HP ML110 G6.. Raid 0+1 nfg?
Not sure if this is supported.. Looks like the card is an HP Smart Array B110i.. But when setup as raid 0+1 in the bios.. FreeBSD 8.0 amd64 says it can not find any disks.. Anyone have this working? Thanks in advance. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
Re: Data Redundancy RAID 0+1 Vs 1+0 FREEBSD 7.4 STABLE
Jeronimo Calvo wrote: Hi folks, I'm thinking to build a Raid on my system and I'm getting documented abut 0+1 and 1+0 RAID systems. As far as I can see the best option is a 1+0 is the best option as if one of the from mirrors fails, the RAID still be redundant and in the case of a 0+1 RAID, If one of them fails, the RAID will be down until u replace the HD in question... So my question is, if I am right above, what are de benefits of mounting a RAID 0+1 ?? For workloads that involve streaming large volumes of sequential data RAID 0+1 can perform very well. However, for the typical sort of workloads seen on a general purpose workstation, it offers no advantages over RAID10, and for the sort of workload you get with RDBMSes -- lots of randomly scattered small IOs -- RAID10 does the business. Given the poor resilience characteristics of RAID 0+1 the sequential data streaming workload would be better handled by RAID5(0) or RAID6(0) plus a good hardware RAID controller with plenty of battery backed cache RAM. Cheers, Matthew -- Dr Matthew J Seaman MA, D.Phil. 7 Priory Courtyard Flat 3 PGP: http://www.infracaninophile.co.uk/pgpkey Ramsgate Kent, CT11 9PW signature.asc Description: OpenPGP digital signature
Data Redundancy RAID 0+1 Vs 1+0 FREEBSD 7.4 STABLE
Hi folks, I'm thinking to build a Raid on my system and I'm getting documented abut 0+1 and 1+0 RAID systems. As far as I can see the best option is a 1+0 is the best option as if one of the from mirrors fails, the RAID still be redundant and in the case of a 0+1 RAID, If one of them fails, the RAID will be down until u replace the HD in question... So my question is, if I am right above, what are de benefits of mounting a RAID 0+1 ?? Cheers! -- () ASCII Ribbon Campaign | Against HTML e-mail /\ www.asciiribbon.org | Against proprietary extensions ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
RE: RAID 0+1
> -Original Message- > From: Nejc Škoberne [mailto:[EMAIL PROTECTED] > Sent: Friday, May 23, 2008 2:06 AM > To: Ted Mittelstaedt > Cc: 'User Questions' > Subject: Re: RAID 0+1 > > > Hey, > > > don't use gmirror and atacontrol at the same time. Use one or > the other. > > I don't. As I said: > > > Then I would merge the second > > slice of all 4 drives into a 0+1 array (first gstriping and > > then gmirroring them). I somehow succeeded this, but I also > > get a WARNING when booting the system: > > So I am gstriping first and then gmirroring the stripes. > And I am getting this warning message: > > >> WARNING: Expected rawoffset 0, found 63 > > I don't know if I should worry about this or not. > You mentioned you used atacontrol to create the array, that's not part of gmirror. However in any case, you are not done. What you need to do now is write some data to the array, then unplug one of the sata connectors to the drive. The array will go into fault mode. Then you need to add the disk back in and see if the array will accept it, or if the array ends up scotching everything. A mirror is no good if it can't actually survive a fault. I used to do this when selling servers to HP Proliants. I'd have a customer with me and go to one of our production, running HP servers, eject a drive from the array, give it to the customer for inspection, then plug it back in. Other than the red light appearing on the drive for a few minutes, the rebuild operation was entirely in hardware, the server wouldn't even blink. If your array can't do that, your just basically technically masterbating with your system to feel good - it is in actuality a completely worthless mirror setup. Ted ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: RAID 0+1
Hey, don't use gmirror and atacontrol at the same time. Use one or the other. I don't. As I said: > Then I would merge the second > slice of all 4 drives into a 0+1 array (first gstriping and > then gmirroring them). I somehow succeeded this, but I also > get a WARNING when booting the system: So I am gstriping first and then gmirroring the stripes. And I am getting this warning message: WARNING: Expected rawoffset 0, found 63 I don't know if I should worry about this or not. Thanks, Nejc ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
RE: RAID 0+1
don't use gmirror and atacontrol at the same time. Use one or the other. Ted > -Original Message- > From: [EMAIL PROTECTED] > [mailto:[EMAIL PROTECTED] On Behalf Of > Nejc Škoberne > Sent: Thursday, May 22, 2008 1:51 AM > To: User Questions > Subject: RAID 0+1 > > Hello, > > I have FreeBSD 7.0 and 4 250GB SATA disks and I would like to > make one big 500GB > 0+1 RAID array. My hardware is HP ProLiant ML110G5. > > First I tried creating ATA RAID arrays with BIOS tools, but > FreeBSD wouldn't recognize the arrays. > > Than I decided to create the RAID-0 arrays with atacontrol. I > succeeded, ending up with ar0 and ar1 each 500GB big. My plan > was to install the FreeBSD as usual on ar0 and then add both > into a gmirror array. However, I BootMgr seem not to be able > to boot from RAID-0 array created with atacontrol? Is this > correct? I guess this is because BootMgr knows nothing about > striping and cannot read the kernel. > > Finally I tried to create two slices on each drive. The plan > was to join the first slice of all 4 drives into a gmirror > array and mount it as / partition (I guess BootMgr would boot > normally from such an array). Then I would merge the second > slice of all 4 drives into a 0+1 array (first gstriping and > then gmirroring them). I somehow succeeded this, but I also > get a WARNING when booting the system: > > ad0: 238475MB at ata0-master SATA150 > ad1: 238475MB at ata0-slave SATA150 > ad2: 238475MB at ata1-master SATA150 > GEOM_STRIPE: Device gs0 created (id=2160028923). > GEOM_STRIPE: Disk ad0s2 attached to gs0. > ad3: 238475MB at ata1-slave SATA150 > GEOM_STRIPE: Disk ad1s2 attached to gs0. > GEOM_STRIPE: Device gs0 activated. > GEOM_STRIPE: Device gs1 created (id=3269017453). > GEOM_STRIPE: Disk ad2s2 attached to gs1. > GEOM_MIRROR: Device mirror/gm0 launched (4/4). > GEOM_STRIPE: Disk ad3s2 attached to gs1. > GEOM_STRIPE: Device gs1 activated. > WARNING: Expected rawoffset 0, found 63 > GEOM_MIRROR: Device mirror/gm1 launched (1/2). > GEOM_MIRROR: Device gm1: rebuilding provider stripe/gs1. > > Is this critical? It looks like the system is fine, though. > > Is there any "more proper" way to build such a system with RAID 0+1? > > Thanks, > Nejc > > ___ > freebsd-questions@freebsd.org mailing list > http://lists.freebsd.org/mailman/listinfo/freebsd-questions > To unsubscribe, send any mail to > "[EMAIL PROTECTED]" > ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
RAID 0+1
Hello, I have FreeBSD 7.0 and 4 250GB SATA disks and I would like to make one big 500GB 0+1 RAID array. My hardware is HP ProLiant ML110G5. First I tried creating ATA RAID arrays with BIOS tools, but FreeBSD wouldn't recognize the arrays. Than I decided to create the RAID-0 arrays with atacontrol. I succeeded, ending up with ar0 and ar1 each 500GB big. My plan was to install the FreeBSD as usual on ar0 and then add both into a gmirror array. However, I BootMgr seem not to be able to boot from RAID-0 array created with atacontrol? Is this correct? I guess this is because BootMgr knows nothing about striping and cannot read the kernel. Finally I tried to create two slices on each drive. The plan was to join the first slice of all 4 drives into a gmirror array and mount it as / partition (I guess BootMgr would boot normally from such an array). Then I would merge the second slice of all 4 drives into a 0+1 array (first gstriping and then gmirroring them). I somehow succeeded this, but I also get a WARNING when booting the system: ad0: 238475MB at ata0-master SATA150 ad1: 238475MB at ata0-slave SATA150 ad2: 238475MB at ata1-master SATA150 GEOM_STRIPE: Device gs0 created (id=2160028923). GEOM_STRIPE: Disk ad0s2 attached to gs0. ad3: 238475MB at ata1-slave SATA150 GEOM_STRIPE: Disk ad1s2 attached to gs0. GEOM_STRIPE: Device gs0 activated. GEOM_STRIPE: Device gs1 created (id=3269017453). GEOM_STRIPE: Disk ad2s2 attached to gs1. GEOM_MIRROR: Device mirror/gm0 launched (4/4). GEOM_STRIPE: Disk ad3s2 attached to gs1. GEOM_STRIPE: Device gs1 activated. WARNING: Expected rawoffset 0, found 63 GEOM_MIRROR: Device mirror/gm1 launched (1/2). GEOM_MIRROR: Device gm1: rebuilding provider stripe/gs1. Is this critical? It looks like the system is fine, though. Is there any "more proper" way to build such a system with RAID 0+1? Thanks, Nejc ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: Hyperthreading and schedulers [was Re: looking for a good sata or scsi raid 0/1 controller]
On Mon, 25 Apr 2005 17:54:20 -0700, in sentex.lists.freebsd.questions you wrote: >At 05:33 PM 4/25/2005, Mike Tancsa wrote: >>I would start by turning off Hyperthreading as >>that only really works if you are using the ULE scheduler and thats >>broken under RELENG_5. > >I haven't noticed this mentioned before. Is using an SMP kernel on a >system with a single P4 and SCHED_4BSD a waste of time? Is it >detrimental? I've never noticed any problems myself, but I never went >looking for any either. > See http://www.onlamp.com/pub/a/bsd/2005/01/20/smpng.html and http://lists.freebsd.org/pipermail/freebsd-stable/2003-December/005360.html to start. Also, the ULE scheduler is still broken ---Mike >-Glenn > >___ >freebsd-questions@freebsd.org mailing list >http://lists.freebsd.org/mailman/listinfo/freebsd-questions >To unsubscribe, send any mail to "[EMAIL PROTECTED]" Mike Tancsa, Sentex communications http://www.sentex.net Providing Internet Access since 1994 [EMAIL PROTECTED], (http://www.tancsa.com) ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Hyperthreading and schedulers [was Re: looking for a good sata or scsi raid 0/1 controller]
At 05:33 PM 4/25/2005, Mike Tancsa wrote: I would start by turning off Hyperthreading as that only really works if you are using the ULE scheduler and thats broken under RELENG_5. I haven't noticed this mentioned before. Is using an SMP kernel on a system with a single P4 and SCHED_4BSD a waste of time? Is it detrimental? I've never noticed any problems myself, but I never went looking for any either. -Glenn ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: looking for a good sata or scsi raid 0/1 controller
On Mon, 25 Apr 2005 12:22:15 +0200, in sentex.lists.freebsd.hardware you wrote: >Hello, > >I'm looking for a good raid controller, with good driver support, >something I could really rely on ! > > >I have some problems with my 3ware 8006-2 on freebsd 5.3 branch. >Actually the box is a p4-3000 hyperthreaded, the 3ware card is running >raid1 with 2 maxtor 80go drives. Acually the traffic on the card is less >than 2 MB/s. > >When the number of transactions/s a value around 70, the controller >seems to "lag", the average time par transaction can be more than 10 >seconds ! The box also reboot at random intervals giving no messages in >the logs. I have a couple of 3ware cards on RELENG_5 and they dont show such behaviour under load. I would start by turning off Hyperthreading as that only really works if you are using the ULE scheduler and thats broken under RELENG_5. Try rebuilding a uniprocessor kernel and see if the problem is still there. Also, enable crash dumps on the box to see if it's a bug or not. Rarely have I seen random reboots with no crash dump be a software issue. See http://www.freebsd.org/doc/en_US.ISO8859-1/books/developers-handbook/kerneldebug.html for info on how to enable crash dumps. I would also update to 5.4 as there are a number of bug fixes you might benefit from depending on how old your build is. ---Mike Mike Tancsa, Sentex communications http://www.sentex.net Providing Internet Access since 1994 [EMAIL PROTECTED], (http://www.tancsa.com) ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Recreating a software raid 0+1
Hello All, I am unsure where to post this. I have a Maxtor Nas 4000 w/4 drives. It recently took a dive for the worse, maxtor gives no support, except to try to mount it. I tried and was unsuccessful, i consitantly get super block erros. I was able to get this from Maxtor: 1. os = freebsd 2.05.3068 2. IDE software raid (OS) 0+1 I am unable to mount the drives for partitions, except where the OS lays. The rest are superblock errors. I don't know if I need (and how) to recreate the software raid on a new install of freebsd5. Please help! - Do you Yahoo!? Yahoo! SiteBuilder - Free, easy-to-use web site design software ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Recreating a software raid 0+1
Hello All, I am unsure where to post this. I have a Maxtor Nas 4000 w/4 drives. It recently took a dive for the worse, maxtor gives no support, except to try to mount it. I tried and was unsuccessful, i consitantly get super block erros. I was able to get this from Maxtor: 1. os = freebsd 2.05.3068 2. IDE software raid (OS) 0+1 I am unable to mount the drives for partitions, except where the OS lays. The rest are superblock errors. I don't know if I need (and how) to recreate the software raid on a new install of freebsd5. Please help! Neil ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Recreating a software raid 0+1
Hello All, I am unsure where to post this. I have a Maxtor Nas 4000 w/4 drives. It recently took a dive for the worse, Maxtor gives no support, except to try to mount it. I tried and was unsuccessful; I consistently get super block errors. I was able to get this from Maxtor: 1. os = FreeBSD 2.05.3068 2. IDE software raid (OS) 0+1 I am unable to mount the drives for partitions, except where the OS lays. The rest are superblock errors. I do not know if I need (and how) to recreate the software raid on a new install of freebsd5. Please help! Neil - Do you Yahoo!? Yahoo! SiteBuilder - Free, easy-to-use web site design software ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"