Hello, I mentioned this problem a year ago here and filed 6792701 and I know it has been discussed since. It should have been fixed in snv_118, but I can still trigger the same problem. This is only triggered if the creation of a large file is aborted, for example by loss of power, crash or SIGINT to mkfile(1M). The bug should probably be reopened but I post it here since some people where seeing something similar. Example and attached zdb output: filer01a:/$ uname -a SunOS filer01a 5.11 snv_130 i86pc i386 i86pc Solaris filer01a:/$ zpool create zpool01 raidz2 c4t0d0 c4t1d0 c4t2d0 c4t4d0 c4t5d0 c4t6d0 filer01a:/$ zpool create zpool01 raidz2 c4t0d0 c4t1d0 c4t2d0 c4t4d0 c4t5d0 c4t6d0 filer01a:/$ zfs list zpool01 NAME USED AVAIL REFER MOUNTPOINT zpool01 123K 5.33T 42.0K /zpool01 filer01a:/$ df -h /zpool01 Filesystem Size Used Avail Use% Mounted on zpool01 5.4T 42K 5.4T 1% /zpool01 filer01a:/$ mkfile 1024G /zpool01/largefile ^C filer01a:/$ zfs list zpool01 NAME USED AVAIL REFER MOUNTPOINT zpool01 160G 5.17T 160G /zpool01 filer01a:/$ ls -hl /zpool01/largefile -rw------- 1 root root 1.0T 2010-01-22 15:02 /zpool01/largefile filer01a:/$ rm /zpool01/largefile filer01a:/$ sync filer01a:/$ zfs list zpool01 NAME USED AVAIL REFER MOUNTPOINT zpool01 160G 5.17T 160G /zpool01 filer01a:/$ df -h /zpool01 Filesystem Size Used Avail Use% Mounted on zpool01 5.4T 161G 5.2T 3% /zpool01 filer01a:/$ ls -l /zpool01 total 0 filer01a:/$ zfs list -t all zpool01 NAME USED AVAIL REFER MOUNTPOINT zpool01 160G 5.17T 160G /zpool01 filer01a:/$ zpool export zpool01 filer01a:/$ zpool import zpool01 filer01a:/$ zfs list zpool01 NAME USED AVAIL REFER MOUNTPOINT zpool01 160G 5.17T 160G /zpool01 filer01a:/$ zfs -ddd zpool01 <cut> Object lvl iblk dblk dsize lsize %full type <cut> 5 5 16K 128K 160G 1T 15.64 ZFS plain file </cut> |
zpool01.zdb
Description: Binary data
_______________________________________________ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
