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
|
|||
|
|||
Windows XP Driver for Compaq 64 Bit Fibre Channel Adapter
I'm sorry for taking so long to reply. I am afraid I made rather merry
this New Years. I had an identical 1600 sitting on the bench beside the first one (that I mentioned in the prvious post), and I switched a drive from the first system to the scond system (which had an identical RAID arrangement) and the new system assimilated to the established OS installed on the disk that I inserted. I think this corresponds to the experiment that Will suggested. The results: Compaq Hardware RAID works exactly as any of us could possibly hope. The new (completely seperate and unconnected) system BECAME the old system after about two and a half minutes of disk activity. If Will is suggesting a strange redundancy between uncompaible systems, then he may have a point... but as far as Compaq Hardware Raid is concerned, everything WORKS as it has been advertised to work...at least according to my bench tests. |
#22
|
|||
|
|||
Windows XP Driver for Compaq 64 Bit Fibre Channel Adapter
What I have done probably six times in a production environment to recover
in the software RAID scenario (using RAID 1) is the procedure I already spelled out below. I remove the dead drive, swap in the mirror as its replacement, and reboot and then rebuild a new mirror. -- Will "NuTCrAcKeR" wrote in message ... what i mean is this: with software raid, you basically have 1 boot drive. The big hint I am going to give you here is this: ARC Path. If the "boot" drive in a software mirror goes bad ... what are you going to do to boot your system using the other disk (the partner in the mirror)? The following constraints apply: no re-install media only 1 computer at hand (the one with the failed disk) - LC "Will" wrote in message ... I guess I don't understand your question. What do you mean by *primary* here? When I configure a boot volume as RAID 1 using Windows software RAID, either of those two volumes is bootable. When I lose the primary in SCSI ID=0, I simply place the drive that was in SCSI ID=1 into the ID=0 slot and reboot the system. Windows announces that it sees new devices (actually it just sees movement in position of a device), and it asks to reboot. After rebooting, the mirrored drive is now your boot device. To fully recover, you get into Device Manager in Windows, and remove the mirror on the now-missing defective drive. You then insert a new device, rescan to acquire it, mark it as dynamic, and then mirror to it. I've done the above sequence many, many times, and it is infinitely easier than any hardware RAID I have used (and allows for more powerful modes of recovery because you can work on a device from a different computer before reinserting it to reboot from it). -- Will |
#23
|
|||
|
|||
Windows XP Driver for Compaq 64 Bit Fibre Channel Adapter
No, that is cheating. You can't guarantee that every server in your
environment will be identical, and that it will have identical RAID hardware. The beauty of the software RAID approach is that you can take the device to any computer with any hardware and do whatever you need to do. We have Dell and Compaq 1U, 2U, and 4U intermixed in our environment, not to mention workstations from both vendors, and I like being able to take any drive from any of those systems and inspect them across the whole environment. -- Will "Jeffrey Alsip" wrote in message oups.com... I'm sorry for taking so long to reply. I am afraid I made rather merry this New Years. I had an identical 1600 sitting on the bench beside the first one (that I mentioned in the prvious post), and I switched a drive from the first system to the scond system (which had an identical RAID arrangement) and the new system assimilated to the established OS installed on the disk that I inserted. I think this corresponds to the experiment that Will suggested. The results: Compaq Hardware RAID works exactly as any of us could possibly hope. The new (completely seperate and unconnected) system BECAME the old system after about two and a half minutes of disk activity. If Will is suggesting a strange redundancy between uncompaible systems, then he may have a point... but as far as Compaq Hardware Raid is concerned, everything WORKS as it has been advertised to work...at least according to my bench tests. |
#24
|
|||
|
|||
Windows XP Driver for Compaq 64 Bit Fibre Channel Adapter
I'm sorry, but you are not defining premises and I don't think you can reach
any valid conclusions without doing that first. First, I don't think you understand the problem I am trying to explain. Let's say you install software (e.g., a Windows Update) that prevents that computer from booting. It does happen on rare occasion, and when it does the beauty of the software RAID mirror is that you can remove that boot device and insert it to *any* hardware. It's very fast at that point to run a software restore. It's a critical requirement for us that I not have to duplicate exactly the hardware RAID configuration on some new machine just so I can inspect the drive that was in the old machine. Once the drive is mounted on another computer, you can restore from backups, or do other steps to manually recover the system from backups you may already have on disk. Hardware RAID offers no protection at all from this kind of software-related configuration failure. It does however make it very hard to recover unles you have some kind of Disaster Recovery configuration disk at hand. You can do a parallel OS install of course, but all of that takes time to do, time to configure, and time to maintain. Been there done that and it works but just costs too much time. Second, your trying to make a point based on my saying I do something "frequently" is just taking a cheap shot. If I maintain 70 computers and 2% of them have a software-related issue that prevents them from booting, that means once or twice each year I'll need to recover those systems. That's frequent but inconsequential as a percentage of the total time taken to administer the systems. In reality, over seven years, I have had to recover maybe 12 machines whose software configurations had problems, and I've certainly had plenty of those be hardware RAID. Hardware RAID has always always made my life more difficult, not easier. Where hardware RAID is an essential element is for large RAID 5 configurations. But I never use it for the boot device because it just doesn't assist with recovery once the OS becomes corrupted and you cannot boot. If I were running a mission critical database for eBay, or even something with a modest cost in the many thousands of dollars per hour for downtime, then probably I would use hardware RAID for the boot volume, but I would also have a backup server ready to take over from the primary, and I would have plenty of change control in place so that if a software install brings down a system the replacement is there in the ready to take over. All of those steps are time consuming, and the applications we maintain might cost about $1000/hour for downtime. Since in every case to date we have been able to recover a software mirrored boot device within 10 to 20 minutes, the reward to risk ratio has been reasonable. In every case where I have had a hardware RAID mirror fail the boot device and not had an identical - unused - server with identical RAID hardware available, the recovery has taken more than three hours. I'm not trying to be difficult about this. I used to believe that hardware RAID was the only way to go as well. But after seven years of working with them I'm simply interested in minimizing the organization's costs. They have their place for data and application volumes, but for us the boot device isn't the right place for hardware RAID, in a small company. -- Will "NuTCrAcKeR" wrote in message ... software raid often necessitates that, so dont use it. It is the way of the amateur. I have never had raid volume corruption on a RAID1 volume (harware raid) and I have used the following controllers: Mylex AMI Dell PERC HP NetRAID IBM ServeRAID Adaptec and Compaq Smart Array (as far back as the IDA and IDA II models). I have to site your approach to raid as being serious flawed. Especially if you find yourself "frequently" haveing to swap drives between systems to get data from them. I dont fully understand what you are doing, but it just doesnt sound right. "Will" wrote in message ... The problem I have with hardware RAID is that quite often to recover a system I need to remove the drive, mount it on a different computer, and then work with its file system. Hardware RAID makes that next to impossible to do easily. -- Will "Nut Cracker" wrote in message ... interesting ... i would still prefer a hardware based solution. Just out of curiosity, if your primary drive were to fail under the mirror, do you know what you would need to do to boot the system again ? - LC "Will" wrote in message ... I found the article that talks about how to make RAID-5 on Windows XP work, and I assume once you did that tweak you would be able to do mirroring as well. http://www.tomshardware.com/2004/11/...pen/index.html -- Will "NuTCrAcKeR" wrote in message ... Yes, I know. I posted that to illustrate that you cant do it on XP. The workstation products used to be able to do simple stripes and mirrors. im rather suprised they dropped the support for it. In the meanwhile, use a hardware based RAID controller. Hardware is always better than software, in my opinion. I wipped out my googfing and went to work. I didnt find anything, and i doubt it would be a simple as a registry tweak. - LC |
#25
|
|||
|
|||
Windows XP Driver for Compaq 64 Bit Fibre Channel Adapter
anyone that thinks they can take a compaq raid volume and move it to an IBM
or Dell array controller (or any other interoperable scenario) and have it work should have thier head checked. I really hope that isnt what I think I am hearing the OP complain about. - LC "Jeffrey Alsip" wrote in message oups.com... I'm sorry for taking so long to reply. I am afraid I made rather merry this New Years. I had an identical 1600 sitting on the bench beside the first one (that I mentioned in the prvious post), and I switched a drive from the first system to the scond system (which had an identical RAID arrangement) and the new system assimilated to the established OS installed on the disk that I inserted. I think this corresponds to the experiment that Will suggested. The results: Compaq Hardware RAID works exactly as any of us could possibly hope. The new (completely seperate and unconnected) system BECAME the old system after about two and a half minutes of disk activity. If Will is suggesting a strange redundancy between uncompaible systems, then he may have a point... but as far as Compaq Hardware Raid is concerned, everything WORKS as it has been advertised to work...at least according to my bench tests. |
#26
|
|||
|
|||
Windows XP Driver for Compaq 64 Bit Fibre Channel Adapter
you still havent answered the question, which is a very basic one at that.
with only 1 system, if you lose your boot drive in a software mirror set, what do you need to do to get the system to boot from the other drive? you reputation as a contributor here is on the line, Will. Just answer the question. - LC "Will" wrote in message ... No, that is cheating. You can't guarantee that every server in your environment will be identical, and that it will have identical RAID hardware. The beauty of the software RAID approach is that you can take the device to any computer with any hardware and do whatever you need to do. We have Dell and Compaq 1U, 2U, and 4U intermixed in our environment, not to mention workstations from both vendors, and I like being able to take any drive from any of those systems and inspect them across the whole environment. -- Will "Jeffrey Alsip" wrote in message oups.com... I'm sorry for taking so long to reply. I am afraid I made rather merry this New Years. I had an identical 1600 sitting on the bench beside the first one (that I mentioned in the prvious post), and I switched a drive from the first system to the scond system (which had an identical RAID arrangement) and the new system assimilated to the established OS installed on the disk that I inserted. I think this corresponds to the experiment that Will suggested. The results: Compaq Hardware RAID works exactly as any of us could possibly hope. The new (completely seperate and unconnected) system BECAME the old system after about two and a half minutes of disk activity. If Will is suggesting a strange redundancy between uncompaible systems, then he may have a point... but as far as Compaq Hardware Raid is concerned, everything WORKS as it has been advertised to work...at least according to my bench tests. |
#27
|
|||
|
|||
Windows XP Driver for Compaq 64 Bit Fibre Channel Adapter
You keep making my points for me and then trying to make them arguments
against me. I don't think you get it at all. I don't expect drives from a Compaq hardware RAID to work in IBM or Dell hardware RAID. That's one reason why we prefer software mirrors for boot volumes. Because we the software mirror we can move the drive across different vendor's hardware without any issue, and work on the drive to make it bootable again. -- Will "NuTCrAcKeR" wrote in message ... anyone that thinks they can take a compaq raid volume and move it to an IBM or Dell array controller (or any other interoperable scenario) and have it work should have thier head checked. I really hope that isnt what I think I am hearing the OP complain about. - LC "Jeffrey Alsip" wrote in message oups.com... I'm sorry for taking so long to reply. I am afraid I made rather merry this New Years. I had an identical 1600 sitting on the bench beside the first one (that I mentioned in the prvious post), and I switched a drive from the first system to the scond system (which had an identical RAID arrangement) and the new system assimilated to the established OS installed on the disk that I inserted. I think this corresponds to the experiment that Will suggested. The results: Compaq Hardware RAID works exactly as any of us could possibly hope. The new (completely seperate and unconnected) system BECAME the old system after about two and a half minutes of disk activity. If Will is suggesting a strange redundancy between uncompaible systems, then he may have a point... but as far as Compaq Hardware Raid is concerned, everything WORKS as it has been advertised to work...at least according to my bench tests. |
#28
|
|||
|
|||
Windows XP Driver for Compaq 64 Bit Fibre Channel Adapter
I think I have answered this question now three times. I don't understand
why it is you don't see the response as a response. If I loose the boot device in a Windows 200x software mirror, I perform steps similar to the following: 1) I remove the dead drive. 2) I take the mirrored drive from the SCSI ID=1 position and move it to the SCSI ID=0 position. 3) I boot the system. The system will claim to have found new devices on login, and I reboot again. From this point on everything else is cleanup work. 4) I start Disk Manager and I remove the mirror from the missing drive. 5) I remove the missing drive from the configuration entirely. 6) I insert a new drive and make it Dynamic. 7) I mirror to it. -- Will "NuTCrAcKeR" wrote in message t... you still havent answered the question, which is a very basic one at that. with only 1 system, if you lose your boot drive in a software mirror set, what do you need to do to get the system to boot from the other drive? you reputation as a contributor here is on the line, Will. Just answer the question. - LC |
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
nVidia GeForce4 MX 4000 - Jerky Video | John in VA | Nvidia Videocards | 12 | June 6th 05 12:00 AM |
Game freezes system - possible ATI issue? | Blaedmon | Ati Videocards | 1 | February 12th 05 06:20 PM |
IDE Channel driver for 8KNXP | Frank Buff | Gigabyte Motherboards | 1 | December 29th 04 11:16 AM |
Radeon 7500 Saphire Windows ME Problem | Pamela and Howard Signa | Gateway Computers | 5 | February 17th 04 11:07 PM |
9700 Pro Crash during benchmark with 3dmark03 | Sean | Ati Videocards | 5 | December 17th 03 04:47 PM |