Sponsored Content
Operating Systems Solaris Reboot of Unix servers - recommended? Post 302104699 by Tornado on Friday 26th of January 2007 04:18:50 PM
Old 01-26-2007
Quote:
Originally Posted by mr_manny
Do you want to deal w/trying to identify what changed in the last how many months that could contributing to why your box goes into a panic during boot-up?

once a month in this shop Smilie
Thats what change control is for. If you are maintaining servers in a 24x7 enviroment you need proper change control, even if its not a 24x7 enviroment you should have something in place to track changes. There is no need to have a scheduled reboot. I bet everyone that recomends scheduled reboots has a linux background or has not maintained any 24x7 mission critical servers. Places like Banks, Telcos, Airports and large corporates require 5 9's availability(99.999% uptime).

If you don't have change control and you have a sysadmin in the team that is a cowboy and changes things without documenting them then your solution isn't a scheduled reboot(Sun Solaris doesn't need it). Your work practices need to change to include proper change control and system monitoring to make sure sysadmins are not making changes on the fly without following the change control proccess.

I have maintained servers with uptime over a couple of years that have rebooted fine.
Tornado
 

10 More Discussions You Might Find Interesting

1. UNIX for Dummies Questions & Answers

Reboot the Unix work station

Hi I'd like to know if to reboot a work station I need the root permission. If not what is the procedure ( commands or steps ) to reboot a work station. Cheers Mohsen (5 Replies)
Discussion started by: mohsen
5 Replies

2. UNIX for Dummies Questions & Answers

Any recommended book on Unix & Shell Programming

Any recommended book on Unix & Shell Programming (7 Replies)
Discussion started by: endeavour1985
7 Replies

3. UNIX for Dummies Questions & Answers

Unix recommended reading

Hello I was wondering if anyone had a Unix book recommendation. Just one book, that can give me a good foundation in Unix and has plenty of hands on exercises to follow along with on my Unix terminal? Thank you (1 Reply)
Discussion started by: vedder191
1 Replies

4. UNIX for Dummies Questions & Answers

best or recommended unix os for new user

hello, I want to learn more on unix, and being use to windows (VERY SAD i know) i want to learn more, but which unix is recommended for a new user to become novice to advance? also what is a great website to look into to gain more knowledge of unix, i have the commands pretty good in my head,... (0 Replies)
Discussion started by: Siten0308
0 Replies

5. AIX

"Out of Memory" after reboot for only Tuxedo DB servers on AIX 5.3

Hello all, I am working on an AIX 5.3L machine. I am currently trying to work on getting all of my Tuxedo servers booted for my app. However, some of these servers also use the Oracle 10g R2 client to connect to the 10g2 DB. AT FIRST, everything was working fine. Anyway, one night all... (3 Replies)
Discussion started by: swvahokie
3 Replies

6. UNIX for Advanced & Expert Users

shutdown and reboot unix server

Hi, please could someone advise the best command to shutdown and then for it to reboot back online again. Note: I shall be doing this from a telent session. regards venhart (3 Replies)
Discussion started by: venhart
3 Replies

7. Solaris

how to Implicitly execute some unix commands during/after Solaris reboot

I have to execute 2 commands everytime i reboot my Solaris machine. Simple "share" and "chown" commands. Is there anywhere (maybe a config file or something) where I can include these commands so that I do not have to manually run them everytime I reboot the machine? Any help appreciated... (2 Replies)
Discussion started by: lydiaEd
2 Replies

8. Shell Programming and Scripting

How to stop monitoring of servers at the time of reboot through shell scripting?

We have number of servers which belongs to platforms Solaris, AIX,HP-UX and LINUX. Monitoring tool 'Patrol Agent' process run on the servers to check for the server health and communicate with the Patrol server through the port 5181. During scheduled reboot and maintenance of servers we do receive... (1 Reply)
Discussion started by: subharai
1 Replies

9. UNIX for Dummies Questions & Answers

Recommended book to learn about unix administration?

My knowledge is weak in understanding NIS servers, setting up user accounts, mounting network file systems, clearing stale NFS handles, those sorts of things that I encounter but my IT server team handles. Can anyone recommend a good book on the subject? Something to demystify mount points,... (1 Reply)
Discussion started by: srhadden
1 Replies

10. Solaris

Unexpected error on reboot in UNIX

Hi all, My remote unix machine failed unexpectly, and I am unable to login to it. Here is what I can see on the screen - > Boot device: .... File and args: -i > Boot load failed. > The file just loaded does not appear to be executable. > {1} ok How can I fix this problem? Has... (1 Reply)
Discussion started by: bhakti.gandhi
1 Replies
stmsboot(1M)															      stmsboot(1M)

NAME
stmsboot - administration program for the Solaris I/O multipathing feature SYNOPSIS
/usr/sbin/stmsboot [-d | -e | -u | -L | -l controller_number] The Solaris I/O multipathing feature is a multipathing solution for storage devices that is part of the Solaris operating environment. This feature was formerly known as Sun StorEdge Traffic Manager (STMS) or MPxIO. The stmsboot program is an administrative command to manage enumeration of fibre channel devices under Solaris I/O multipathing. Solaris I/O multipathing-enabled devices are enumerated under scsi_vhci(7D), providing multipathing capabilities. Solaris I/O multipathing-disabled devices are enumerated under the physical controller. In the /dev and /devices trees, Solaris I/O multipathing-enabled devices receive new names that indicate that they are under Solaris I/O multipathing control. This means a device will have a different name from its original name (following installation) when it is under Solaris I/O multipathing control. The stmsboot command automatically updates /etc/vfstab and dump configuration to reflect the device names changes when enabling or disabling Solaris I/O multipathing. A reboot is required for changes to take effect. The following options are supported: -e Enables Solaris I/O multipathing on all fibre channel (fp(7D)) controller ports. Following this enabling, you are prompted to reboot. During the reboot, vfstab and the dump configuration will be updated to reflect the device name changes. -d Disables Solaris I/O multipathing on all fibre channel (fp(7D)) controller ports. Following this disabling, you are prompted to reboot. During the reboot, vfstab and the dump configuration will be updated to reflect the device name changes. -u Updates vfstab and the dump configuration after you have manually modified the configuration to have Solaris I/O multipathing enabled or disabled on specific fp(7D) controller ports. This option prompts you to reboot. During the reboot, vfstab and the dump configura- tion will be updated to reflect the device name changes. -L Display the device name changes from non-Solaris I/O multipathing device names to Solaris I/O multipathing device names. -l controller_number Display the device name changes from non-Solaris I/O multipathing device names to Solaris I/O multipathing device names for the speci- fied controller. Along with its primary function of enabling or disabling Solaris I/O multipathing, the stmsboot command is used to update vfstab and the dump configuration to reflect device name changes. For a system to function properly, you must configure the applications that consume the devices by old names to use the new names. The -L and -l options display the mapping between the old and new device names. These options work after the changes made to the Solaris I/O multipathing configuration have taken effect. For example, you can use these options following the reboot after invoking stmsboot -e. The old device names must exist in order to display the mappings. Example 1: Enabling Solaris I/O Multipathing Following OS Upgrade To enable Solaris I/O multipathing on all fibre channel (fp(7D)) controller ports run: # stmsboot -e Example 2: Disabling Solaris I/O Multipathing To disable Solaris I/O multipathing on all fibre channel (fp(7D)) controller ports, run: # stmsboot -d Example 3: Enabling Solaris I/O Multipathing on Selected Ports You want to enable Solaris I/O multipathing on some fibre channel controller ports and disable the feature on the rest. You edit the fp.conf file (see fp(7D)) to enable or disable Solaris I/O multipathing on specific controller ports. You then run the following command to have vfstab and the dump configuration updated to reflect the new device names: # stmsboot -u See attributes(5) for descriptions of the following attributes: +-----------------------------+-----------------------------+ | ATTRIBUTE TYPE | ATTRIBUTE VALUE | +-----------------------------+-----------------------------+ |Architecture |SPARC | +-----------------------------+-----------------------------+ |Availability |SUNWcsu, SUNWcslr | +-----------------------------+-----------------------------+ |Interface Stability |Obsolete | +-----------------------------+-----------------------------+ SEE ALSO
dumpadm(1M), ufsdump(1M), dumpdates(4), vfstab(4), fcp(7D), fctl(7D), fp(7D), qlc(7D), scsi_vhci(7D) Consult the Sun StorEdge Disk Tray [or Subsystem] Administrator's Guide for the T3, 3910, 3960, 6120, and 6320 storage subsystems. Sun StorEdge Traffic Manager Installation and Configuration Guide Solaris I/O multipathing is not supported on all devices. After enabling Solaris I/O multipathing, only those devices that Solaris I/O mul- tipathing supports are placed under Solaris I/O multipathing control. Non-supported devices remain as before. For Solaris releases prior to the current release, the -e and -d options remove the mpxio-disable property entries from fp.conf file (see fp(7D)) and add a global mpxio-disable entry to fp.conf. The current release of the Solaris operating system does not support the mpxio-disable property. Solaris I/O multipathing is always enabled. If you want to disable multipathing, you must use the mechanisms provided by the HBA drivers. See fp(7D). Enabling Solaris I/O Multipathing on a Sun StorEdge Disk Array The following applies to Sun StoreEdge T3, 3910, 3960, 6120, and 6320 storage subsystems. To place your Sun StorEdge disk subsystem under Solaris I/O multipathing control, in addition to enabling Solaris I/O multipathing, the mp_support of the subsystem must be set to mpxio mode. The preferred sequence is to change the subsystem's mp_support to mpxio mode, then run stmsboot -e. If Solaris I/O multipathing is already enabled but the subsystem's mp_support is not in mpxio mode, then change the mp_support to mpxio mode and run stmsboot -u. Refer to the Sun StorEdge Administrator's Guide for your subsystem for more details. ufsdump Users The ufsdump command keeps records of the filesystem dumps in /etc/dumpdates (see dumpdates(4)). Among other items, the records contain device names. An effect of the "active" stmsboot options (-e, -d, and -u) is to change the device name of a storage device. The stmsboot command does not modify the dumpdates file. Because of this, the dumpdates records will refer to the old device names, that is, the device names that were in effect before you ran stmsboot. The effect of this device name-dumpdates disagreement is that, following use of stms- boot, ufsdump will be processed as if no previous dump had ever been made, thus dumping the entire filesystem (effectively, a level 0 dump). Procedure to Use stmsboot in Sun Cluster Environment If possible, use stmsboot -e before you start installing Sun Cluster software. After you run stmsboot, you install Sun Cluster software as you normally would. If you install Sun Cluster software before running stmsboot, you must use the following procedure. On each machine in the cluster on which you want to enable the Solaris multipathing feature, enter: # stmsboot -e ...and allow the system to reboot. When the system comes up, enter the following two commands: 1. # /usr/cluster/bin/scdidadm -C 2. # /usr/cluster/bin/scdidadm -r The preceding commands update did mappings with new device names while preserving did instance numbers for disks that are connected to multiple cluster nodes. did instance numbers of the local disks might not be preserved. For this reason, the did disk names for local disks might change. 3. Update /etc/vfstab to reflect any new did disk names for your local disks. 4. Reboot the system. To disable the Solaris multipathing feature, use stmsboot -d (instead of stmsboot -e), then follow the procedure above. To view mappings between the old and new device names, run stmsboot -L. To view did device name mappings, run /usr/cluster/bin/scdidadm -L. 3 Mar 2005 stmsboot(1M)
All times are GMT -4. The time now is 03:03 AM.
Unix & Linux Forums Content Copyright 1993-2022. All Rights Reserved.
Privacy Policy