Sun - Week Fortythree

Monday starts off OK. Coming into work to find some tickets on the queue and some work to be done from last week. My main ticket is setting up an x4100 over in the 02 lab for an engineer who wants to install VM’s. There is a problem with this, he wants to use Windows and whenever someone says that, there are always groans, why? Because Windows == licensing. He wants to use MSDN licenses, however, he isn’t the only user of the systems-to-be, so he needs to use volume licenses. We have an old VM machine which has valid volume licenses on, but I can’t access the VM files so I can’t transfer them…

Tuesday was a terrible day, it started terrible and ended terrible. I woke up late, to Tom asking if I was ready to go at 8:30. After telling him I’d just got up and rushed to get ready and get the next bus. When I got in the ticketing system was down, when it came back up one of my tickets had gone from being a normal priority to a high priority…The ticket was to set up two Intel boxes, an x4150 and an x4250, each with a 10 gigabit Ethernet card. This should have been a really easy ticket. However it turned into a nightmare. Installing the 10Gb card gave me PCI option ROM errors during POST, I wouldn’t usually care, but this meant the BIOS couldn’t see the RAID card and so couldn’t use the disks. After flashing the BIOS and trying the on-board disk controller and still getting no where, it was 6:25pm and I decided enough was enough.

Wednesday was significantly better than Tuesday. I got in and disabled PCI option ROM for the PCI slot the 10Gb card was in, this allowed the BIOS to see the RAId card and the disks, I booted from the disks and used `ifconfig -a plumb` to see all the interfaces and found the 10Gb card! I also managed to install 64-bit Java to a shared directory (on a SPARC host!) and get it to work on an x64 system (simply install on an x64 host, then copy the files over). The afternoon was mostly consumed by a meeting with our new boss…All the labstaff from around the globe jumped on a call which was just as exciting as I’d predicted…

Thursday didn’t start well, I woke up a little later than usual and felt disorientated most of the day, I think it’s a lack of sleep… The morning was spent assigning some IP’s for an engineer and set up a T2000 sc for him. The afternoon was spent troubleshooting a 15K domain. Solaris would ping the local production server, but OBP couldn’t see any network activity…The interface with the cable in told me that I should check for a problem with the cable, and I did. Moving the cable around made this message follow the cable to different interfaces and using a known working cable didn’t make a difference. It’s very weird, it’s like BIOS not seeing your disk, but Windows being able to see and access it fine…In the end I delved into /etc/path_to_inst and copied the device path, under OBP I then did: boot /pci@9c,700000/pci@1/network@0 - install This seems to have worked and the domain is net booting as I type this!

Friday was spent fixing console access for an engineer and finding a PSU for a customer system. We also moved some old racks into the campus data center, they were very kind to let us use some of their space. The old racks contain machines which have been end of service lifed, but may still be called up if a big customer offers to pay the right amount of money.

Are blogs a fad?

Why does one start to write a blog? I know why I started mine, I wanted some representation on the web which was a bit more organised than the sites I tend to cobble together…Did I share much with you all?  Looking back, no.  OK, so I’ve blogged about at least one thing every month…But that’s just it, some months I’ve completely neglected this site…

I’ve been unable to sleep tonight, so thought I’d have a look around at friend’s blogs.  It’s amazing to see how many aren’t updated on a regular basis, that have been left since the end of last year, or longer…There are a couple who surprise me too, a few people I would expect to update their blogs more often than I, they generally seem very interesting people, and I’m surprised that they don’t update more often.

So why is it?  I know I often forget to update everyday after work, and tend to look back at the week in my weekly blogs.  I think, though, when I leave Sun, I’ll probably drop back to a couple of entries a month (which will be a shame).  I think people get busy, I’ll often think that the I’m doing right now is interesting, and it would be good to blog about it, but obviously I’m busy doing it, too busy to stray off to this place…

I’ve been looking for a Folding@Home client for OpenSolaris, the number of blogs (and forum posts) I came across from 2006 is unbelievable…And often the whole blog has been abandoned…I know I abandoned my old blogger blog, Wordpress has a much better feature list in my opinion.

I hope that I’ll be able to make some time to update this place next year, in amongst the rush of uni work, a home life, and other bits and bobs I’m hoping will come together, and I hope you lot keep reading, I enjoy looking at my stats on a weekly (sometimes daily) basis and seeing that you lot have pushed for another all time high…

Sun - Week Fortytwo

Monday morning starts with us meeting a newtern for next year. Alex is down for a weeks experience to get used to working at sun. As a result, I’ve not actually done that much this week.

Monday morning begins with Alex coming in to the office, I start looking at a SunBlade 2000 which belongs to an engineer, the machine won’t boot, nothing out the graphics card, nothing out the serial port…Before I get any further with the machine, I got an email from Jonathan Schwartz. Sun has been bought by Oracle…I blogged about it too, found here. Feeling a little weird, we went off to lunch to come back to a meeting that afternoon. The meeting was the usual Monday afternoon affair, however, we all knew one note on the agenda was Oracle. Paul doesn’t know any more than we do, and confirmed that. After the meeting we tidied the lab as we have two student interviews and some guests coming for lab tours over the next few days.

Tuesday starts with me looking more into this SunBlade 2000, I debate with Paul as to whether we should sacrifice a lab box and just get this machine up and running. He agrees and I set about finding another box, thankfully there is one over in GMP02 lab. Since I was going over there, I took Alex with me to look at a problem, it would appear a PSU is playing up on one of the machines over there. I was sure that the power lead had been pulled out from PSU0, I was correct and plugged it back in…I recovered the SunBlade 2000 and took it over the GMP03 lab. The afternoon for dominated with a call by one of the executives. He wasn’t letting anything out the bag, it was basically a Q&A session, but he kept telling us he couldn’t answer this question, or that he couldn’t comment on rumors…It wasn’t so bad, someone had used the meeting room before us and ordered food from the canteen…We had platefuls of sandwiches and loads of things on sticks and baguette…

Wednesday was spent fixing up that SunBlade 2000, I took various bits from the old box, popped them in the new box and it seemed to work, this, however, took most of the morning…It wasn’t an easy task trying to troubleshoot some of the problems. I spent the afternoon with Alex looking at JLT and jump starting machines.

Thursday is here! It’s felt like a long week, although nothing official has been mentioned about the takeover, it’s still looming. I’ve spent most of the morning with Alex looking more at jumpstart and other UNIX bit and bobs, it’s currently lunch time, and I need to get some coding done before the end of Friday…

Friday came and went frankly…I spent the day doing some coding and surfing of the web, the Dennis forums went down last night :( We’ve set ourselves back up over on x404.co.uk! Also, Aled went to pick up my “new” iMac… I say new, it’s a CRT G3 iMac, 350MHz, 256MB RAM and a whopping 20GB hard drive, all for £15! Pictures of it here courtesy of Aled!

Opensolaris Live USB Creator

Just been alerted to this: clicky

Though if you don’t have access to a Windows box, but do already have access (or know someone with access) to an OpenSolaris box, then this method works too.

Oracle buys Sun Microsystems

Well, the news hit the two company sites at around 12:30 BST by my reckoning.  Sun.com and Oracle.com both reveled the news at around the same time (sun clicky and oracle clicky <- pics of the sites on the day).

I think this is a much better acquisition than what IBM offered, personally, I think IBM were looking to buy up competition, Oracle have a number of reasons to buy Sun as I see it:

  • Expanding the business into new markets means more fingers in more pies, hopefully those pies have got meat in them otherwise you’ve wasted your time.
  • Fishworks, the ability to create appliances.
  • A large base of customers running Oracle on Solaris and Sun kit.
  • Solaris technologies such as ZFS and DTrace The Oracle site has an amount of information for existing customers etc. It can be found here. There is a discussion over at opensolaris.org which echo’s some of my thoughts, found here.

We’ll have to wait and see what happens, but it’s going to be an odd couple of months these last two working for Sun/Oracle.  I’m glad this has happened while I was on an industrial year, I’ve gained the experience of a potential take-over and that of a take-over with little risks attached for me, far more than I can say if it happens in a couple of years time when I potentially have mouths to feed and a mortgage to pay!

My weekly blog will continue, due Friday, I will expand on the other events of today and probably link to this article…

Sun Week Fortyone

This week starts on Tuesday, Monday being Easter Monday and a day dedicated to getting over a “chocolate hangover”, whatever one of those is…

Tuesday started with me trying to figure out how to send an email using Java. But before I sunk my teeth into it, I realised I had promised to change the hardware of the stable zones box for some engineers. The stable zones box is just a box which is set up to create and test Solaris zones on, the current machine (A Sun Blade 2000) was restarting every 24 hours, seemingly this has something to do with Solaris 10 update 5 and above…. So I got an Ultra 80, set up the zones stuff and made it ready to take the place of the current zones box. At 4pm I scheduled some downtime and swapped the boxes, they are heavy things, and powered up the Ultra 80. I had console access, I had changed the details in NIS+, and I just had to change the hostname, which was easy thanks to this page :D At 5pm an engineer decided he really want to play with domains on the stable 15K downstairs, the SC’s use our production password so I can’t give him access. On top of this, the console to the main SC was dead, it wasn’t allowing SSH or telnet connections and wouldn’t failover. I didn’t have time to look at it, so emailed the guy and left it for Wednesday

Wednesday! Feels like a normal Wednesday too, yesterday felt very much like two days… The morning was spent fixing that 15K, the console wasn’t broken, someone had consoled from the SC to an inactive domain, so I had to ~. out of that. After some investigating, it looked like NIS+ wasn’t running on the machine, this being Solaris 9 I’ve got no idea…I’ll have to grab David or James at some point…

Gyaaaargh! It’s only Thursday :( The morning went slow…I didn’t get any development done because an engineer wanted me to move a storage array and kept complaining that it didn’t work. In the end it was his fault for not enabling something or other… I also got a new email from the guy who I’m doing this programming for.

I’ll update this later to say how I got on…
I’ll update this later to say how I got on…
I’ll update this later to say how I got on…
I’ll update this later to say how I got on…
I’ll update this later to say how I got on…
I’ll update this later to say how I got on…
I’ll update this later to say how I got on…
I’ll update this later to say how I got on…
I’ll update this later to say how I got on…
I’ll update this later to say how I got on…
I’ll update this later to say how I got on…
I’ll update this later to say how I got on…
I’ll update this later to say how I got on…
I’ll update this later to say how I got on…
I’ll update this later to say how I got on…
I’ll update this later to say how I got on…
I’ll update this later to say how I got on…
I’ll update this later to say how I got on…
I’ll update this later to say how I got on…
I’ll update this later to say how I got on…
I’ll update this later to say how I got on…
I’ll update this later to say how I got on…
I’ll update this later to say how I got on…
say how I got on…

:wq

*note to self, Wordpress does not except Vim key bindings - FAIL!*

Sun - Week Forty

This week I’m going to try and finish that email script…A while ago I was assigned the task of writing a module for our booking system.

The Problem

Engineers can book hardware to escalation numbers (customer cases), however, we know that a number of engineers take advantage of this, and book systems for other, more personal, uses under the same escalation numbers. These bookings don’t end unless: 1) The engineer or a lab team member removes the booking, or 2) The customer case is closed.

The Solution

Create something that will check the database for bookings which have been active for more than X number of days and bug the engineer via email to see if they still need the system.

What I’m doing

So, after achieving this once in Perl (*spit*) and it being little good (because we want to trigger it from the database), I gave it another bash: This time I’m using a stored procedure (from now on referred to as an SP) which will collect a list of bookings. This will then run a Java application, passing in the booking number as an argument. The Java application will then connect to the database again and call another SP which will recover the booking details (I couldn’t pass them in as arguments as the method of calling external scripts from the database is limited to 255 characters).

This is basically where I am now, at the end of the week. Next week I’ll need to find a way to get the Java application to email the user of the system, and maybe look at a better way of passing in a list of booking numbers, as having loads of these processes running could seriously slow things down…

I will, at some point, blog about connecting to a Sybase database using Java, this should be fairly generic, I just found other documentation for doing this very poor :(

Sun - Week Thirtynine

Monday morning is here, we are an hour ahead of last week with the clocks going forward. It seems really sunny today, shame that my day starts off with unblocking the toilet, yuch. Anyhoo, I get into the office as usual, grab a coffee from the Java cafe and spend most of the morning setting up some of the French kit. I’ve set up console entries in JLT, and all console connections appear to work. I’ve just got to set up IP addresses for the kit and their SP’s, configure all the SP’s to work on this network and boot net the hosts. However I must configure boot net for any hosts that are booked, but not jump start them…Something which is very tempting! The afternoon was spent in a meting, and finding three machines. I want to test some thign out: we have had complaints that from Solaris 10 machines, people can’t access one of our share through /net though Nevada machines seem to be unaffected…I have three v210’s which I have installed Solaris 10 update 4 and 6, and Nevada build 110 on, I have found that I can reproduce the issue, just a case of finding out why!

Tuesday came and went with me looking over the machines again, and looking into some problems as to why an engineer couldn’t get console to work.

Wednesday, 1st April…April fools day… Again, I mainly looked at kit, labeling them up etc today. I also fixed an issue in the states where an engineer couldn’t boot net - install a machine, because they had changed the devalias’ in OBP sigh*

Thursday morning was spent fixing some console connections to those French machines, they are doing my head in! There are still two to fix :( The afternoon was more exciting, to an extent….An engineer wanted to customise a boot image, I had done this on Solaris 10 update 5, but in update 6, the game had changed…. We had to unpack the minitroot, apply the patch, the repackage the minroot…I’ll post the steps later in a separate entry I think…

Friday! End of the week! I spent today catching up on some other bits and bobs, as well as sorting a couple of bookings for engineers.

My blog, here in 1969!

As you can see from the image (below), I have the stats for my blog on Thursday, 1st January 1970!  I also have stats for ‘yesterday’ presumably meaning Wednesday, 31st December 1969!

Obviously those of you that are more observant, anyone can claim this: simply put an invalid date into the address bar!  Though interestingly, 2009-04-31 give you the date: 1st May 2009.

Just something to blog about before bed!

rm -rf tragedy

It happened a couple of weeks ago now.  I plugged in my external hard drive “POCKET” and OpenSolaris mounted it under /media/POCKET.  I was going to copy my films and TV series over to POCKET as I was travelling up to see Faye in Aber. 

I started the copy from my Mac, however halfway through my Mac crashed :(  After a restart I decided to start the copy again.  So, my plan was to cd to POCKET and `rm -rf` everything…However, by some horrible fate, I typed the path to my media folder by default and went ahead with `rm -rf *`.

I only noticed about 30 seconds later what was happening…Bit by bit, my data was being de-referenced, left for ZFS to write over it with something else…My first thought was “It’s OK, I’ve got a mirror!”, however, the mirror is, of course, in real time…My second thought was “ZFS snapshots!”, however, they are not enabled by default :(

I have spent the last few weeks slowly re-ripping my collection (though thanks to James as he had copies of a lot of content), I’ve still got a few to do, but I’m mostly there.  I turned on ZFS snapshots too.  I access my server though SSH from my Mac, so here’s what I did:

ben@faegilath>$ ssh -X beleg-ia -l admin Password: admin@beleg-ia:~$ time-slider-setup

I was then presented with the screen on the left

As you can see, you can just tick the box saying “enable time slider” and the choose to snapshot all filesystems or select what filesystems you do want to snapshot.