Comment In the land of the blind... (Score 1) 238
Or maybe that should be blindfolded.
Or maybe that should be blindfolded.
OpenVZ is not the same as the rest. It takes the host O/S and creates skinny containers from it that appear as individual systems; but that's not the same as "virtualization" in this context. You can't run an alternate O/S using OpenVZ.
What you actually said was:
Has anybody here on Slashdot had any experience with one or more of these clustered file systems?
The rest of my comments, WRT mdadm etc, we're not related to clustering, or sharing direct attached raid devices between systems - because as others have also said, I don't think that a cluster best suits your requirements. Unless of cause you just want to do it for the fun of it, in which case go for it.
Or alternatively you back everything off to tape (rotating sets) and store them in a fireproof safe.
How is this an answer to your question? You identified 3 cluster filesystem types that protect against hardware loss by distributing the data over a cluster of systems - but OCFS2 isn't like that. It's a filesystem that's designed to provide concurrent shared access to a filesystem by a cluster of servers, which in combination with a HA framework can provide a platform that applications can use to protect against node failure, not disk failure. With OCFS2 you still have to make the storage highly available with a RAID solution plus manage concurrent connectivity via a SAN, iSCSI, etc. So unfortunately this is not an answer to your question at all. The filesystem types you've identified would do what you want, but they're also expensive for a home solution because you have to throw more computers at the problem to increase redundancy and performance.
Have you considered using software RAID (mdadm) on Linux instead of a hardware RAID controller? It has a useful feature that allows you to grow existing raid volumes by adding more disks. Maybe combine that with a small UPS to allow your system to shutdown gracefully in the event of a power failure. Alternatively, if you want to stick with a hardware solution have you taken a look at Drobo? I have no personal experience with Drobo, but from what I've read their proprietary RAID solution allows you to grow your array by just popping new disks in or increase capacity by replacing existing disks with larger ones on the fly. They have a couple of different models that can scale to 16TB. Best of luck with your search.
Not one x86_64 laptop is certified on that list.
Last update for those unixutils tools was 8 years ago. I hope there are no nasty security bugs lurking in there or you've just compromised your entire estate.
Last time I checked my eyeballs didn't work very well with objects. They handle text perfectly well though. Seriously, you're missing the point. The command line is for running commands and eyeballing text output. Bash functions perfectly for this. It's also perfect for writing small quick scripts that do simple tasks. If I need to do something ore complicated that requires an OO approach, then I have python or ruby to play with. Use the right tool for the job.
That's because creationist Christianity (usually evangelical) is a cult. It's not recognised as such by most because it's become so main stream; but any religious sect that preaches absolutes (e.g. the bible is inerrant) and is blind to any other way of thinking is a cult.
I'm not familiar with how Ubuntu package and push their software. Do you get everything inc the kitchen sink with each crank of the wheel whether you want it or not? Red Hat, in between major/minor releases bundle up new RPMs into errata. An errata may contain one or more RPM packages that are tested/required to be installed together, and they're delineated into three types, prefixed with:
RHBA - Red Hat Bugfix Advisory
RHSA - Red Hat Security Advisory
RHEA - Red Hat Enhancement Advisory
In our production environment, we avoid anything that's an enhancement (too risky) and consider each bugfix and security errata to see if it's application to our builds, and whether we want to install them or not, and whether the fixes would warrant a new QA effort (time consuming and costly). We're not forced to, nor are we forced to take each point release in turn if we don't want either. That's the way it should be done.
And oh so bloody convenient. I'm with hairyfeet on this one, though we're never going to know as they quickly dumped the corpse at sea. I've long believed that he copped it pretty early on in the campaign. But announcing his death at that point would have muted the ire of public opinion, and that didn't suit the political aspirations of the war mongering neo loony right in US government at that time. Much better to have his ghostly presence live on as a target for public hatred. Now, with the coffers empty and US dollars needed at home to try and right the economy, and with continuing loss of US soldier lives, it's the right time to announce his death. Something the government had to do in order to claim victory. But there are no pictures; a body quickly and conveniently disposed of in such a manner that it can never be recovered, and no evidence that can be ratified by an independent third party. This whole scenario is a work of fiction, now that the US government have decided it's the right time to kill him off.
The reason may have an element of FUD to it, but the consequences aren't. People don't buy computers to just sit there running an O/S, and the majority of HP-UX Itanium builds I used to be involved with were commissioned to run Oracle. So this is going to hit HP Itanium sales hard and HP Itanium customers even harder.
Oracle seem to be getting more ruthless in recent times. They aren't playing very nice with Red Hat either as they've killed off support for ASMLib and OCFS2 on RHEL 6. Clearly an attempt to force customers wanting Oracle RAC clusters on Linux off RHEL onto Oracle's own Red Hat clone. Small wonder that Red Hat have reciprocated by making it harder for Oracle to get a free ride on the work Red Hat put into the kernel.
Personally I think it's pushing it to classify RIM's BlackBerry as a smart phone. I have both an iPhone 4 and a company Blackberry given to me for work. The Blackberry is a clunky unwieldy toy compared to the iPhone with a fraction of the capability. I use it as little as I can possibly get away with.
I still think a better solution is to make it impossible to write malicious software in the first place
If it was that easy it would have been done already.
Maybe an in-between solution is for Google to vet apps that request more sensitive permissions.
And how do you determine if an app is going to request sensitive permission without umm, vetting it in the first place? Chicken and egg situation there mate.
Perhaps another step is to make it so that by default the app asks for the more sensitive permissions but the user has to confirm them individually
So you have a situation where the app is constantly asking the user for confirmation before doing things, kind of like how MS Vista used to do. We all know how well that was received.
I guess that Apple obviously thought this through properly before they released their product. Maybe Google should eat some humble pie and just emulate what Apple have done.
I heard once that London Irish rugby club used to get theirs shipped direct from Dublin because it was better than the stuff the UK distributors had.
It is a little extra special when it's fresh. If you're ever in Dublin visit the Guinness Museum at the Guinness brewery. At the end of the tour make sure you visit the Museum bar and sample a few jars. It's a real treat
The only possible interpretation of any research whatever in the `social sciences' is: some do, some don't. -- Ernest Rutherford