There are several places where the transition from one RAID state to another results in Capacity changes, or partition table changes. Each change in state would walk the device nodes, and quite frankly, at a rapid pace. There are several steps when creating an array, each step requires the updated information to be part of the system. These changes are initiated asynchronously by the Firmware (only access to kernel space), and not the Management Software (user space). The events that trigger these changes are reported as AIF (Adapter Initiated FIB) and trigger the rescan and target determination via APIs in the kernel. In one of the more critical states of array creation the partition table of the original targets can show on the array, then be cleared as the build starts. Hotspares have a similar action where the legacy partition table information is erased, and this needs to be reflected on the system, without a node walk (the alternative to re-reading the partition tables, WWN and Capacity information is a node walk). It is obvious the effect thin-provisioning would have on node-walking, snapshot is less obvious. The Firmware actually creates a new R/O device, but the original R/W device is replaced internally with the R/O device with a Journal to track and manage the R/W nature. There is a 'glitch' where the device changes state three times and one of those states resulted in a momentary increase in Capacity. They were not going to fix this, but no bother, the fact that the device is listed as Removable hid this problem from Linux. Adaptec RAID controllers are not the only ones with this problem. I am sure, given time, I can provide some additional cases. Sincerely -- Mark Salyzyn -----Original Message----- From: Ankit Jain [mailto:jankit@xxxxxxx] Sent: Friday, May 20, 2011 9:00 AM To: Mark Salyzyn Cc: Christoph Hellwig; linux-scsi@xxxxxxxxxxxxxxx; aacraid@xxxxxxxxxxx; Hannes Reinecke Subject: Re: [PATCH] aacraid: Add module param to export disks as 'fixed' instead of 'removable' On 05/19/2011 01:51 AM, Mark Salyzyn wrote: >> On Wed, May 18, 2011 at 01:05:29AM +0530, Ankit Jain wrote: >>> aacraid: Add 'aac_export_fixed' module param to export disks as >>> 'fixed' instead of 'removable'. >>> Adaptec controllers set the removable flag on devices to >>> allow for reconfiguration. The fear is that _not_ setting this >>> flag would inhibit the OS from detecting any device reconfiguration. >> The flag should not be set at all and just be removed. > > This will break the RAID Management Software, and various Firmware > features (snapshot, thin-provisioning etc) I'm trying to get a better understanding. How do the RAID management software and firmware features depend on this? -- Ankit Jain SUSE Labs ______________________________________________________________________ This email may contain privileged or confidential information, which should only be used for the purpose for which it was sent by Xyratex. No further rights or licenses are granted to use such information. If you are not the intended recipient of this message, please notify the sender by return and delete it. You may not use, copy, disclose or rely on the information contained in it. Internet email is susceptible to data corruption, interception and unauthorised amendment for which Xyratex does not accept liability. While we have taken reasonable precautions to ensure that this email is free of viruses, Xyratex does not accept liability for the presence of any computer viruses in this email, nor for any losses caused as a result of viruses. Xyratex Technology Limited (03134912), Registered in England & Wales, Registered Office, Langstone Road, Havant, Hampshire, PO9 1SA. The Xyratex group of companies also includes, Xyratex Ltd, registered in Bermuda, Xyratex International Inc, registered in California, Xyratex (Malaysia) Sdn Bhd registered in Malaysia, Xyratex Technology (Wuxi) Co Ltd registered in The People's Republic of China and Xyratex Japan Limited registered in Japan. ______________________________________________________________________ -- To unsubscribe from this list: send the line "unsubscribe linux-scsi" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
