Solaris 10 svc:/milestone/single-user timeout during patch run

Hello all,
when doing a patch run (Oct 07 baseline) on a Solaris 10 client the machine rebooted into single user as per normal. However the time for the patches to be applied in single user took so long the single-user milestone timed out and the machine went into maintenance mode. The machine did continue to complete the rest of the patch run. Anyone else seen this and have a work around? I don't really want to blanket increase the service timeout. Selected console output below..
( 27 / 37 ) Installing '127111-02' - succeeded.
( 28 / 37 ) Performing 'Mandatory reboot with reconfigure and action as single-user'
( 29 / 37 ) Installing '127753-02' Nov 22 10:30:08 svc.startd[7]: svc:/milestone/single-user:default: Method or service exit timed out. Killing contract 31.Nov 22 10:30:08 svc.startd[7]: milestone/single-user:default failed: transitioned to maintenance (see 'svcs -xv' for details)
Requesting System Maintenance Mode
(See /lib/svc/share/README for more information.)
Console login service(s) cannot run
Root password for system maintenance (control-d to bypass): Nov 22 10:30:20 svc.startd[7]: network/ntp:default failed repeatedly: transitioned to maintenance (see 'svcs -xv' for details)
Nov 22 10:30:20 svc.startd[7]: failed to abandon contract 45: Permission denied
- succeeded.
( 30 / 37 ) Installing '119757-09' - succeeded.
( 31 / 37 ) Installing '127737-01' - succeeded.
( 32 / 37 ) Installing '127718-01' - succeeded.
( 33 / 37 ) Installing '125184-04' - succeeded.
( 34 / 37 ) Installing '124444-01' - succeeded.
( 35 / 37 ) Installing 'postpatch' - succeeded.
( 36 / 37 ) Performing 'Reboot'
( 37 / 37 ) Performing 'Reboot with reconfigure'

Hi, yes I see it all the time and it's normal and in fact documented in the release notes:
http://docs.sun.com/app/docs/doc/820-0428/6nc5u3kou?l=en&a=view#smftimeout
One workaround is to ignore it and not do anything! In fact I think that is the preferred workaround.
Another question, just out of curiosity, what does your "postpatch" Post action script do? Just wondering if it might be something useful for other sites too?

Similar Messages

Unpredictable behaviour of svc:/milestone/multi-user-server:default

I have a Solaris 10 U4 running on SPARC(PrimePower H/W).
The machine is not going to run-level 3 (i.e the service svc:/milestone/multi-user-server:default ) is in maintenance mode inspite of all the required dependency for the specified services are met.
This is the output of svcs -l /milestone/multi-user-server command.
root@jupiter>svcs -l /milestone/multi-user-server
fmri svc:/milestone/multi-user-server:default
name multi-user plus exports milestone
enabled true
state maintenance
next_state none
state_time December 19, 2008 3:40:11 PM CET
logfile /var/svc/log/milestone-multi-user-server:default.log
restarter svc:/system/svc/restarter:default
dependency optional_all/none svc:/system/cluster/cl-svc-enable:default (online)
dependency require_all/none svc:/milestone/multi-user (online)
dependency optional_all/none svc:/application/management/dmi (disabled)
dependency optional_all/none svc:/application/management/snmpdx (disabled)
dependency optional_all/none svc:/network/rpc/bootparams (disabled)
dependency optional_all/none svc:/network/samba (disabled)
dependency optional_all/none svc:/network/nfs/server (disabled)
dependency optional_all/none svc:/network/winbind (disabled)
dependency optional_all/none svc:/network/wins (disabled)
dependency optional_all/none svc:/network/rarp (disabled)
dependency optional_all/none svc:/network/dhcp-server (disable)
Regrads,
Rahul.

rahulsingh wrote:
I have a Solaris 10 U4 running on SPARC(PrimePower H/W).
The machine is not going to run-level 3 (i.e the service svc:/milestone/multi-user-server:default ) is in maintenance mode inspite of all the required dependency for the specified services are met.Dependencies aren't sufficient for online. They're sufficent to begin bringing it online.
This is the output of svcs -l /milestone/multi-user-server command.
root@jupiter>svcs -l /milestone/multi-user-server
fmri svc:/milestone/multi-user-server:default
name multi-user plus exports milestone
enabled true
state maintenanceThis milestone isn't "offline", it's in "maintenance". So it tried to come online (after the dependencies were met), but failed for some reason. You should check the log file for the service.
/var/svc/log/milestone-multi-user-server:default.log
Darren

Solaris 10 Milestone Dependency

Hi all,
I am running Oracle Solaris 10u8 on a x86 Machine.
I have read in Solaris Admin 2 document and see that there is a special service in Solaris 10 named "Milestone".
SA2 document saids:
"A milestone is a special type of service which is made up of a defined set of other services.
A milestone can be regarded as a system state to reach. This system state requires a defined set of services to be running. These services depend on other services being available. Hence, there is a hierarchy of dependency relationships."
In my machine, I tried to issue below command to find all the milestones in Solaris OS:
root@x4150 # svcs | grep milestone
online Oct_14 svc:/milestone/name-services:default
online Oct_14 svc:/milestone/devices:default
online Oct_14 svc:/milestone/network:default
online Oct_14 svc:/milestone/single-user:default
online Oct_14 svc:/milestone/sysconfig:default
online Oct_14 svc:/milestone/multi-user:default
online Oct_14 svc:/milestone/multi-user-server:default
However, I see that the "svc:/milestone/name-services:default" milestone is ONLINE despite of the services on which it depends are all disabled.
root@x4150 # svcs -d svc:/milestone/name-services:default
STATE STIME FMRI
disabled Oct_14 svc:/network/rpc/nisplus:default
disabled Oct_14 svc:/network/nis/client:default
disabled Oct_14 svc:/network/dns/client:default
disabled Oct_14 svc:/network/ldap/client:default
I do not understand why the "svc:/milestone/name-services:default" milestone can be ONLINE. I think that all 4 services (nisplus, nis, dns and ldap) should be ONLINE before this milestone can be ONLINE.
Please help me to understand this situation. Is the any misunderstand of me about Milestone in Solaris here?
Thanks so much for any help and suggestion.
BR,
HuyNQ.

The service manifest for milestone/name-services defines that the grouping for all 4 services it depends on is "optional_all".
That means these services must be online, disabled or in maintenance. The other types of grouping are require_all, require_any and exclude_all.
# svcs -l milestone/name-services
fmri svc:/milestone/name-services:default
name name services milestone
enabled true
state online
next_state none
state_time Sun Aug 01 16:17:38 2010
logfile /var/svc/log/milestone-name-services:default.log
restarter svc:/system/svc/restarter:default
dependency optional_all/none svc:/network/dns/client (online)
dependency optional_all/none svc:/network/ldap/client (disabled)
dependency optional_all/none svc:/network/nis/client (disabled)
dependency optional_all/none svc:/network/rpc/nisplus (disabled)

Automatic scrip execution on the console in single user mode on Solaris 10

Hi All,
On Solaris 8, I am able to launch my script automatically on the console when the system comes up into single user mode without having to enter the password. However, on Solaris 10, the script is display in the foreground on the console. When i do a ps -ef , I see the process running. Can someone please tell me how to get this to work on solaris 10?
To automatically boot the system into single user mode:
# svcadm milestone -d milestone/single-user:default
Files edited for this to work on solaris 8:
To login without a password:
# cat > /etc/default/sulogin <<EOF
PASSREQ=NO
CONSOLE=/dev/console
ALTSHELL=YES
EOF
# echo "/sbin/su -" > /sbin/rc1
(Solaris 10, I used rcS instead of rc1)
my script is called from the /.profile when the system comes up into single user mode.
I did the same on solaris 10, but no success.
Please let me know if you have any suggestions.
Thank you,
Brian

try
- add 'set -x' to your /etc/profile
- add 'set -x' to /etc/rc? scripts
- check /etc/vfstab for a nfs mountpoint not using 'bg'

/var/run/name_service_door - zones stuck in single user mode

Hello,
I have two Sol10 boxes - e4500 & x86. Both have multiple zones configured which have been working for quite a while. I was enabling audit logging on these machines (perzone) which required a reboot.
After a reboot, the zones won't come out of the singleuser milestone. Trying to ping the global zone by name (nsswitch is files, dns), it hangs. A truss reveals:
open64("/var/run/name_service_door", O_RDONLY) = 3
fcntl(3, F_SETFD, 0x00000001) = 0
door_info(3, 0xFF16F7A8) = 0
door_call(3, 0xFFBFD790) (sleeping...)
and hangs forever.
I threw the latest patch cluster at both and it had no effect.
Each zone has a /etc/sysidcfg, /etc/.NFS4inst_state.domain, and proper nsswitch.conf/resolv.conf that had been working (rebooting) until I tried to enable auditing. This happened a while ago and a total uninstall/reinstall of the zones fixed the issue.
svcs reveals
offline 13:11:49 svc:/network/nfs/client:default
offline 13:11:49 svc:/system/filesystem/autofs:default
offline 13:11:49 svc:/system/system-log:default
offline 13:11:49 svc:/network/smtp:sendmail
offline 13:11:49 svc:/milestone/multi-user:default
offline 13:11:49 svc:/milestone/multi-user-server:default
offline 13:11:50 svc:/application/print/ipp-listener:default
offline 13:11:53 svc:/network/ssh:default
#svcadm milestone svc:/milestone/multi-user:default
Requesting System Maintenance Mode
(See /lib/svc/share/README for more information.)
Console login service(s) cannot run
Root password for system maintenance (control-d to bypass):
At this point I'm not sure how to debug this issue. Any ideas?
Thanks
Zetan

Hi, snikpmup.
Your panic log cites a page fault:
panic(cpu1 caller 0x001A3135): Unresolved kernel trap (CPU 1, Type 14=page fault)…
This is a common fault type, indicating the processor could not access either data memory or the next instruction to be executed. Potential causes include RAM problems, hard disk corruption, and programming errors.
The panic log you posted is incomplete, so there's not much else to say based on what you've posted. I do note some messages prior to the panic indicating M-Audio devices: you might want to check with those folks re: Mac Pro compatibility, any necessary updates, etc.
I suggest following the full set of troubleshooting steps cited in my "Resolving Kernel Panics" FAQ. The FAQ includes step-by-step instructions for identifying and resolving some of the most common causes of kernel panics. The FAQ is a roadmap: start at the beginning and work systematically through the end, i.e. the "If all else fails..." section.
Good luck!
Dr. Smoke
Author: Troubleshooting Mac® OS X
Note: The information provided in the link(s) above is freely available. However, because I own The X Lab™, a commercial Web site to which some of these links point, the Apple Discussions Terms of Use require I include the following disclosure statement with this post:
I may receive some form of compensation, financial or otherwise, from my recommendation or link.

Can't enable root user in single user mode

hey, im having a problem. I cant enable the root user in single user mode. This is what I do:
1.boot in single user mode
2.I run the command /sbin/fsck -fy
3.I run the command /sbin/mount -uw /
4.I run the command passwd
5.I type the password and then I retype it, but it gives me an error saying passwd: cant change password for root the account is not enabled (thats not exactly what it said, but you get the point)
this worked in leopard, why dosent it work now? can someone give me an alternative way or tell me what im doing wrong?
P.S. give me an actual answer not a why enable root? and I dont want to use directory utility.

The command is wrong man. Do the following:
Boot into S.U. Mode (Cmd+S).
once the terminal prompt opens, type the following:
mount -uw / ("mount"+space+"-uw"+space+slash)
launchctl load /System/Library/LaunchDaemons/com.apple.opendirectoryd.plist
passwd root
(after you press enter it will ask the password for the root user. Type it, even if it won't show any character. Then type again to confirm. Once you had finished type: exit).
Test it and post the results.

How to restart computers in single user mode?

I 'd like to use ARD to send a command to reboot target computers in Single User mode (and then run line of command).

have you considered just using safeboot? safeboot dose many of the same things applejack dose, and it's built into the mac os x.
Safeboot using ARD:
You can start a safeboot by send unix command, as root, command:
nvram boot-args="-x -v"
send command reboot computer
After the reboot, you could then send unix command, as user root, command:
sudo nvram boot-args=""
send command reboot computer
NOTE: some features do not work while the computer is in safeboot, so it's a good idea to have the computer plugged in by ethernet, because most macs still work over ethernet in safeboot.
Or you could just tell the user to reboot the computer holding the shift key, and when they get to the login window with the words "safeboot" to reboot the computer again.
FYI: in the past i've ran safeboot, and watching the output, I could see it repaired the disk and permistions. soo if I had a disk that needed repairing, you could in thory try safeboot to fix the issue. The computer already runs a RAM test every time you power on the computer.

Can't 'reboot' out of single-user mode.

Went into single-user mode earlier to run fsck. After everything had finished, and a couple of minor things repaired, typed in 'reboot' and... nothing happened. There didn't seem to be a prompt anymore to accept an input. I eventually just used the master switch. This was the first time I'd tried this since updating to Tiger. Is this a Tiger thing?

reboot is not what you type at single user mode. That is not a Unix command. In single user mode you type:
/sbin/mount -uw /
followed by
exit
or
logout
to leave it. You may be confusing it with open firmware, which you type:
mac-boot
to exit.

Applying Solaris Patches that Require Single User Mode

Some OS patches require single user mode and reboot. Does the standard Solaris plugin patch management handle this? If not, could I write my own patch component and simply add "init 1; patchadd; init 6" to the execnative?
Thanks - Jay

I can't speak about the Solaris plugin, unfortunately, but a single execnative call wouldn't cover what you're trying to accomplish. There's an RFE for the core SPS system to add this capability, but I don't know that it will be added soon. Remember that when you switch runlevels, a number of start and stop scripts are invoked, shutting down and starting up a number of services. To do what you want to do, you'd have to have a start script for the RA in runlevel 1. You'd also have to have a way for the RA to execute a command on startup which isn't possible today. There'd have to be a way for the MS to deposit a command for the RA to execute, restart the RA in runlevel 1, the RA would have to read the command on startup and execute it, then restart the RA to runlevel 3 (or 6, whatever is your default or previous runlevel). It's all doable, to be sure, but not with changes to the core product.

How to mount USB & CDROM drives from single user mode - Solaris boot disk?

Hi All,
I need to carry out ufsrestore on a single newly replaced system disk (no redundancy / mirroring) from either USB or CDROM drives from the following steps:
( i ) GRUB => e, e, cdrom –s, Enter, b (boot from Solaris 10 x86 media).
( ii ) Choose menu 1 for Install Solaris Interactive Mode to reach single user mode.Below are the commands that I have tried in single user mode without success:
mount –r –F pcfs /dev/dsk/c1t0d0p0 /cdrom (after unmount Solaris CD)
mount: /dev/dsk/c1t0d0p0 is not a DOS filesystem.
svcadm –v enable smserver
svcadm: Pattern ‘smserver’ doesn’t match any instances
svcadm enable autofs
svcadm: Pattern ‘autofs’ doesn’t match any instances
devfsadm
devfsadm: mkdir failed for /dev 0s1ed: Read-only file system
mkdir /mnt
mkdir: Failed to make directory “/mnt”; Read-only file systemPart of the problem is due to the current READ ONLY filesystems which does not allow the creation of mount point such as /mnt.
I have been able to create both / root (/dev/dsk/c1t0d0s0) and /export/home (/dev/dsk/c1t0d0s0) filesystems with newfs that are awaiting for restore using ufsrestore.
The only option I am left with is to attach a secondary disk to install Solaris 10 on either of them before bring the system to multi-user mode so that service such as volmgt, autofs, volfs are available to access USB & CDROM drives.
Any suggestion on how to resolve this issue?
Many thanks,
Jack

Hi JKGN,
Is the directory a located under the root filesystem. ie /a? Are there any other mount point for say USB drive as well? I am not in a position to try it out right now but will get back
to you soon on whether /a exist or not.
Btw, the last thing I managed to do on this system at the time was added a secondary 1TB internal disk with the intention to restore both / and /export/home data onto this disk while
making use of all the disk device management services such as the following services only available in a fully installed Solaris system (both single & multi-user modes) installed on the
primary disk:
# svcs smserver
STATE          STIME    FMRI
online         Jun_08   svc:/network/rpc/smserver:default
# svcs autofs
STATE          STIME    FMRI
online         Jun_08   svc:/system/filesystem/autofs:default
# devfsadm
# iostat -En
c1t0d0           Soft Errors: 0 Hard Errors: 0 Transport Errors: 0
Vendor: ATA      Product: SAMSUNG HD321KJ Revision: 0-11 Serial No:
Size: 320.07GB <320072932864 bytes>
Media Error: 0 Device Not Ready: 0 No Device: 0 Recoverable: 0
Illegal Request: 41 Predictive Failure Analysis: 0
c0t0d0           Soft Errors: 0 Hard Errors: 5 Transport Errors: 0
Vendor: PHILIPS Product: DVD+-RW DVD8801 Revision: AD21 Serial
Size: 17.54GB <17538875392 bytes>
Media Error: 0 Device Not Ready: 5 No Device: 0 Recoverable: 0
Illegal Request: 10 Predictive Failure Analysis: 0
c2t0d0           Soft Errors: 0 Hard Errors: 0 Transport Errors: 0
Vendor: JetFlash Product: Transcend 16GB   Revision: 1100 Serial No:
Size: 0.00GB <0 bytes>
Media Error: 0 Device Not Ready: 0 No Device: 0 Recoverable: 0
Illegal Request: 7 Predictive Failure Analysis: 0
# rmformat
Looking for devices...
     1. Volmgt Node: /vol/dev/aliases/cdrom0
        Logical Node: /dev/rdsk/c0t0d0s2
        Physical Node: /pci@0,0/pci-ide@1f,1/ide@0/sd@0,0
        Connected Device: PHILIPS DVD+-RW DVD8801 AD21
        Device Type: DVD Reader/WriterHowever, I was disappointed that "iostat -En" has not found the secondary disk (1 TB) even though it was available earlier, for Solaris installation as the only disk on the same system.
As a result, I am left with no option but the need to rebuild this system with only a single primary disk while in single user mode. However, the single user mode that I am in is the
one from Solaris installation disk with restrictive (_cannot create folder for mount point or running disk management utilities_) capability such as those services just covered, compared to one from a completed Solaris installation system with full access to all filesystems and utilities / commands in general.
I would very much value your assistance on how to mount both CD & USB in this restrictive limited Read Only OS (assume that it is running from memory) in order to get complete access to
the blank primary disk so that full restore with ufsrestore could take place.
Thanks in advance,
Jack
Edited by: 797805 on 9/06/2012 04:15

Proper procedure for patching from Single User mode

Typically when I install a patch cluster from Sun, I do a sanity reboot from the console of the server using:
shutdown -y -g0 -i6
When the system comes back online, I log into the console again and then do:
shutdown -y -g0 -i0 (to go into OBP)
then
boot -s (to go into single user mode)
The procedure above was given to me from a Sun technician.
Then I install the patch cluster and reboot. It has come to my attention that Sun recommends breaking any mirrors between your disks before patching. I wanted to know what is the best way to do this for both Veritas Volume Manager and Solaris Volume Manager. For Veritas Volume Manager, I was thinking of going into the vxdiskadm menu driven utility and choosing the option to "Remove a disk for replacement" for the rootmirror disk and then after a reboot to check that the patches did not cause a problem, go back into vxdiskadm and choose the option "Replace a failed or removed disk" and select the rootmirror which should then begin to automatically resync itself to the primary rootdisk. Any comments on if this is a proper way to do this or if someone has a better method, I would love to hear it. I am assuming a system with just two internal disks: c1t0d0s2 and c1t1d0s2
Also, if anyone can comment on how to do this with Solaris Volume Manager or if it is required would be great also.
Thanks much for any advice.

Typically when I install a patch cluster from Sun, I
do a sanity reboot from the console of the server
using:
shutdown -y -g0 -i6
When the system comes back online, I log into the
console again and then do:
shutdown -y -g0 -i0 (to go into OBP)
then
boot -s (to go into single user mode)
The procedure above was given to me from a Sun
technician.Not a bad thing to check reboot before patching, but I don't think it's in any official documentation that I'm aware of.
Then I install the patch cluster and reboot. It has
come to my attention that Sun recommends breaking any
mirrors between your disks before patching.Again, I don't know if it's "official", but if you have a backup copy that you could boot from, it does reduce the possibilities of critical problems from a bad patch.
I wanted
to know what is the best way to do this for both
Veritas Volume Manager and Solaris Volume Manager.
For Veritas Volume Manager, I was thinking of going
into the vxdiskadm menu driven utility and choosing
the option to "Remove a disk for replacement" for the
rootmirror disk and then after a reboot to check that
the patches did not cause a problem, go back into
vxdiskadm and choose the option "Replace a failed or
removed disk" and select the rootmirror which should
then begin to automatically resync itself to the
primary rootdisk. Any comments on if this is a proper
way to do this or if someone has a better method, I
would love to hear it. I am assuming a system with
just two internal disks: c1t0d0s2 and c1t1d0s2
Also, if anyone can comment on how to do this with
Solaris Volume Manager or if it is required would be
great also.Well, it'll work as you've described, but what if the patches fail? The disconnected mirror is not bootable. You'd have to go through an unencapsulation and other things from a CD.
I've often simply pulled one side of the mirror while the machine was shutdown. Since the mirror was valid prior to pulling, it will boot. If there's a problem, I shut down, swap disks, and boot from the untouched mirror. If no problem, I re-insert, reattach the disk to the diskgroup, then recover the volumes.
I don't know of any nice supported method of booting from an offline VxVM mirror that doesn't involve a very long series of steps. My method isn't supported, but it does work. If you have both disks in the machine at the same time though, it'll update the private regions. Don't do that until you're ready to sync up one way or the other. Test before doing it in production.
In any event, you should have a backup ready to go.
Darren

Patching zfs root in single user mode

Hi
I have to apply the Oracle recommended patchset to a server running on zfs root with zone roots also on zfs. The usual method of patching with Live Upgrade is giving some problems after the activation of the ABE. Can the patchset be applied to the server in the traditional single user mode
Any inputs would be appreciated

Yes, you can apply the patchset in single user mode, just keep in mind that your zones need to be bootable when patching as the patch installation will "partly" boot each zone when applying the patches. So if your zones rely on any NFS mounts etc, you will need to mount these first.
An alternative is to detach your zones during patching and then attach them using "zoneadm -z zonename attach -u" (or -U see the zoneadm(1M) man page for details on the differences).

Solaris 10 single user mode

Hi,
I had just install a solaris 10 update 9 in the SUN v240. Do i need this same version of update in order for me to run in Single User Mode?
Or it can do also with the current latest version= solaris 10 update13?
Thanks,

994409 wrote:
Hi,
I had just install a solaris 10 update 9 in the SUN v240. Do i need this same version of update in order for me to run in Single User Mode?
Or it can do also with the current latest version= solaris 10 update13?
Thanks,If you are asking whether you can boot that system using a CD/DVD of a newer release...
... then go do system maintenance tasks...
Yes, that is perfectly acceptable.
If you are asking whether you can boot that box using a CD/DVD of a newer release...
... then copy files fom that CD/DVD and overwrite existing files...
No, that would possibly muck up your system and you would have to wipe it and reinstall fresh.

I can only start my iBookG4 after deleting preferences. I have to go into single user mode to do so. If I don't go into single user mode, the computer shuts down during boot up. What might be the problem?

I can only start my iBookG4 after deleting preferences. I have to go into single user mode to do so. If I don't go into single user mode, the computer shuts down during boot up. What might be the problem? How can I fix this?

I do not have a startup disk. I did try fsck (based on what I learned from your reply-thank you). This failed (fsck); here is the message:
** Checking catalog file.
Invalid key length
(4, 8198)
** Volume check failed.
/dev/rdisk03 (hfs) EXITED WITH SIGNAL 8.
I am sure this is an easy fix; I just have no clue how.

Latest Patch cluster (Feb 11, 2010) unable to run in Single user mode.

Hi ,
This is my first post in this forum.
Hope you people will help to resolve the issue.
My issue - I have downloaded the latest patch cluster (Feb 11, 20100) and tried to install it from single user mode.
Nothing happened. I have waited for more than 30 mins, still no progres.
Then I killed the process, and tried to run it in Multi user mode (To verify any issue with extraction or file) -- Its working(and killed the process) !!!
Anyone facing this strange issue ?
Appreciate your help to complete the patching.
Thanks.
sas.

Hi,
If you're encountering issues with a patch cluster bundle your best bet is to contact [email protected] From the description it sounds like there may be a filesystem not mounted in SUM that the cluster is trying to patch.

Solaris 10 svc:/milestone/single-user timeout during patch run

Similar Messages

Maybe you are looking for