Today's Page Hits: 1444
I have more hair and it isn't so grey. :->
This page validates as XHTML 1.0, and will look much better in a browser that supports web standards, but it is accessible to any browser or Internet device. It was created using techniques detailed at glish.com/css/.
With the OpenSolaris pnfs binary drops, things are still pretty raw. (Hey, we could hold onto the code until it is fully baked!) A common problem is telling which machines plays which role - i.e., I've got a pNFS community set up, I've done some testing, and I've rebooted the machines. How do I remember which machine is which?
The dservadm command will help you. If it shows nothing, you have either a client or the MDS:
# dservadm listpools zpools:
Okay, we find the DS:
# dservadm listpools
zpools:
train
# uname -a
SunOS silent 5.11 pnfs-curr i86pc i386 i86pc
Now, how do I tell which of the remaining two are the MDS and client? Well, dservadm can tell me a bit more:
# dservadm listmds
mds:
192.168.2.119.8.1
# host 192.168.2.119
119.2.168.192.in-addr.arpa domain name pointer stealth
Don't worry about the extra numbers at the end of the mds identifier - just realize the first four are octets. Okay, we can now start back up with our testing.
I've given up trying to explain to people what it is I do for a living. But I think I'm going to keep on trying to explain what a BakeAThon is to them.
How can I do one without the other? Well, I can abstract the process.
So a BakeAThon (and a ConnectAThon) is an interoperability testing event. And the best way to describe it is that you have 10 different people with a set of rules for a game (if you say AD&D edition 3 rules, you have another set of problems to deal with). Each of them has read the rules and believes that they know all of the intricacies. But none of them have played the game with anyone else.
So they all get together and start to play each other. And they start to argue about each and every move. Sometimes it is pretty obvious whose interpretation is wrong. And sometimes they call someone else over to help decide.
As soon as player A is done with player B, they start with player C. Except sometimes they are also playing with player D at the same time. Or player B comes back to see if they have gotten rule 5.3 correct now.
Sometimes they all vote on how to interpret a rule and even change the rule book. And sometimes player F was at the bathroom when that happened and causes the debate to start back up again.
Then they all go away again for 3 months, promising to play games against each other remotely. They meet back up again at the next BakeAThon - sometimes there is a new player or someone didn't show up. But they are willing to chime in over email.
But they have to start all over from scratch because no one played remotely and they've been busy playing with themselves.
A further complication is that some people only play defense and some only play offense. Sometimes you get a team where they split those duties. So when you talk to one person about how they run their offense, they shrug and say that they only do defense. And the problem that arises here is when the team's offense only plays against their defense - they get pretty good at it and understand some simple shortcuts that make it easy. But when they play another team, those same shortcuts cause problems.
So a BakeAThon is pretty much like that. The major difference is that the competitiveness isn't in winning a game but in getting the game adopted by other people. I.e., Foo Inc. and Bar Inc. may have differences and fight over customers, but while at a BakeAThon, they work together to make NFS a better protocol.
The best way to rewrite the partition table on a Mac Mini is with the OS X boot DVD. Insert it, reboot, and press C until it boots from the DVD. At that point, get to where you can start Disk Utility, then:
I also partitioned it as type Unix, but I think you can skip this as long as you get the partition table written. I couldn't figure how to eject the Tiger disk at this point. So, reboot!
At this point, your disk is hosed as far as booting. You need to get the DVD out and the best way to do that is to be pressing your mouse button while it boots. (I did the left mouse button on a Logitech mouse.) When the DVD ejects, insert your OpenSolaris DVD. I can't remember if it booted then or I had to reboot. If you need to reboot, remember to press C.
The other thing to note is that I selected the standard installation (and not the developer's installation). I partitioned the drive as follows:
0 / 20480 1 swap 4096 3 /a1 20480 4 /a2 20480 5 /zap 38632 6 ---- ----- 7 /export/home 10240
/a1 and /a2 allow me to do Live Upgrades if I need to. I can also scavenge /export/home if needed (i.e., I'd have a NFS homedir server). These are mainly test machines.
And when it boots and is networked:
partition> p Current partition table (original): Total disk cylinders available: 14590 + 2 (reserved cylinders) Part Tag Flag Cylinders Size Blocks 0 root wm 7054 - 9664 20.00GB (2611/0/0) 41945715 1 swap wu 3 - 525 4.01GB (523/0/0) 8401995 2 backup wm 0 - 14589 111.77GB (14590/0/0) 234388350 3 unassigned wm 4443 - 7053 20.00GB (2611/0/0) 41945715 4 unassigned wm 1832 - 4442 20.00GB (2611/0/0) 41945715 5 unassigned wm 9665 - 14589 37.73GB (4925/0/0) 79120125 6 unassigned wm 0 0 (0/0/0) 0 7 home wm 526 - 1831 10.00GB (1306/0/0) 20980890 8 boot wu 0 - 0 7.84MB (1/0/0) 16065 9 alternates wu 1 - 2 15.69MB (2/0/0) 32130
The IOGear KVM I have is both a lifesaver and the bane of my existence. OpenSolaris does not recognize the Apple wired keyboard under GRUB. (And it also fails to recognize it under Ubuntu.) If I use the KVM, I can use the keyboard. So that is how it is saving my life.
It is slowly killing me because of tech support. The two add-on cable clusters are not working. IOGear says it is because I am using HDMI converters on the Minis. Strangely enough, it does not matter on the embedded cable clusters. I finally found a machine in the house which has a pure VGA connection. Guess what, the cables fail to work there as well.
I'm frustrated because the tech support crew is not listening to me. They are fixated on the HDMI converters and not on the fact that the box is broken. I've just got them to admit that the thing is broken. Now I get to send them the unit and they'll send me something back, sometime.
I think I'll head over to the Apple store to see if they have anything.
I've done enough swapping of cables to be undaunted about having 3 computers and only 1 KB, mouse, and monitor. Anyway, once the systems have OpenSolaris on them, they should run smooth, right?
In Getting together a pNFS storage community I explored putting OpenSolaris on a Mac Mini. Here I'm going to talk about reversing that.
First things, the mini I am working with seems to have a screwy interaction with both my IOGear kvm and my wired Mac keyboard. The other two minis I have work fine with both.
I'm reverting back to see if I can get the machine to work with the kvm in the native OS.
Okay, pop in your installation disk - in my case it is Tiger. The first thing you will notice is that the install can not find a disk to install on to. Start Disk Utility, then:
And then Bob is your uncle! At this point you can change the partition to 1 huge one and select your other options. Then partition it. If you skip the steps to change to a GUID table, then this partitioning will not work. Once done, exit Disk Utility and go about a normal install.
Note, I bet I could have used Disk Utility to have avoided Ubuntu when going the other route.
We decided to get some Mac Minis as test machines for pNFS. We needed small and quiet machines.
The first problem was that the minis do not have a serial port for a console. I bought an IOGear kvm that supported 4 machines - right now I have 3. It fits right in the mini stack. The front view:
And the side view:
I know I can install Nevada 79 (the pNFS codebase is currently ontop of 79) on them - Trebor has done it - but he won't blog about it. So off I go. The first issue is that in boot up, the USB keyboard is not seen, so I have to take the default install of the developer edition. Note that I like a finer grained install which allows you to setup multiple slices. In fact, I will need that for the DS machines - they need ZFS.
Anyway, onward. Note that there are some resources for doing this with bootcamp. I want the whole disk, so I'm trying to treat this like any other x86 box.
The next hurdle is when partitioning it craps out right away with a warning that the '/' slice extends beyond HBA cylinder 1023. The nice thing about the installer is that I do not have to reboot right here. Hmm, trying a 32G root partition does the same thing. I wonder if I'm hitting the bug Paul reports in How to Dual Partition a MacBook Pro with MacOS and Solaris, e.g., CR6413235?
Nope, I bet it is the EFI issue. Okay, I asked Trebor (dark side of Robert Gordon ) if he hit this and the answer was yes. His suggestion was to let the install work until it failed, and then in the resulting terminal do fdisk -E blah, where blah for me was either 'c1d1' or 'c1d1p0'. Both gave the same output for me.
I still selected a 40G partition because I want to have a ZFS partition on the rest of the disk. And now the installation is just proceeding. And it dies trying to install the boot blocks. This was after installing everything.
I'm trying Ubuntu right now to get a different perspective. It also had an issue with the keyboard at first. Note that I can reconfigure the nevada ISO to boot grub up into the regular install if needed.
Anyway, the Ubuntu install is chugging away. It is done and appears to start to reboot. Very slow compared to OS X. And it is networked!
Now I want to see what happens when I install Nevada (snv 79) on top of Ubuntu. The first difference is that I do not get that annoying '/' 1023 cylinder message. Umm! And that works!
Okay, I want to avoid the booting Ubuntu step in the future. I'm pretty sure the problems I am seeing are with the disk label. I guess I really need to pay attention to steps 6-9 of Alan Perry's Setting up a Mac Mini for dual booting Solaris and MacOS X - note that neither Alan or Paul had problems with their keyboards. Paul wouldn't on a laptop. And Alan may have been doing this before the Developer installation was an option.
To get networking going, you need a Marvell Yuokon driver, which you can get from this entry: Solaris 10 U3 on Gateway MX6453 - be sure to follow the advice to get the 64 bit version. And note, it downloaded as skgesol_x64v8.19.1.3.tar.Z.tar for me, so rename it to get rid of the trailing '.tar'.
Read that blog entry again, I needed to add to my '/etc/driver_aliases':
skge "pci11ab,4362"
Okay, I'll continue this later...