Monitoring disk failures

[gwlogue]

Are there any tools or techniques that can monitor disks for problems before they fail?

I run a small business on my system just like several other trades guys I know in my town. I have seen two of them go through quite a disruptive process trying to get their systems back after a disk failure. They had adequate backups in place so they didn't loose any data but they were out of business for a day or two while they had the drives replaced and their systems rebuilt. One guy is having problems now that I suspect are a failing disk and he's too darn busy with his regular job, (an electrician), to take the time to have it looked at.

I'm looking for some way to be proactive. The ideal would be to have some sort of tool that would raise flags before things got critical so I could aquire another drive and migrate my data in a controlled manner before the drive fails. I could continue to use the wonky drive as a slave for low impact data until it decided to give up the ghost.

If the technology is not at that stage yet then I will be investing some time in documenting and testing a recovery plan to minimize the outage when it comes my turn.

Thanks in advance for your opinions.

[Midknyte]

http://grc.com/spinrite.htm

you also need to check the event viewer logs. Win2k/2k3 will give messages of possible failure.

why don't you use RAID1 or RAID5? that pretty standard practice for servers nowadays.

[Midknyte]

speed fan has some SMART detection:
http://www.almico.com/speedfan.php

[gwlogue]

The keyword here is SMALL business. We are all 1-man operations running our business's on standard wintel desktop systems.

If servers and RAID are the answer maybe we should consider banding together and finding someone who could provide centralized server support for guys like us?

Can that sort of thing be done remotely? We all operate out of our basement offices at home.

Thanks

[Midknyte]

RAID1 is pretty common on motherboards now. check your specs. That would be a minimum for your server or main workstation. This won't protect you from viruses or worms, but it would save you if one drive suddenly died.

When you said small business, you didn't say one man. That's more like home business to me.

I wouldn't outsource unless you really need to. You never know who's going to be looking at your data if you enable remote access.

[davewuk]

I second Speedfan as its monitoring utility allows you to visualise overall health as a percentage and technically lists the features of the drive with corresponding manufacturer limits against what is recorded by S.M.A.R.T. this will flag a failing drive. However, with S.M.A.R.T. enabled in the BIOS, at post the same failing will be shown here with an advisory to backup.

Midknyte's suggestion for RAID is a good idea and as stated does not have to be expensive. If your mobo is not equipped, then add on cards are obtainable for very little outlay, drop in another drive and mirror the system. Having said that, there is always the slightest possibilty of both drives going down so buy reliable brands and not at the same time as the main drive.

[Sterling_Aug]

Maybe the best plan of attack would be for you all to band together and have one person provide the backup solutions.

While I was in school 4 years ago (for job retraining) one of the teachers was involved in remote backup for several local businesses. He provided all of the hardware, time, support, software, etc to remotely backup the systems. He would then have the backups stored in a bank vault for safe keeping in case any one system failed.

Depending on where you are located, I have most of the hardware and software here at home already. It would only require a small investment in new/bigger SCSI drives for my RAID5 server and I could send you the free remote control software that would allow the secure file transfers over the internet.

Let me know if you and your other businessmen are interested and we can discuss the goals, define the costs, and finalie the terms.

[davewuk]

Though the remote backup scenario is a sound idea, this still fails to address the OP question on tools for determining if a drive is going south. He has stated that the other traders all had backups, it was the downtime to getting their systems up again which proved frustrating. Probably what is needed here is the solution you propose and a utility similar to Intel's Landesk software which given the right environment and setup allows admin monitoring of remote systems, whether drive status is one of the features I've no idea!

IMO the mirrored RAID is the way to go, at least the main drive is duplicated which with a simple change of jumper will get a rig up in minutes.

[Sterling_Aug]

The problem is more like this: The time it takes to keep the systems running and the time it tskes to recover from a crash. The businessmen all have jobs to do, not take time to redo the computer systems.

You could also "hire" a local computer tech to roam around between sites to monitor the PCs.