Odd hang on a 5.6 system

Home » CentOS » Odd hang on a 5.6 system
CentOS 1 Comment

Hi All, I’ve been trying to trace the cause of a hang on a 5.6 i386 system.

After running for almost a year, it hung last week, when I plugged in a screen it was blank, machine was unresponsive to the keyboard, over the network ssh and other daemons didn’t respond but the thing has two network cards and routing from one to the other was still working. So the kernel was up and I suspected a dying disk. But smartctl -a revealed
nothing untoward. Didn’t see anything significant in the log files at the time. Logging had stopped when the machine hung.

The machine rebooted normally and has run for almost a week and hung again with the same symptoms. Again rebooted, nothing untoward in the logs and smartctl still OK. But shortly after I left site this was logged in /var/log/messages

Mar 27 16:52:04 cjcsrv kernel: INFO: task hald-addon-stor:2179 blocked
for more than 120 seconds.
Mar 27 16:52:04 cjcsrv kernel: "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
Mar 27 16:52:04 cjcsrv kernel: hald-addon-st D 00000197 2552 2179 2160 2170 (NOTLB)
Mar 27 16:52:04 cjcsrv kernel: d895bbbc 00000086 a3374d80 00000197 e89436f0 d895bbbc c084e894 0000000a
Mar 27 16:52:04 cjcsrv kernel: d891aaa0 a3468fc0 00000197 000f4240 00000000 d891abac c1506800 e43383c0
Mar 27 16:52:04 cjcsrv kernel: 00000000 00000086 00000000 e7ad2b80 c061f7ca 00000000 c1506844 d895bc0c
Mar 27 16:52:04 cjcsrv kernel: Call Trace:
Mar 27 16:52:04 cjcsrv kernel: [] cdrom_do_pc_continuation+0x0/0x2c [ide_cd]
Mar 27 16:52:04 cjcsrv kernel: [] schedule+0x9c6/0xa4f
Mar 27 16:52:04 cjcsrv kernel: [] wait_for_completion+0x6b/0x8f
Mar 27 16:52:04 cjcsrv kernel: [] default_wake_function+0x0/0xc
Mar 27 16:52:04 cjcsrv kernel: [] ide_do_drive_cmd+0xd7/0xfa
Mar 27 16:52:04 cjcsrv kernel: [] cdrom_queue_packet_command+0x35/0xbc [ide_cd]
Mar 27 16:52:05 cjcsrv kernel: [] poll_freewait+0x18/0x4c
Mar 27 16:52:05 cjcsrv kernel: [] do_sys_poll+0x314/0x339
Mar 27 16:52:05 cjcsrv kernel: [] cdrom_check_status+0x52/0x5d [ide_cd]
Mar 27 16:52:05 cjcsrv kernel: [] blk_end_sync_rq+0x0/0x1d
Mar 27 16:52:05 cjcsrv kernel: [] ide_cdrom_check_media_change_real+0x1a/0x34 [ide_cd]
Mar 27 16:52:05 cjcsrv kernel: [] media_changed+0x40/0x6e [cdrom]
Mar 27 16:52:05 cjcsrv kernel: [] check_disk_change+0x13/0x3b
Mar 27 16:52:05 cjcsrv kernel: [] cdrom_open+0x833/0x876 [cdrom]
Mar 27 16:52:05 cjcsrv kernel: [] avc_has_perm+0x3c/0x46
Mar 27 16:52:05 cjcsrv kernel: [] avc_has_perm+0x3c/0x46
Mar 27 16:52:05 cjcsrv kernel: [] __d_lookup+0x98/0xdb
Mar 27 16:52:05 cjcsrv kernel: [] avc_has_perm+0x3c/0x46
Mar 27 16:52:05 cjcsrv kernel: [] inode_has_perm+0x54/0x5c
Mar 27 16:52:05 cjcsrv kernel: [] kobject_get+0xf/0x13
Mar 27 16:52:05 cjcsrv kernel: [] get_disk+0x35/0x6e
Mar 27 16:52:05 cjcsrv kernel: [] exact_lock+0x7/0xd
Mar 27 16:52:05 cjcsrv kernel: [] kobj_lookup+0x10d/0x168
Mar 27 16:52:05 cjcsrv kernel: [] idecd_open+0x7b/0xa8 [ide_cd]
Mar 27 16:52:05 cjcsrv kernel: [] do_open+0x89/0x2cc
Mar 27 16:52:05 cjcsrv kernel: [] blkdev_open+0x0/0x44
Mar 27 16:52:05 cjcsrv kernel: [] blkdev_open+0x1c/0x44
Mar 27 16:52:05 cjcsrv kernel: [] __dentry_open+0xc7/0x1ab
Mar 27 16:52:05 cjcsrv kernel: [] nameidata_to_filp+0x19/0x28
Mar 27 16:52:05 cjcsrv kernel: [] do_filp_open+0x2b/0x31
Mar 27 16:52:05 cjcsrv kernel: [] do_sys_open+0x3e/0xae
Mar 27 16:52:05 cjcsrv kernel: [] sys_open+0x16/0x18
Mar 27 16:52:05 cjcsrv kernel: [] syscall_call+0x7/0xb
Mar 27 16:52:05 cjcsrv kernel: =======================
Mar 27 16:52:19 cjcsrv kernel: ide1: reset timed-out, status=0xd0

ide1 has a CD attached. Not essential, the CD could be unplugged. ide0 has the hard disk, hda, attached. Looking back through the logs there is another of these recorded before the previous hang. Any clues as to what this is telling me – other than something crashed.

One thought on - Odd hang on a 5.6 system

  • The process “hald-addon-storage” got stuck while trying to access the cd
    (“ide-cd”). It probably tried to poll the drive to check if there was a cd
    inserted. Unplugging the drive should do the trick although you could try
    to disable the polling by creating a file
    “/etc/hal/fdi/policy/99-custom.fdi” with the following content:

    < ?xml version="1.0" encoding="UTF-8"?>



    hald-addon-storage


    After doing so restart hald or reboot. hald should no longer poll the drive
    after this.

LEAVE A COMMENT