The risk to using swap on raid is that resync and swap operations can not go on at the same time so that if the swap raid set has been damaged and is resyncing, swap can not be allowed. There are several workaround scripts that detect the resync and allow it to finish before proceeding. Would not a better (faster) solution be to detect the out of sync condition for a dedicated swap raid set, halt the resync and do a mkraid -f followed by a mkswap, swapon. This should be faster right?? There was some talk at one time of modifying the kernel raidstart code so that the order of resyncing could be controlled in some fashion when it was needed. This would seem to be a perfect application for this. Perhaps resyncing should be delayed or disallowed until started by something in the start up code, just to accomodate the ordering of resync in the event swap on raid is in use. Michael [EMAIL PROTECTED]
