Solaris / OpenSolarisThis forum is for the discussion of Solaris, OpenSolaris, OpenIndiana, and illumos.
General Sun, SunOS and Sparc related questions also go here. Any Solaris fork or distribution is welcome.
Notices
Welcome to LinuxQuestions.org, a friendly and active Linux Community.
You are currently viewing LQ as a guest. By joining our community you will have the ability to post topics, receive our newsletter, use the advanced search, subscribe to threads and access many other special features. Registration is quick, simple and absolutely free. Join our community today!
Note that registered members see fewer ads, and ContentLink is completely disabled once you log in.
If you have any problems with the registration process or your account login, please contact us. If you need to reset your password, click here.
Having a problem logging in? Please visit this page to clear all LQ-related cookies.
Get a virtual cloud desktop with the Linux distro that you want in less than five minutes with Shells! With over 10 pre-installed distros to choose from, the worry-free installation life is here! Whether you are a digital nomad or just looking for flexibility, Shells can put your Linux machine on the device that you want to use.
Exclusive for LQ members, get up to 45% off per month. Click here for more info.
I currently am in the position that i need to setup a Solaris server to use on our network (our being where i work). Unfortunately i am not terribly familiar with Solaris and how it functions. So, this brings me here.
I am currently in the process of trying to set up mirrors of two of the disks. There are four physical disks in the server. I want to mirror the first to the second and third to the fourth. Now, i have setup six metadbs in total. Two on each of the non-system disks. Now this is where i am getting lost. What's my next step? How do i mirror the currently active system disk? Do i have to mirror each slice independently? Any help would be greatly appreciated.
What's the best and most stable way of doing it? This server will be under lots of stress and needs to be able to withstand lots of torture. The way it is implemented currently (with a much older server) is with meta devices.
Performing tests at my Job; ZFS is faster than meta devices...There is some overhead involved (but you had that anyway with meta devices ) But it's minor.
Our primary Web server (which we do most of our business on) is in the high 90s in idle time. zpool iostat shows little "disk trashing" (or in this case pool thrashing ); But we are running on T2000 servers...so it depends on your hardware (if you get a chance check out the Sun SPARC Enterprise T5220 Server...I like to get my hands on that )
Based on the original email, you said to mirror the first with the second disk. If that is the boot disk, you CAN'T use zfs and you have to use either solaris volume manager or veritas volume manager.
Also, with zfs, the syntax looked wrong. It should be zpool create mypool mirror first_disk second_disk (not first_disk mirror second_disk).
With svm, you need state database replicas which you create with the following command:
metadb -afc 3 c0t0d0s3 (example here of course. 3 copies in one slice in this example; put a few more on the other drives if you can).
To mirror a file system (let's say /export/home on slice 7):
Do the same for the others. Root, /usr and var can't be unmounted so it will require a reboot. Also look at the man page for metaroot (updates /etc/vfstab and /etc/system for you for the / file system.
Also, with zfs, the syntax looked wrong. It should be zpool create mypool mirror first_disk second_disk (not first_disk mirror second_disk).
*sigh* you unix people and the syntax...
Fist off you're right; the command was wrong (slightly...but still wrong nonetheless)
Why do we geeks feel the need to "show up" other people. We cannot memorize EVERY COMMAND IN EXISTANCE...that's why the man pages are for...it's sad when people are annal like that....
Next time iI'll put a disclaimer saying "I'm not at work right now so there is no way to test my commmands but the theory is all there..."
Also I thought it was common knowledge that the boot drive cannot be a ZFS filesystem...so I assumed that I did not have to mentioned that.
If you want slow performance...use metadbs; also you shouldn't use metadbs on your boot drive because you will "thrash" your boot drive...which is bad...Good luck...I'm unsubscribing to this thread now...
Distribution: Solaris 10, Solaris Express Community Edition
Posts: 547
Rep:
Quote:
*sigh* you unix people and the syntax...
Fist off you're right; the command was wrong (slightly...but still wrong nonetheless)
Why do we geeks feel the need to "show up" other people. We cannot memorize EVERY COMMAND IN EXISTANCE...that's why the man pages are for...it's sad when people are annal like that....
Sorry for commenting this but there's nothing wrong in pointing out wrong syntax: it's something that could help *others* looking LQ for help, rather than pointing out that you have written something wrong, which is human and comprehensible.
If someone studying mathematics looks for help, and you help them saying that what you're saying is not correct but... there's a book for it, why should I look for help? Maybe I'm stuck with something I cannot see.
I did not mean to show up anybody here, I was just trying to help the person that opened this thread.
Also, if I was trying to show up anybody, I would have mentioned that you can also use zfs on 06/06 as well, not only
on 11/06. The only things missing on 06/06 that appeared on 11/06 were raidz2 which is not what we were talking about
or the concept of the spare (also new in 11/06).
So let's chill out, it is about helping people "custangro". LOL.
What do you mean about not putting the the metadbs on the boot drive and trashing the disks? If the user has 4 disks, he should put the replicas on all 4 drives for high availability and redundancy. Even if you have two disks and you want to mirror the boot disk, you put the state databases on both disks. There is a parameter to put on /etc/system that allows
the system to boot up even if you lose one disk (meaning you come back with only 50% of the replicas). You can google that
if you want.
LinuxQuestions.org is looking for people interested in writing
Editorials, Articles, Reviews, and more. If you'd like to contribute
content, let us know.