Sorry read "The following services while running the command" as "The following services failed to stop while running the command"

Hi Tony

I tried your suggestion stopping services from command line using service {servicename} stop command following the sequence in /etc/rc0.d.. The following services while running the command

snort
fetchmail
avahi-daemon
avahi-dnsconfd
snortsam
cyrus-imapd
nmb
smb
l7-filter
freshclam
winbind
lvm2-monitor
mdmpd
multipathd
netplugd


Note that the sequence in my case was pretty much simiar to urs (74 was HAL etc), so u can see some of the services before (and after) HAL gave an error.. So I'd assume when issuing a shutdown command, this is not the case, as I don't see any services failing to stop, prior to system hanging after HAL...

And the errors are also varying.. in some cases it says file not found, where as in some other it just says and error occurred.. While trying to stop lvm2-monitor, it gave an error saying "this is a dangerous operation and advised to use force-stop instead"..

I shall try to a do a BIOS upgrade following Alan's suggestion soon and let you know the result..

Cheers

Praveen

Hi Praveen,
Well, the BIOS update did seem to work initially but now I'm back to hanging shutdown intermittently. So I'm going to go through Tony's instructions as soon as I can to try to find the hanging process.

To update the BIOS on my ThinkServer I downloaded the BIOS update utility from Lenovo and followed the readme on that page. (Unzip the file, use the included rufus.exe to create a DOS bootable USB, start the server to BIOS settings and move USB device to first in the boot list, then boot from USB and run the update utility.)

(As a reminder I'm on COS 6.5)

Hi Alan and Tony...

Had been out of station and as such could not try out Tony's suggestions, but planning to do it today or tomorrow.. Good to know Alan's issue is resolved, hoping the same is the case with mine too.. Alan can you let me know the procedure for updating BIOS from clear OS command line. Have done this from Windows before, but not sure how to, from clear OS...

Thanks
Praveen

Thanks for the great information Tony. A BIOS update did the trick. That probably would have been the first thing I tried but I have two identical servers and the other worked fine. So I checked and yep the rev's were off. I'll note this page because the codes in the etc/r0.d are new to me and maybe Praveen's issue can be fixed with this information.

In my limited experience with this problem there have been two causes of this shutdown problem...

1. Buggy acpi or apm routines in the BIOS.

Make sure you have the latest BIOS for the mother-board. If no fix then try adding certain command line options to the "kernel" line in "/etc/grub.conf". A Google search on shutdown problems using linux with provide a plethora of suggestions. A BIOS upgrade to my old Atom D945GCLF2 board minimised, but not eliminated, shutdown hangs. Having easy physical access never pursued this further...

If you look in /etc/rc0.d you will see the shutdown sequence. The number after the "K" shows the sequence (ascending i.e. lowest number first off the block). On this machine here the HAL demon is 74 (K74haldaemon). Also at 74 are K74acpid, K74lm_sensors, K74nscd and K74ntpd. Interesting that acpid is part of the '74' series. Without issuing the shutdown command you could try working your way along from K01 upwards to see if you can locate what is hanging using the physically attached screen and keyboard. For example "service dropbox stop" would be the first from the list below, then "service smartd stop" etc working your way through it, but using the entries from your machine as your services will probably be a little different to those listed here...

[root@may rc0.d]# ls

K01dropbox       K20hddtemp        K39clamd       K74nscd         K88rsyslog

K01smartd        K25clearsyncd     K45arpwatch    K74ntpd         K89portreserve

K01snort         K25squid          K50dnsmasq     K75netfs        K89rdisc

K01syswatch      K25sshd           K50netconsole  K75ntpdate      K90network

K02avahi-daemon  K30postfix        K50vnstat      K75storage      K91firewall

K02snortsam      K30spamassassin   K50xinetd      K75udev-post    K92pppoe-server

K10saslauthd     K31amavisd        K60crond       K85mdmonitor    K99sysstat

K15gkrellmd      K35cyrus-imapd    K73slapd       K85messagebus   S00killall

K15htcacheclean  K35ddclient       K73winbind     K87named        S01halt

K15httpd         K35nmb            K74acpid       K87restorecond

K15svnserve      K35smb            K74haldaemon   K88auditd

K15webconfig     K36system-mysqld  K74lm_sensors  K88nslcd

[root@may rc0.d]# 

2. Samba or NFS shares talking too long to shutdown.
Haven't seen this in a long time. A type of reverse "chicken and egg" deadlock. The remote share hasn't shutdown, but the network has. System is waiting for confirmation from the remote end that the share has terminated, but as the network has stopped it will never happen. Simple test is to manually ensure all shares have terminated before issuing the shutdown command.

I am also having this problem, but with one of two servers running ClearOS 6.5. Both are Lenovo TS140's with two PCIe NIC's. I've been trying to determine if it's an ACPI / BIOS problem, or an OS issue. When I shut down or restart, the uncooperative system hangs after the HAL screen as you have described. So remote restarts aren't possible.