Using file system information in raid data reconstruction and migration

Apparatus including

• means for identifying a first group of data blocks, said first group including a block capable of maintaining redundant information associated with said data blocks;
• means for determining if there is a failed data block in said first group;
• means for reconstructing data from said failed data block in response to a result of said determining operation;
• means for sending said data to a data block associated with a second group, said second group having at least one non-failed data block;
• means for maintaining information about a set of failed data blocks on a first disk, said information including a stripemap with entries indicating which groups of data blocks have been migrated to one or more second disks;
• and
• means for logically setting said failed data block to a selected value, whereby later calculation of said redundant information for said first group can proceed without reconstructing said data.

Memory or mass storage including instructions, said instructions being capable of being performed by one or more computing devices, said instructions directing said one or more computing devices to

• identify a first group of data blocks, said first group including a block capable of maintaining redundant information associated with said data blocks;
• determine if there is a failed data block in said first group;
• in response to a result of said instructions to determine, reconstruct data from said failed data block and send said data to a data block associated with a second group, said second group having at least one non-failed data block;
• maintain information about a set of failed data blocks on a first disk, said information including a stripemap with entries indicating which groups of data blocks have been migrated to one or more second disks;
• and
• logically set said failed data block to a selected value, whereby later calculation of said redundant information for said first group can proceed without reconstructing said data.

A method including

• selecting a set of data blocks to send from a source device to a target device;
• copying said data blocks to said target device;
• copying parity associated with said data blocks from said source device to said target device;
• setting any unallocated blocks at said target device to a selected value, whereby said parity for said data blocks is accurate on said target device;
• whereby a minor relationship between said source device and said target device is established.

Memory or mass storage including instructions, said instructions being capable of being performed by one or more computing devices, said instructions directing said one or more computing devices to

• select a set of data blocks to send from a source device to a target device;
• copy said data blocks to said target device;
• copy parity associated with said data blocks from said source device to said target device;
• set any unallocated blocks at said target device to a selected value, whereby said parity for said data blocks is accurate on said target device;
• whereby a mirror relationship between said source device and said target device is established.

Apparatus including

• means for selecting a set of data blocks to send from a source device to a target device;
• means for copying said data blocks to said target device;
• means for copying parity associated with said data blocks from said source device to said target device;
• means for setting any unallocated blocks at said target device to a selected value, whereby said parity for said data blocks is accurate on said target device;
• whereby a minor relationship between said source device and said target device is established.

An apparatus coupled to a RAID group storing datablocks, the apparatus comprising:

• a data storage device and a processor capable of accessing the data storage device;
• wherein the processor maintains on the data storage device information regarding the datablocks, the information including
• a zeromap table, including a set of entries describing which datablocks have been set to zero;
• a blockmap table, including a set of entries describing which datablocks are in use by allocated file system data;
• and
• a stripemap table, including a set of stripemap entries that correspond to one or more RAID stripes in a failed disk in a RAID group that have been migrated to a nonfailing disk in said RAID group.

An apparatus coupled to a RAID group storing datablocks, the apparatus comprising:

• a data storage device and a processor capable of accessing the data storage device;
• wherein the processor maintains on the data storage device information regarding the datablocks, the information including
• a zeromap table, including a set of entries describing which datablocks have been set to zero;
• a blockmap table, including a set of entries describing which datablocks are in use by allocated file system data;
• a stripemap table, including a set of stripemap entries that correspond to one or more RAID stripes in a failed disk in a RAID group that have been migrated to a nonfailing disk in said RAID group;
• and
• said set of entries in said stripemap table is updated after reconstructing said set of entries in said blockmap table.