- BZ#631541, BZ#631561
This updated package provides three new tools:
hyptop — a dynamic of a System z hypervisor environment in real time.
lsmem — lists the ranges of available memory and their status, whether online or offline.
chmem — sets a particular size or range of memory online or offline.
- BZ#631546
The s390-tools package now provides cmsfs-fuse, a tool that can mount a CMS disk as a writeable file system using the FUSE
infrastructure on Linux. This new tool allows you to read and write configuration files on CMS disks directly.
- BZ#597360
Previously, the wrong control unit type for CTC devices was used in the udev rules. As a result, CTC devices were not detected being detected. In this updated package, the rules are corrected and devices are detected properly on system startup.
- BZ#619504
Previously, the xcec-bridge utility did not process link-level headers in incoming packets. The utility therefore could not forward multicast network traffic. Now, xcec-bridge are taken in to account and multicast forwarding works.
- BZ#623250
When the ziomon
command exits successfully, it should return a value of 0
. Previously, faulty logic caused the command to return an exit status of 1
when run with the --help
or -v
arguments. The logic is now corrected and ziomon
returns 0
when run successfully with --help
or -v
.
- BZ#627692
Previously, the code used by qethconf to process devices assumed that the device subchannel was set to 0
. Consequently, devices where the subchannel was set to any other value were not processed and did not appear on the IPA list. Now, qethconf recognizes devices where the subchannel is set to values other than 0
and these devices appear correctly in the IPA list.
- BZ#631527
The cio_settle
kernel facility is a new mechanism by which processes in user space can monitor CIO actions. Previously, user-space processes could not wait for devices to become available, leading to possible race conditions, particularly as the system started and started processing CIO requests. This updated s390-utils package uses this mechanism to enable user space processes to wait for devices to become usable. Now that processes can wait for device availability, handling of all CIO actions is ensured and race conditions are avoided.
- BZ#633411, BZ#676706
Previously, triggering a dump too soon after a kernel panic could lead to an infinite panic–dump–IPL loop. This updated package introduces DELAY_MINUTES
as a new keyword for the etc/sysconfig/dumpconf
configuration file, and updates the dumpconf
manual page to describe its use. When configured, the new keyword delays the dump and therefore help to avoid situations where triggering the dump leads to a re-IPL loop.
- BZ#633420
Linux on System z might not provide a particular terminal or console. This updated package provides a new tool, ttyrun, which safely starts getty
programs and prevents re-spawns through the init program if a terminal is not available.
- BZ#633437
Previously, the zfcpdump utility could only read and write s390 format dumps. This updated package adds two new dump formats: ELF (source/target) and LKCD (target). Therefore the tool can now read ELF, s390 and LKCD and write ELF and s390 format dumps, allowing it to be used for dump format conversion. The ELF target format can be used to run the makedumpfile tool as a second step to compress the dump.
- BZ#633534, BZ#636849
OSX and OSM are new network interface types from zEnterprise for hybrid data (management) networking. Previously, znetconf did not handle OSX and OSM devices, and could not be used to configure them. This updated package updates the tool's internal tables so that znetconf handles these devices correctly. Additionally, new udev rules ensure that these devices come up when the system starts.
- BZ#636204
Previously, iucvtty passed the z/VM user ID of the originating guest virtual machine as an argument to the -h
option of the login program. Depending on the implementation of the login program, passing the user ID to login -h
can cause timeouts when the target system does not have a working network connection. Now iucvtty no longer passes the user ID and therefore avoids timing out during login.
- BZ#644935
This updated package adds -Q
as a new option to the tunedasd tool that allows it to show the reservation status of a given DASD in relation to the current Linux instance. Used on the command line, tunedasd -Q
returns the reservation status to standard out.
- BZ#649787
Previously, the format 7
label written by fdasd and dasdfmt was incorrect. Therefore, backups of Linux on System z disks from z/OS did not work when the disk was not fully partitioned. libvtoc now writes the format 7
label correctly and backups work correctly.
- BZ#651012
Previously, the cmsfs utilities crashed when they were used on filesystems with block sizes different from the underlying device. Users had to work around the issue by creating a filesystem with the same block size as the device. With this updated package, cmsfs utilities report the mismatch in block sizes but still work.
- BZ#658517, BZ#693365
Previously, cpluplugd contained incorrect checks. Therefore, when cpluplugd exited, it restored /proc/sys/vm/cmm_pages
to 0
, regardless of its previous value. Also, when where cmm_pages
was equal to cmm_inc
, cmm_pages did not correctly reach a cmm_min
of 0
during run-time. The incorrect checks in the cpluplugd utility are now fixed, so that /proc/sys/vm/cmm_pages
maintains its correct value, and the evaluation of cmm_min
is now correct.
- BZ#659828
Previously lsluns failed to report LUNs from the SAN Volume Controller (SVC). The strategy of lsluns is now changed to check if LUN 0
or the WLUN
is already available. If both LUNs are not available, LUN 0
is tried first; if this fails, WLUN
is tried.
- BZ#660361
Previously lsluns did not accept uppercase letters for hex digits in the FCP device or WWPN. With this update, lsluns accepts uppercase and lowercase letters.
- BZ#688140
The mon_statd
script contained a call to udevsettle
instead of udevadm settle
, which failed because udevsettle
doesn't exist. The call is now corrected and mon_statd
works correctly.
- BZ#688340
Previously, when fdasd
tried to write to an read-only disk, it would attempt to format an error message through libvtoc
, where it would cause a buffer overflow. Therefore, rather than a useful error message, users were presented with an error about a buffer overflow. The fdasd
tool now prints the error message directly and therefore avoids the buffer overflow.