to another hardware vendor that uses a different RAID controller (HP is a possibility - though we're an all Dell shop) or - sniff - leave FreeBSD in favour of a Linux distribution. I realise that FreeBSD is a volunteer project, and that users can't have any specific expectations - this is not a threat, as I like FreeBSD and want to remain in the community, but I would need a new server to work properly!
Looking at CVSweb, it seems that ambrisko@ and scottl@ are the two people most closely associated with the mfi code - I've added them into the CCs. My apologies if that was unwelcome.
_______________________________________________
freebsd-stable@xxxxxxxxxxx mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-stable
To unsubscribe, send any mail to "freebsd-stable-unsubscribe@xxxxxxxxxxx"
- References:
- Dell Perc 6 disk geometry problem with RAID5 (both 6.3 final and 7.0 RC1)
- From: Aldas Nabazas
- Re: Dell Perc 6 disk geometry problem with RAID5 (both 6.3 final and 7.0 RC1)
- From: David Wood
- Re: Dell Perc 6 disk geometry problem with RAID5 (both 6.3 final and 7.0 RC1)
- From: Erik Trulsson
- Dell Perc 6 disk geometry problem with RAID5 (both 6.3 final and 7.0 RC1)
- Prev by Date: Re: Scheduler in Various Docs
- Next by Date: Re: Dell Perc 6 disk geometry problem with RAID5 (both 6.3 final and 7.0 RC1)
- Previous by thread: Re: Dell Perc 6 disk geometry problem with RAID5 (both 6.3 final and 7.0 RC1)
- Next by thread: Re: Dell Perc 6 disk geometry problem with RAID5 (both 6.3 final and 7.0 RC1)
- Index(es):
Relevant Pages
|
|
