If this is your first visit, be sure to check out the FAQ by clicking the link above. You may have to register before you can post: click the register link above to proceed. To start viewing messages, select the forum that you want to visit from the selection below. |
|
|
Thread Tools | Display Modes |
#21
|
|||
|
|||
In article ,
Malcolm Weir wrote: The company that I'm using to post this tried a few. They sucked. And Apple has no idea how to support serious operations. Their response seems to be "reformat/reinit and try again". An obviously fabricated tale. |
#22
|
|||
|
|||
In article ,
"Fatboy40" no one here, sorry wrote: I've never heard of the Apple Xserve RAID but I'm glad I now know about it, for the price it seems damn good. It has most of the redundancy features I'm after however it doesn't look like the controller of the Xserve has a backup You can two controllers, but I don't think they talk to each other, so there is no failover if one controller just quits. |
#23
|
|||
|
|||
In article ,
"Fatboy40" no one here, sorry wrote: I've got to ask, why exactly did they suck, was the hardware prone to fail Apple's consumer level hardware is regarded as the best in the industry by Consumer Reports' readers. I don't know there are long term reviews of the enterprise level hardware. or was Apple support just **** poor ?. If they're a pile of crap then surely Apple must know this, there must be some people out there who are happy with it ?. Apple's consumer level tech support is regarded as the best in the industry by Consumer Reports' readers. I don't know if their enterprise level support has been reviewed. Apple also has a public forum for the Xserve hardware and RAID at http://discussions.info.apple.com/[email protected] I imagine if there were unhappy customers, at least some would show up here to let Apple know it. |
#24
|
|||
|
|||
On Thu, 16 Dec 2004 21:24:25 -0000, "Fatboy40" no one here, sorry
wrote: On paper they seem like a reasonable bit of kit with dual everything (PSU's, controllers etc.). I've got to ask, why exactly did they suck, was the hardware prone to fail or was Apple support just **** poor ?. If they're a pile of crap then surely Apple must know this, there must be some people out there who are happy with it ?. RAID looks easy, but it isn't. The tricks lie in determining *how* you know you've got a failure, what you do when you detect a possible failure, what you do when you get a second fault, etc. One common mistake that novice vendors often make is to omit a mechanism to force a disk *back* into being valid after they've marked it bad. If you've got a bad (set of) sector(s) somewhere, a common error is to fault the entire array and make it inaccessible, despite the fact that the rest of the disk isn't bad, and the problem sectors lie in a chunk of disk space that you can happily ignore (e.g. it's static data that you can recreate). These things happen... and then aggravated by inept support who haven't quite grown out of the "reformat and start again" mindset... Malc. |
#25
|
|||
|
|||
On Fri, 17 Dec 2004 05:26:25 GMT, flux wrote:
In article , Malcolm Weir wrote: The company that I'm using to post this tried a few. They sucked. And Apple has no idea how to support serious operations. Their response seems to be "reformat/reinit and try again". An obviously fabricated tale. Flux is not only infantile, but stupid and a liar. Try this, you idiot: http://www.computerworld.co.nz/cw.ns...256E27000D5707 And then apologize for your lie. Malc. |
#26
|
|||
|
|||
On Fri, 17 Dec 2004 04:56:52 GMT, flux wrote:
In article , Malcolm Weir wrote: The silly little boy doesn't understand the difference between the Toshiba 1.8 inch MK2004GAL drives used in an iPod (3600rpm) and the drives used in high-end storage (15,000rpm). The difference is that Toshiba is like a Timex, takes a licking and keeps on ticking. Now trying throwing one of these unnamed "high-end storage" devices on the couch on a regular basis-sometimes while it's spinning--and seeing how long it lasts. Unnamed? Are you *really* this ignorant? Hitachi and Seagate are the common 15K drives. As to your "point", why would anyone vaguely competent throw enterprise storage on "the couch"? Where would a competent data center employee *find* a couch in a data center? And I'm mightily amused at the idea of using 20gb or 40gb 3600 rpm drives as data center storage! Still, that shows what sort of environment this flux kid has knowledge of! Malc. |
#27
|
|||
|
|||
Malcolm Weir writes:
On Fri, 17 Dec 2004 05:26:25 GMT, flux wrote: In article , An obviously fabricated tale. Flux is not only infantile, but stupid and a liar. Try this, you idiot: http://www.computerworld.co.nz/cw.ns...256E27000D5707 And then apologize for your lie. A relevant quote here is "Never argue with an idiot. They drag you down to their level and then beat you with experience." Ignore him and move on. *p |
#28
|
|||
|
|||
Faeandar,
For me it's all about the best components at the cheapest price and to be honest it really should be SCSI for the disks otherwise I know I'm asking for trouble. My no. 1 choice is the EMC CX300 but I think that's out of my league, my no. 2 now seems to be the HP MSA1000 (specifically the kit version they produce with a switch and two HBA's (which seem to be rebranded Qlogic)). I think I'll start a seperate thread about the HP hardware to see if anyone has any opinions on it (and use Google Groups for research), for around £10,000 I could setup a pretty good SAN using the MSA1000 and HP resellers/support companies are all over the place and at a later date add backup switches/controllers etc.. "Faeandar" wrote in message ... On Tue, 14 Dec 2004 09:39:24 -0000, "Fatboy40" no one here, sorry wrote: I think this is the perfect forum, and honestly it could use more traffic so post away.... I didn't catch how large you wanted to make this but unless you're really tied to EMC or HP there are alot of other array mfg's that make very good array's. And all that makes a SAN is the switch, so buy what you can get or what is supported. Brocade, McData, and Cisco fit just about everyone's support matrix. ~F |
#29
|
|||
|
|||
"flux" wrote in message
... In article , "Fatboy40" no one here, sorry wrote: Years ago, I would say there was a significant distinction between the "high-end" drives and the "low-end" But "low-end" drives are much better and are really no longer "low-end". For example, even the "low-end" drives used in consumer devices like Tivo and the iPod have decent throughput and reliability. Tivo is intended to operate 24/7 and iPods are MP3 players, which are prone to being dropped and tossed around. Anyway, here is an excellent article about a "beer-budget" SAN: http://www.nwc.com/showitem.jhtml?docid=1519ws1 I am not a storage guru, I run a world class hpc cluster and have had to get a grip on storage in our environment. Most storage vendors we have talked to call them "tier-2" storage and we have talked to a bunch. Our leveraged storage group calls them tier-2 storage (usually looking down their nose) and most of my peer in other companies call them the same thing. So I will most likely keep calling them that. While I think it is beginning to change, corporations-specially large ones have a tendency not to want to use ATA drives for their primary storage needs. Ask around and you will be told how these drives just do not measure up to fibre and SCSI drives. You can never get fired getting a Fibre Netapp or a Fibre SAN. Cluster guys are statistics whores. We analyze and test everything to death before it goes on-line-cause one you are on-line you can't stop production. Part of our acceptance procedure is to run anything we get through a beater job composed of a home grown disk tester that keeps statistics and an mpi linpack that kill memory. We always bust cache when testing performance. And we run it for at least 7 days-even under unbearable pressures to get it on-line. We have some 150TB going on 180TB - what we will retire. Unlike most environments we ride our storage very very hard-and of course we have the stats to prove it! We also do not do any backups-because with this amount of data it is actually easier to recreate it (and a whole lot cheaper) so not losing data is very important. Here's what I have learned about ATA drives do fail a little more often. But if we have our vendor burn it in, and we run through our process (cluster guys are also very anal about the process) we will fail the weak drives (and CPU and memory etc) before they go into production. Make sure you scrub the disks (3ware calls it verify) on a weekly basis. Going backwards, we have a CX500 test box w/ 18TB which we have beating like a red-headed step child with 2 Opterons. We have pushed some 330MB on on controller. This box helped get a good build for the head nodes and for the last 90 days have spent their entire lives in our office getting hammered. We lost one drive the first week. Going back over this year, we have 50TB of Netapp R200. We have had to replace 4 disks. 2 failed, 2 were going to fail. Over 18 months, 50TB of commodity ATA drives w/ 3ware cards purchased 30 and then 20 6 months later. These boxes were vendor supplied and maintained as NAS boxes. It used RH9 and XFS and we broke it. It took and 80 node job 3.5 days but it broke it 100% of the time. So we had to develop our own build and we really cleaned up the installation process. Our second 20 TB we had them burn in on their end for 72 hours and then we run our beater jobs. We lose one or two drives a month in this group (320 or so total) So what is our new beer-budget SAN? A CX700 with 30+ TB (usable) of ATA drives!!! Merry Xmas to ya'll. -- Wolf ---------------------------------------------------------------- Please post all responses to UseNet. All email cheerfully and automagically routed to Dave Null |
#30
|
|||
|
|||
Hitachi and Seagate are the common 15K drives.
Hitachi? Ex-IBM UltraStars? -- Maxim Shatskih, Windows DDK MVP StorageCraft Corporation http://www.storagecraft.com |
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
Abit Forums? Can't post.. | Raphael Bustin | Abit Motherboards | 1 | July 14th 04 03:46 PM |
Free Support forums various categories. | Askhow.net | General | 0 | January 12th 04 07:12 PM |
Gateway computer web-based forums? | Andy Roberts | Gateway Computers | 2 | December 5th 03 12:13 PM |
Forums for cheap PC components | harp | General | 2 | July 11th 03 10:08 PM |