Technology Tidbits

This page is designed primarily to be Google fodder - occasionally in my travails with computers I come across the odd quirky problem that isn't readily soluable by a straight search engine query. In an attempt to save others the same search it is my intention to document some of these now solved problems here so that others might find the answer more readily. As such this is going to be somewhat dry reading and without further ado...

IDE hard-drive failing on resume from S3 (Suspend to RAM) in Windows XP with a JMicron JMB36X Controller

This was a somewhat scary one simply due to the way it presented. My Windows XP box was upgraded from a 160 Gb IDE primary OS drive to a 320 Gb SATA2 drive. Due to the general phasing out of IDE drives at the time the motherboard I bought was designed with SATA in mind foremost and IDE was an afterthought and only available via the JMicron RAID controller. While I wasn't entirely happy with that solution of running two drives in JBOD mode through a RAID controller it seemed to work pretty well. Till I ran out of drive space on the C: drive.

Okay so I upgraded the drive. Naturally the OS drive being the primary and first drive put in I made it the Master drive. Cloning it to the new SATA drive after having booted Windows XP with the SATA drive in (to force it to load the drivers for SATA devices) I simply removed the older IDE Master drive and began using the new SATA drive as normal. This presents as being entirely fine - the machine boots, runs and both drives were entirely happy. It wasn't till the next day that the oddness began.

I awoke to the machine being active, when it should have been sleeping, with the hard drive activity light jammed solidly on. Even more oddly Windows was hitching becoming inactive for long periods then returning for a brief period which allowed me to coax a graceful power down. Reseting the machine and all returned to normal. Having work to go to I didn't think too much more till returning home I find the machine once again on with a jammed drive activity light and hitching. This got me to check the System Event log to find the follwoing two error messages :-

Event ID 9 JRAID The device, \Device\Scsi\JRAID1, did not respond within the timeout period.
Event ID 51 Disk An error was detected on device \Device\Harddisk1\D during a paging operation.

This are somewhat scarey error messages to see. Generally these are the heralds of a dying hard drive and NT alerting you to this. However that this had started exactly the day after I pulled an IDE drive out I figured it was far more likely that my messing around had upset something. The other big pointer is the fact that the second drive is entirely used for data storage and has no pagefile on it or any applications run from it. Demand paging simply shouldn't be an issue.

To cut a long story short chatting with a friend, thanks Ewen, he pointed out that IDE technically doesn't like not having a Master device on the bus. Although most modern IDE controllers are not at all fussed with this is seems that the JMicron controller has managed to replicate this much older, more 1990 style, behavior. One link inserted later to force the Slave drive into a Master and the machine became nice and stable again waking from S3 without issue.

As a further side note to this, if you can't force the drive to become a Master but still need some form of sleep mode then consider Hibernate rather than Suspend to RAM. Hibernate allowed the machine to wake and sleep without issue, presumably allowing the BIOS a brief chance to swift kick the hardware into shape which the more Windows only S3 doesn't permit.

Philip R. Banks
Send Email

Return to the Archives
Return to Fortress Entrance