Forgot your password?
typodupeerror

Comment: Re:Unfamiliar (Score 1) 366

by tinker_taylor (#47889849) Attached to: The State of ZFS On Linux

Cons:
- cpu and ram overhead (even by current standards, uses a tonne of resources)

You can tune the size of the ZFS arc cache, thereby optimizing RAM use. In the Solaris world, upto 25% of RAM is used by ZFS by default, unless we throttle it. If you have SAN LUNs as the underlying block storage for ZFS, it is better to reduce the arc cache size. I suspect same thing is possible in Linux port as well.

- doesn't like hardware raid (apparently a lot of the pros rely on talkign to an actual disk)

It is not recommended to use hardware raid, but that's because ZFS has superior FT mechanisms (RAIDZ2/Z3 etc). And if you use a JBOD, you can leverage things like L2 ARC (using flash devices), SSD based ZFS write-ahead-logs (what's called the Logzilla in the ZFS appliance world)

- expandability sucks (can be done, but weird rules based on pool sizes and such) compared to most raid levels where you can easily toss a new disk in there and expand.

This is incorrect. Expanding ZFS pools is as simple as adding additional devices to the pool. Depending on your underlying striping strategy, you would have to add storage in commensurate manner of course. It's literally -- "zpool add "

Comment: Re:above, below, and at the same level. ZFS is eve (Score 1) 366

by tinker_taylor (#47889749) Attached to: The State of ZFS On Linux

The RAID, LVM, Filesystem approach is defunct in the modern world. Also, ZFS already incorporates multi-protocol support, ability to turn any host with local storage into a target (via the COMSTAR framework). Not sure how much of this is in the linux port, but I suspect that if it's close enough to Illumos, it should have these features.

ZFS is not in the microsoft tradition, it is a departure from 20th century storage design/architecture. The very idea that there has to be a RAID/LVM/FS is archaic and has been thoroughly disproven. In my previous shop we had petabytes of storage in ZFS pools and hardly ever lost data.
The Pool-based model that eliminates the layers of RAiD/LVM/FS results in better performance, easier supportability and superior diagnostics capabilities.

Do you realize that almost every major storage vendor first bashed ZFS and then about 3-4 years later started building architecture that was eerily like ZFS?

My shop was one of the early adopters of ZFS since back 2007. There were a few bugs then, but over the years I have been absolutely impressed with the efficiency and stability of ZFS.

Comment: Re:An F- for the handling of Solaris (Score 1) 223

by tinker_taylor (#45955251) Attached to: James Gosling Grades Oracle's Handling of Sun's Tech

[[[Only major shortcoming with pkg manager in solaris' 11 was with the fact that there was no directly accessible repository of software that could be used (albeit the dependency checks etc were missing too).]]]
This should read "only major shortcoming with pkg mgr in solaris less than or equal to 10"

Comment: Re:An F- for the handling of Solaris (Score 1) 223

by tinker_taylor (#45953113) Attached to: James Gosling Grades Oracle's Handling of Sun's Tech

I think there is some sort of censorship going on (thereby forcing me to get my slashdot password reset) so i can repost my comments, this time not as an AC.

- Package manager was brain dead. apt, yum are far better. ( Sorry Solaris 11 was too late. Too much legacy out there. )

Only major shortcoming with pkg manager in solaris' 11 was with the fact that there was no directly accessible repository of software that could be used (albeit the dependency checks etc were missing too). That has been fixed in Solaris 11. You have to remember that Solaris 10 was released in 2004 (10 years ago). If you want to use Solaris today, you should switch to an opensolaris variant or Solaris 11.1

- Patching made no sense. You have no idea what packages are patched with a patch. Patches were just binary disk vomit that spewed crud all over the system. Impossible in the real world to build any sort of verification around them. ( Sorry Solaris 11 was too late. Too much legacy out there. )

It's an easy enough thing to fix. Use zones to convert physical servers into solaris 10 branded zones. All of a sudden you patching becomes non-sequiter and generally painless. In my shop we automate patching of 100s of solaris (yes 10) boxes at a time and using ZFS root + Iive upgrade we patch them all under 30 minutes. With Solaris 11, this is down to a 2 minute window. We literally patch Solaris 11 LDOMs in under 2 minutes flat, including reboot. And Solaris 11 has gone away from Patch IDs etc...they are all pkgs, just the same as Linux.

- Zones: Are a nightmare of security and privilege. I don't care what any says a zone is just a change root jail. Which means you will only every be as up-to-date as the host system. And it means you must be compatible and tested against the host system. Which is really no different than not having zones. Zones are a horrible horrible mess.

Zones are a very very cool technology. You can do a lot of powerful things with zones at zero additional cost. And with the advent of Solaris 11, you can now P2V physical hosts into Solaris 10 branded zones, and circumvent the whole patch/package issue. I have used zones with great success (100s of them running multi-TB DWH on M-class, T-class and x86 hardware). They are not a nightmare of security and privilege. The rules are very very clear and concise. There are things you can do within zones and things you cannot. It calls for a little old thing called RTFM, that's all.

- No dependable only repository of packages that is robust or up to date. Far to much package hunting still required to locate software for solaris. Most packages are
months to years behind there linux counterparts.

Not anymore. Solaris 11 SRUs are released at regular intervals and upgrading is as easy saying 1, 2, 3 ;-)

- Java performs better on x64 than Solaris/SPARC. This has boggled me for years. Only recent sparc architectures let java and other highly threaded applications stacks really perform well. Why do I even have to know about processor binding for processes?

Try running you java apps within zones with dedicated vcpu and memory grants. And Solaris runs on x64 as well. Consider running java on the T4s and T5s and see how they run.

Comment: The Model for IT is wrong (Score 1) 277

by tinker_taylor (#44445841) Attached to: Ask Slashdot: Is Tech Talent More Important Than Skill?

The whole concept of x AM - y PM is wrong when it comes to IT. IT calls for flexible work schedules (unless one is doing pure development work) and IT professionals should be given the flexibility to choose deliverable deadlines and left to their own devises, accountable for their promised work. I have found that in organizations that focus on end results as opposed to strict work schedules, IT is far more successful, because such environments tend to nurture innovation and talent. Another thing is to encourage lateral thinking (out of the box) and problem solving skills. It is amazing how quickly even an "average" skill level can turn into high-level, when provided with the opportunities and the freedom to think.

Most enterprise-level organizations don't allow that of their employees (at least most of those that are IT user as opposed to IT creators). The employees then, with the lack of freedom then become drones who regurgitate mindless, automaton-like work year on end.

IT is an intellectual profession and IT professionals should be encouraged and nurtured to be thinkers, philosophers even. They should be given an environment where they can question every decision, know every reason why a decision was made and how their work affects the bottom-line for their organizations' business. That's when a narrow focused, unimaginative IT professional can start developing and exhibiting real talent and develop real skill.

Comment: The future doesn't look that multipolar (Score 1) 119

by tinker_taylor (#37571500) Attached to: Oracle: Proud, Self-Reliant, Increasingly Isolated

With the state of affairs (financially) with players such as HP, the future doesn't look very multipolar. I suspect that the Enterprise computing market will become more and more bipolar (pun intended) and the focus will shift to selling "platform as a service" or "Infrastructure as a service" solutions which hook seamlessly into public cloud offerings...
HP will get bought, EMC will get bought, Netapp will get bought and then there will remain only 4 main players in the Computing platform market -- IBM, Oracle (Private Clouds) and Google and Amazon (Public clouds)!
Everyone's personal computing will be powered by an i or a Droid. Offices will become completely virtual, UNIX will (REALLY) be dead (this time), Systems admins will either have to work for the big four or start flipping burgers. All code will be automatically generated by some code vending virtual kiosks...

Hey wasn't there another thread about Operating Systems for Cities?!? :o

Comment: Define Newspaper...moot points (Score 1) 412

by tinker_taylor (#31776782) Attached to: Rupert Murdoch Hates Google, Loves the iPad

Based on Murdoch's track-record, it seems like none of what he would define as "News outlet" would actually qualify as being bona-fide Journalistic entities. Take Fox News for that matter...they don't report news, they manufacture it. I am certain all of Murdoch's "News outlets" are manufacturers of News. I guess that does qualify them as "News outlets" sans the Journalism.

So it is not very clear why he is mad at Search engines for distributing. Isn't the fact that one is able to distribute one's propaganda for free a dream-come-true for a Propagandist? What else does he expect now? Google should pay him $$$ for the garbage that his "News Outlets" produce?!? :o

Comment: Re:That's fine (Score 1) 392

by tinker_taylor (#31691154) Attached to: Solaris No Longer Free As In Beer

Maybe this is not such a bad thing afterall. The Linux kernel seemed to do well without any corporate tinkering. Why can't OpenSolaris simply carry on into whichever direction things take it, independent of Oracle.

Projects like Crossbow, NPIV etc are integral parts of OpenSolaris, stuff that is missing from stock Solaris afaik, so what is to prevent the community from building on this solid base and reaching new heights?

Comment: Database programming needs Math (Score 1) 609

by tinker_taylor (#31617236) Attached to: Math Skills For Programmers — Necessary Or Not?

One major component of mathematics is Set Theory and database programming uses Set Theory for unions, joins, etc. So yeah, Math knowledge is needed. On the other hand, mundane programming doesn't really need too much Math background, especially of the formulaic/formula-driven part of Math. What is needed is a sound understanding of Boolean algebra (which is what Digital Electronics --> Building blocks of Electronic computing needs), and thereby Logic. Do you need too much math background to do bubble-sorting, etc? Perhaps not.

Applications that are calculation intensive (statistical programs, scientific applications etc) do need Math background, but not necessarily beyond basic University level (provided there are statisticians, scientists who provide the underlying Math to the developer).

Comment: Re:Absurd! (Score 1) 238

by tinker_taylor (#31601316) Attached to: Oracle/Sun Enforces Pay-For-Security-Updates Plan

Would that then qualify Security patches as being "enhancements" or would they still be "bug-fixes" (irrespective of whether the bugs were hitherto known or unknown)?

Since we live in a Global marketplace and all major players are playing at the global level, the playing field is automatically leveled! Companies have to follow and adhere by a common set of rules (moral & ethical first, eventually legal)...which would remove the gray area vis-a-vis ethics and their subjectivity.

There are far better avenues of revenue-generation/augmentation than nickel&diming the user community for security patches.

Comment: Absurd! (Score 2, Insightful) 238

by tinker_taylor (#31587122) Attached to: Oracle/Sun Enforces Pay-For-Security-Updates Plan

This is the most absurd piece of news I've come across this year! Why on earth should I pay to have Oracle/Sun fix their own bugs?
Obviously Security flaws are bugs. If any security vulnerabilities are identified, they should be ethically and morally obligated (ie assuming that the legal angle is unenforceable) to fix these and distribute the patches for free.

Isn't there anything called accountability/responsibility left any more?!? We are a huge Sun shop and one of the reasons we loved Sun so much is the fact that it was not a blood-sucker when it came to things like patches, software, etc. Unlike a company like HP, who charged for everything from multipathing software to UNIX resource mgt tools (which should be defacto standard of any mature OS).

Comment: Re:Don't be fooled (Score 1) 403

by tinker_taylor (#31045862) Attached to: India Ditches UN Climate Change Group

Indeed. There is in fact an entire industry revolving filtration of water in India, using various technologies such as reverse osmosis filtration and UV radiation, made affordable to most of India's urban and well-to-do rural sections.

There are indigenous and traditional techniques being used today (as sort of a revival movement) in Rainwater harvesting etc to offset the effects of Global Warming. In India's case specifically global warming is a modern phenomenon. The British rule of India resulted in massive restructuring of India's native infrastructure systems. One such case is the destruction of a very elaborate and intricate irrigation network (using canals and small dams) to introduce the Railway system. A direct result of this was the death of millions in the Famines of Bengal in the 19th century (a holocaust-proportion extermination due to "Globalization") that no one talks about these days.

What does this have to do with Global Warming? It is simply an example to demonstrate that "Industrial" activity has been very detrimental to climate since early 19th Century (won't mass-scale Famines qualify) and is not really that new a phenomenon. What's new is that there is awareness among the victims of Global warming, which naturally draws the line between them and the culprits (and their representatives).

COBOL is for morons. -- E.W. Dijkstra

Working...