Chapter 24. Storage
e4defragutility in the e2fsprogs package is not supported in Red Hat Enterprise Linux 7.0, and is scheduled to be removed in Red Hat Enterprise Linux 7.1.
- If XFS metadata checksums are enabled specifying the
-m crc=1option to the
mkfs.xfscommand, the kernel prints the following warning message when the file system is mounted:
Version 5 superblock detected. This kernel has EXPERIMENTAL support enabled! Use of these features in this kernel is at your own risk!Note that this CRC functionality is available for testing in Red Hat Enterprise Linux 7.0 and is planned to be fully supported in Red Hat Enterprise Linux 7.1.
- When the
systemddaemon parses the crypttab file in case of a non-LUKS device for swap with a random key, it uses the
ripemd160hash by default, which is not allowed in FIPS mode. To work around this problem, add the
hash=setting with an algorithm approved for FIPS to the particular crypttab line. For example:
swap /dev/sda7 /dev/urandom swap,hash=sha1.
- If the kernel thin provisioning code detects a device failure or runs out of metadata space, it sets a flag on the device to indicate that it needs to be checked. Currently, LVM tools do not perform this check automatically. To work around this problem, execute the
thin_check --clear-needs-check-flagcommand to perform the check and remove the flag. Then run the
thin_repaircommand if necessary. Alternatively, you can add
thin_check_optionsin the global section of the
/etc/lvm.confconfiguration file to run the check automatically.
- In Red Hat Enterprise Linux 7, the behavior of the
kpartxutility has been changed, so it no longer adds the letter
pas a delimiter between a device name and a partition suffix unless the device name ends with a digit. When a device is renamed using the
multipathutility, the previous device name is simply replaced by a new name while the suffix stays unchanged, regardless of the delimiter. Consequently, the
kpartxbehavior is not followed, and changing device names ending in a digit to names ending in a letter, or the other way round, works incorrectly when using
multipathto rename devices. To work around this problem, choose one of the following three options:
- Remove the
multipathddaemon and add the devices again;
- Remove the devices manually by using the
kpartx -dcommand and then add them by running the
- Rename the devices by using the
kpartx -p pcommand for device names that are supposed to contain the delimiter and they do not, and the
kpartx -p ""command in cases when the delimiter is used redundantly.
empty-pre-postcleanup algorithm, which is used for deleting pre-post pairs of file system snapshots with empty diffs, does not work in Red Hat Enterprise Linux 7. To work around this problem, remove empty pre-post snapshot couples manually by using the
- The bigalloc feature for the
ext4file systems, which enables
ext4to use clustered allocation, is not supported in Red Hat Enterprise Linux 7.
- Placing the
/bootpartition on an LVM volume is not supported. Placing the
/bootpartition on a Btrfs subvolume is not supported either.
- While the rpc.svcgssd binary is included in the nfs-utils package in Red Hat Enterprise Linux 7.0, its use in new deployments is discouraged in favor of gssproxy. The rpc.svcgssd binary may be removed in later Red Hat Enterprise Linux 7 releases.
kernelcomponent, BZ#1061871, BZ#1201247
- When a storage array returns a CHECK CONDITION status but the sense data is invalid, the Small Computer Systems Interface (SCSI) mid-layer code retries the I/O operation. If subsequent I/O operations receive the same result, I/O operations are retried indefinitely. For this bug, no workaround is currently available.