Too many DASDs accessible to Linux for read

Edit online

Linux® system restart might be made slower due to processing many DASDs. This processing can cause network time-outs and other failures.

There is a problem that might occur when Linux is started on a server in an enterprise environment, in which the system can reach and identify many disks in the storage network. If the access for a system is not limited by a restrictive authentication mechanism, it is possible that when Linux starts, the scanning of all the storage information could take a long time due to the large number of disks.

Linux system restart time in this situation could be hours. The external symptoms can be erratic behavior immediately after system restart, including unsuccessful network connections and various failures. The problem could stop suddenly when the scanning of the storage information in the background completes.

The problem was resolved on the test system by modifying the /etc/zipl.conf file to exclude access to all disks by device number, except those disks that are of interest. The exceptions must also include Fibre Channel Protocol (FCP), cryptographic, and networking devices, as required. If your system is running Linux on z/VM®, the device address from the z/VM console (typically 0009) must be added back, or the system cannot restart.

An important caution is to include all system disks, including back up system disks, because Linux accesses only those disks specified in the zipl.conf file (and made active with the zipl command before shutdown).

In this example, in the parameters section, the cio_ignore=all command makes all devices inaccessible to this system. Then the devices are added back individually, by listing them with each device number preceded by an exclamation mark (!). Bold is used to illustrate these devices.
IMPORTANT: The list of disks specified with the cio_ignore= command must all appear on one continuous line. This list is illustrated with line breaks for the documentation, but in the file it must be specified without line breaks from cio_ignore to TERM=dumb". Incorrect configuration information in the zipl.conf file might prevent the system from restarting. 
[SLES_10_SP2]
    image = /boot/image-2.6.16.60-0.42.4-default
    target = /boot/zipl
    ramdisk = /boot/initrd-2.6.16.60-0.42.4-default,0x1000000
    parameters = "root=/dev/disk/by-id/ccw-IBM.750000000L5611.2be4.35-part1 
    cio_ignore=all,!7137,!7220,!7239,!7339,!7435,!7436,!743a,!7535,!7536,!7539,
!753a,!7735,!7736,!773a,!78b2,!78b3,!78b4,!79b2,!79b3,!79b4,!7ab2,!7ab4,!7ab6,!
7bb2,!7c82,!7c83,!7c84,!7c85,!7caf,!7cb0,!7cb3,!7cb4,!7db4,!7db6,!7eaf,!7eb0,!7
eb4,!0420,!0421,!0422,!423,!0424,!0425,!0426,!0440,!0441,!0442,!443,!0444,!0445
,!0446,!0600,!0601,!0602,!0610,!0611,!0612,!0620,!0621,!0622,!0630,!0631,!0632,
!0780,!0781,!0782,!0783,!0784,!0785,!0786 dasd=7435,7436,743a,7535,7436,7436,7c
84,7c85,753a,78b4,79b2,7ab2,7c83,7ab4,7caf,79b3,79b4,7cb0,7cb4,7db6,7eaf,7eb0,7
ab6,7137,7239,7339,7539,7bb2 TERM=dumb"

After changing the /etc/zipl.conf file, enter the zipl command to make the change effective for the next system restart.