The Grumpy IT Admin

Here's the thing, if I order something physical like a computer, router or network switch the supplier has to send that to me in a box. When I order something less tangible like software or a license to use something or even swap money for some virtual points (Microsoft Live points etc.) I don't actually need to have anything more than a license key or at most a PDF that says I am ok to use that item and somewhere to get the software from. You may be saying yes that sounds about right to me, however, as astounding as it may seem in this day and age, many things are still delivered as physical when a simple e-mail with a license key would suffice. I recently purchased a couple of XBOX live gold memberships with 800 MS points (don't get me started on that rip off) from Play.com and had to wait 6 days for the four 25 character keys to be delivered by Royal Mail. Why? Just e-mail them to me, save some carbon and I get fulfilment in 2 minutes and everyone including the frogs

Just been over to Anandtech where Anand has a review of the latest Sandforce controllers in a beta of OCXs Vertex 3 drive. http://www.anandtech.com/show/4159/ocz-vertex-3-pro-preview-the-first-sf2500-ssd EDIT: That was in fact the Enterprise version, here's the skinny on the consumer SKU and what the differences are. http://www.anandtech.com/show/4186/ocz-vertex-3-preview-the-first-client-focused-sf2200 He clears up the issue of SSD life span very well. To sum up the original spec for 50nm Nand was 10,000 erases, 35nm is generally around 5,000 and the new 25nm is around 3,000, this doesn't sound like a lot of lifespan but once you understand how these things work yu will appreciate it is quite a long time. SSDs only wear when individual cells are wiped so basically the controller tries do everything it can not to do that, however the wipe cycle takes time so there is a balancing act between wiping early to avoid slow downs and leaving it to the last possible minute to a

Did you know that if you snapshot a server in VMware the snap files are kept with the server? So what you say. Well if you have some tiered storage and your nice super-duper SQL box has system drive on some slow SATA storage and your Logs and Data are nice fast SAS or FC storage when you snap all the writes will go to the slow storage and your SQL box will run like a complete dog, not a greyhound but three legged pug with a bad hip. You can if you so wish specify where VMware stores it’s snap files but you will always see performance problems as it switches between locations to provide the latest data. VMware snapshots should be seen as a very short term thing, i.e. snapshot, do update. does it look ok, remove snapshot or revert, anything more than a couple of hours and you are starting to go in to dodgy performance territory particularly on SQL boxes and I would guess nything else running a database like Exchange or Oracle. One useful tool to have in your armoury is RVTools w

Some more Backup exck quickies for you. If you do a full shut down on your tape libraries, the kind where you pull out the power leads, don't forget to re-enable hardware compression, this is turned off be default and full shut down resets its to the default, crazy but that's what happens when developers make descisions. Upgraded from 2010 to 2010 R2, all went well but one duplicate job errored with a mount failure. Couldn't work out what was what, decided to recreate the jobs and it failed with unavailabe media, had to go into the policy and deselect the reselect all the media and drives and then all was well agian. Weird.

Well the SATA channel card replacement went well. The Unisys engineer had some experience from last time and after we checked which card we need to replace about 20 times and read the swap out instructions 10 times which tell you to just pull the failed card, we spotted a little advisory to ring support before just pulling the card. Now our SATA card hadn't actually failed we just had one path that was a bit dicky and this was on the primary card so just pulling it out is not really a good idea. There's no little fail light to show which card needs to be swapped and there is no way to flash a locator light on any parts in an EqualLogic like there is with HP kit which is major failing in my book. I guess you get what you pay for. There is a couple of commands you need to enter from a terminal session plugged in to the live controller, one to shut down the card and one to bring it on line, this went smooth as silk this time, unlike last time when we enabled it and 6 drives fa

Not a big surprise when a big company roles out a product and it does something weird and all the current software that is supposed to work doesn't. In this case it is HPs LTO 5 tape library, this the the FC connected version. A large beast with 4 drives and 96 tape slots. Getting it all hooked up was quite straight forward, which was fortunate as there are no instructions at all anywhere. Backup Exec sees it and you can load up the drivers etc. and all is well, you just can't do anything with the library as you can't open the doors to put tapes in or anything useful like use it. The reason is our own fault as we have a CASO server and second backup server all hooked up through Fibre Channel to the libraries.  Every other library we have had works fine with this set up and the MPIO software from Microsoft. But this new library won't. For some strange reason when you tell the library to Unlock it doesn't. Why, because we have two servers talking to it and it

After a month or so of peace, the truce between EqualLogic and I has been broken. Yes the errant drive 20 has once again gone bonkers dropping in out and finally failing. If you missed the last times marathin fix go here to see it. Basically our EQ was well broken with 6 failed disks and was only saved by the propeller heads in EQ Engineering over in the US after about 9 hours and 15 hours after I arrived at work. EQ phoned home and I waited for the call. This came 12 hours later which is not bad considering we have 24/7 cover with a 4 hour fix. I believe the 4 hour fix actually refers to the time it takes to get the part to you after they have decided what is wrong with the unit, which can take as long as it takes, maybe I should read the small print sometime. Once again the diagnostics were emailed off to Ireland, a new support guy this time, I am working my way through them and rate of knots. The diagnosis is that the SATA card needs to be replaced, this was the one that go

For reasons to do with DR we needed to change our IP addressing scheme, either for our servers or for our PCs and printers and what not. Obviously changing PCs and printers was a whole lot easier due to the possible issues of inept developers’ hard coding IP addresses causing no end of problems with re-addressing the servers. This has entailed changing all the DNS entries for our printers, (whatever happened to the paperless office?) That in itself is no real challenge, they all have web interfaces so a quick connect and amend the setting then next printer. Updating the DNS entries was going to be a bit tiresome so I did a quick Google on how to do this with a script and up popped DNSCMD. You can run this from a CMD box on your PC, the main options you need to know are /recorddelete and /recordadd. To delete a record you simple need to type dnscmd <ServerName> /recorddelete <ZoneName> <NodeName> <RRType> <RRData>[/f] What does all that mean? Well i