Performing offline CDS (Cross-Platform Data Sharing) migration tasks on VxFS file systems using fscdsconv command fails with metadata overflow
Article ID: 100012465
Updated On:
Description
Error Message
UX:vxfs fscdsconv: INFO: V-3-26037: Would you like to proceed further with migration? (ynq) y
UX:vxfs fscdsconv: INFO: V-3-26041: The filesystem would need to be byteswapped for use on this target.
UX:vxfs fscdsconv: INFO: V-3-26037: Would you like to proceed further with migration? (ynq) y
UX:vxfs fscdsconv: INFO: V-3-26130: There are no files violating the CDS limits for this target.
UX:vxfs fscdsconv: INFO: V-3-26047: Byteswapping in progress ...
UX:vxfs fscdsconv: ERROR: V-3-25656: Overflow detected
UX:vxfs fscdsconv: ERROR: V-3-24418: fscdsconv: error processing primary inode list for fset 999
UX:vxfs fscdsconv: ERROR: V-3-24430: fscdsconv: failed to copy metadata
UX:vxfs fscdsconv: ERROR: V-3-24426: fscdsconv: Failed to migrate.
Cause
fscdsconv command takes a filename argument which is used as a recovery failure, to be used to restore the original file system in the event of failure while File System (FS) conversion is in progress. This file has two parts: control part and data part.
Control part is used to store the information about all the metadata like inodes and extents etc. Length of control part is being underestimated for some file systems, where there are few number of inodes, but average number of extents per file is very large (as noticed in fsadm -E report).
# fsadm -E /mnt
Extent Fragmentation Report
Total Average Average Total
Files File Blks # Extents Free Blks
588 1608266 370 79283862
blocks used for indirects: 4632
% Free blocks in extents smaller than 64 blks: 0.01
% Free blocks in extents smaller than 8 blks: 0.00
% blks allocated to extents 64 blks or larger: 100.00
Free Extents By Size
1: 34 2: 30 4: 18
8: 176 16: 181 32: 195
64: 187 128: 166 256: 165
512: 157 1024: 141 2048: 2961
4096: 486 8192: 223 16384: 65
32768: 570 65536: 203 131072: 95
262144: 48 524288: 9 1048576: 4
2097152: 1 4194304: 0 8388608: 0
16777216: 0 33554432: 0 67108864: 0
134217728: 0 268435456: 0 536870912: 0
1073741824: 0 2147483648: 0
Resolution
Fixed the issue by making recovery file sparse, data part starts after 2TB offset, control part can do allocating writes to the hole from the begining of the file. The fix is available in the following VxFS patch releases. The Redhat fix has issue with creating sparese file bigger than 2TB, so need to use vxfs filesystem for the recovery file for below HF, and it will be fixed on public patch later.
Platform Base Version Patch fix
Solaris SPARC - 5.1SP1 - 5.1SP1RP4HF1
RedHat - 6.1 - 6.1.0.085 (one-off HotFix)
RedHat - 6.1 - 6.1.0.202
Please contact Veritas Enterprise Support to obtain these patches.
Issue/Introduction
ERROR: V-3-25656: Overflow detected error when converting from Solaris to Linux platform (and vice-versa).
Additional Information
