* Posts by Nate Amsden

1504 posts • joined 19 Jun 2007

IO, IO, it's profiling we do: Nimble architect talks flash storage tests

Nate Amsden
Silver badge

3par has been saying this for close to 15 years

And built their architecture around it

Picture describing this aspect

http://www.techopsguys.com/wp-content/uploads/2013/05/3par-mixed-workload.png

0
0

Scale-out sister? Unreliable disks are better for your storage

Nate Amsden
Silver badge

3par does this too

Working with SSD vendors to reduce space allocated for bad cells and give it to the array (20% I think). Since 3par operates with chunklets if some flash goes bad it just stops using it. Too many chunklets go bad and the drive is proactively failed. HP calls this "adaptive sparing".

My org's oldest 3par flash is about 22 months old and as far as media life the SSDs have 98% of their life left (if you were to source these SSDs direct from SanDisk they would be read intensive SSDs though HP doesn't limit their workload under the 5 year warranty).

Google could do what XIO did just partner with the HDD makers. Or go buy XIO(and stick to seagate drives last I read XIOs tech was specific to that). XIO can't be too much to buy.

0
0

Cisco preps the P45s for 500 unlucky UK staffers

Nate Amsden
Silver badge

Not that I entirely disagree but it seems kind of funny so many of the "hot big and sometimes new" companies are the ones that constantly lose money, meanwhile the old boring ones that keep laying people off often make tons of money(and have even more in the bank). Sure some of their businesses may be in decline to some extent but if I were a betting person I would think that in many cases these old big companies will run out of cash long after the newer companies that seem to live paycheck to paycheck.

0
0

Video service Binge On 'broke the internet' but 99pc of users love it

Nate Amsden
Silver badge

best of both worlds

People get to watch more video. Those that hate the concept can opt out.

As an ATT subscriber with a 5G data plan who streams maybe 10 minutes of video a month across both wired(200meg comcast with 1TB) and wireless it doesn't impact me either way but sounds like they had a good idea(i think i said that before here)

1
1

Samsung intros super-speedy consumer SSDs, 'fastest M.2s ever'

Nate Amsden
Silver badge

Re: Viable as SAN replacement?

SANs are all about high availability and data services. A lot of which is very sophisticated software and hardware architecture.

You can certainly have a very fast disk array with this kind of storage but it wouldn't really be in the same space as a traditional SAN product. And no the open source storage stacks don't come close to being adequate either unless you are very committed and have significant internal resources for support. Support will of course matter most when shit hits the fan.

0
0
Nate Amsden
Silver badge

Re: Laptops with dual M.2 slots?

My lenovo p50 has dual 512GB 950 pro(pci e 4x 10gig) and a 1TB 850 pro sata. With this config and 3 year on site support about 3500.

I expect to be using it for at least 5 years(previous laptop was a toshiba and was primary computer for 5 years gets light usage still).

I see no difference in performance between 850 pro and 950 pro outside of benchmark scores. Everything is plenty fast.

My normal operating system is linux mint though have windows 7 dual boot in the rare chance i want it for gaming.

1
0

Oracle's cloud strategy is simple – woo and win the latecomers

Nate Amsden
Silver badge

IaaS sucks

but Oracle probably has a decent chance to do some level of SaaS with all of the apps that they have, maybe some PaaS with something like weblogic and oracle db etc..

More value in the higher level services anyway.

IaaS sucks which is why almost everyone is failing at it or those that are not still do a shitty job like amazon (I assume Azure too but can't speak with personal experience).

0
0

Seventeen hopefuls fight for the NVMe Fabric array crown

Nate Amsden
Silver badge

Re: NVMe right, but block fabrics less relevant in the new stack

And those legacy type systems will be around for a very long time yet - even today such systems are still being built as new. The last time I spoke with DBAs about MySQL compression for example the basic suggestion was don't use it. The last time we attempted to use it in a test environment it blew up badly.

You may like new funky DBs that have weird names and are optimized for special edge cases, but the bulk of the world will continue working off of things like Oracle, MySQL, MS SQL and other "legacy" platforms with apps that won't support things like you are talking about for the next decade or more (in many cases they will get along fine without them).

Been working closely with devs on "new" web apps over the past 13 years(across several companies all sort of in SaaS in one way or another) and they still have problems grasping basic things like "single points of failure", or "take the whole app/site down while we make a small schema change". Or let's have microservices, where in many cases there is so many inter dependencies that you are just increasing your points of failure rather than reducing them. Or how about "HEY there are two cpus in this server let's run with more threads to better utilize them, but no the code is too highly serialized and doesn't scale like that..".

I saw one recent app update to a 3rd party component deliver literally an 82% drop in performance(worst case). Can't get the CPU usage above 30% (10 CPUs), vs before could get to 93%. Vendor who makes the software doesn't seem too bothered by the results, or at least not enough to even prioritize a simple investigation into the cause when it's so easy to reproduce.

I remember another time, at another company, in house app - new version comes out - 66% drop in performance due to new serialization in the code. They squeezed an extra 15-20% out of it again(eventually) but admitted that was as fast as it was going to get.

The more I do stuff(and talk with others that do it) the more I realize this is more the norm rather than the exception.

3
0
Nate Amsden
Silver badge

as mr wonderful says

Take violin out behind the barn and shoot it

1
0

Half! a! billion! Yahoo! email! accounts! raided! by! 'state! hackers!'

Nate Amsden
Silver badge

why would people sue

This is an email account, not like they swiped credit cards or social security numbers or something like that(I would expect Yahoo would not need that information for signing up for an account anyway).

(been hosting my own email for roughly 20 years now)

0
9

Intel XPoint over-selling criticism surges as Chipzilla hits back

Nate Amsden
Silver badge

kind of funny

or sad to see some folks here on el reg (in the forums) touting how great Xpoint will be every so often yet expectations keep getting lowered.

For me I'm really not paying attention until the product starts shipping and people see what the real thing is. Until then what's the point of getting excited ? SSDs and NVMe flash are already so damn fast, and durable.

My own org has been using a 3PAR 7450 for nearly 24 months now with what Sandisk would call "Read intensive" SSDs (though HP has no limits as to workloads for their 5 year warranties). After nearly 24 months of usage across hundreds of VMs and MySQL servers the oldest SSDs in the array have 98% of their life left in them according to their internal metrics anyway.

There may be a few extreme edge cases where something like Xpoint will really make a big difference but customers already have a lot of choices for very fast storage. I didn't get the impression that Xpoint will be dramatically cheaper either, which really would be more of something to get excited about I think.

I have a pair of Samsung 950 Pros (NVMe) in my Lenovo P50 laptop(which the Samsung windows tool says has PCIe 4x with 10Gig interface w/ i7-6820HQ CPU), I see absolutely ZERO difference in real world performance between those and the Samsung 850 Pro (2.5" SATA) SSD. Feels kind of crazy to have 3 SSDs in one laptop but I figured what the hell it has the space so I put them in.

I would wager if I replaced all of my SAS SSDs in my 3PAR arrays with NVMe I would see no difference either, because well the individual apps just don't drive that much I/O (much of it is handled in memory caching). Sub millisecond response time is more than good enough for any of the VMs or databases I have worked with over the past 15 years.

5
1

Going hyperconverged? Don't forget to burst into the cloud

Nate Amsden
Silver badge

infrastructure has little to do with bursting to cloud

It's all about the apps and if they can support it.

e.g. it is a stupid idea to scale an application that uses a database by adding web servers in a public cloud and having them connect to a DB over a WAN connection back at the data center.

and most apps (even ones being developed right now) don't support fancy bursting, developers have a hard enough time figuring out multi threading and removing single points of failure from application design. Bursting to cloud will remain a pipe dream for the vast majority for some time to come.

(spoken as someone who has worked closely with developers working on web-based apps for the past 13 years at several companies of various state/scale)

1
0

Electrical box fault blamed for GS2 data centre outage

Nate Amsden
Silver badge

Re: I don't understand

haven't seen a UPS blow up myself but I killed one of my UPSs about a month ago, replaced the batteries, had them wired up wrong. UPS committed suicide I guess to save my equipment. Breaker tripped and started making clicking noises and smelled some smoke. Some internal parts connecting the batteries were melted. This particular UPS had 9 batteries in it, rack mount double conversion sine wave UPS, not a tiny little thing with a single battery. Frustrated that I was too lazy to double check the voltage before I hooked it back up again. I checked it the first time I replaced the batteries, then 2 of the batteries were bad and replaced them, I was over confident I wired it up right and that cost me about $700.

I suppose I did sort of have a UPS blow up about 12 years ago though now that I think more, power failure and one of the pieces of equipment connected to it did not like the voltage the UPS gave off in battery mode so it reacted badly and the UPS sort of blew up, chassis was warped, batteries were leaking, but besides that no damage to anything.

Never seen a big UPS blow though, nor have I heard any stories about such an event. Only been to one data center in past 16 years that suffered power failure on both redundant circuits. Since that time I haven't taken power for granted. I have not had a single feed go down in about 7 years at any facility(which for me has been 4 different facilities in that time).

1
0

CloudFlare offers web encryption up the wazoo

Nate Amsden
Silver badge

I am happy

to browse el reg on regular ol HTTP. Too many paranoid people out there that want to encrypt everything, just a waste of resources for the vast vast majority of things.

Some folks would like to think by using SSL they are safer from things like government snooping.. I don't buy that for a second myself.

I am not fond of my ISP (Comcast today, was a local provider before I moved to new city a couple of months ago) knowing a lot of where I go so I tunnel that to a server I have co-located in a data center for a decent chunk of my traffic at least, and I run my own recursive DNS. But once it gets to the data center if it is not encrypted I don't care anymore, my ISP there(Hurricane Electric) doesn't have a vested interest in trying to "monetize" that kind of data.

I suppose more than anything though I am just riding on top of stuff I setup close to 10 years ago, it doesn't seem to be broken (other than theoretical bandwidth limits are much less due to tunnel limits) so have no real reason to fix it.

In the past tunneling to my co-location server actually sped things up I would assume since HE's peering is better. I have about 21ms ping times to my server, even though the routing runs about 800 miles(very inefficient route but I guess that is how the cables are laid, a very direct route in theory would be about 75 miles). The latency is about the same from my previous home where I was about 25 miles from my server.

0
4

Is Tesla telling us the truth over autopilot spat?

Nate Amsden
Silver badge

i think most people

Would prefer to have them develop the technology in private(like google or others) or at least slap a big BETA sign on their cars. So people that do want to play beta have a chance to, but most of the rest that don't understand will be better informed to look elsewhere.

1
1

Samsung's latest blowup

Nate Amsden
Silver badge

I regularly microwave my Galaxy Note 3s

on cold mornings, never had a problem getting them warmed up for the day.

0
0

Google: There are three certainties in life – death, taxes and IPv6

Nate Amsden
Silver badge

Re: Bridging the gap

Meanwhile millions continue to happily use NAT in either ipv4 or ipv6(most likely NAT to ipv4 networks).

Looking at my ATT branded note 3 it is using a 10.x ip from them (wifi is off). So carrier grade NAT there too, never had an issue(I tether to laptop constantly when I travel ).

1
0

Cloud pause slows server revenues amid price war rumours

Nate Amsden
Silver badge

so the story is

Public cloud is a lot of hype with it only accounting for 8% of server revenue?

0
0

Teenage noughties protocol BitTorrent reinvents itself again

Nate Amsden
Silver badge

why would the network obey a law

That was never written for it. And of course obviously it is not a law just an observation that the number of transistors would double every so often.

Seems that bit torrent is a solution trying to find a problem that people will pay to have their software use.

(Have never used nor needed bit torrent myself)

2
0

Gmail suffers worldwide wobbly Wednesday

Nate Amsden
Silver badge

Re: This is why the "cloud" is a terrible idea

A big issue for me and cloud isn't that shit fails. Is that they are constantly fucking with it. When you have control over things you generally schedule things like upgrades and code changes so at least key folks know it is happening and have the option of providing input as to changing the schedule if needed.

I haven't seen a data center power outage on two redundant circuits in 9 years (and that was a poorly managed facility that had several outages). I have seen one ISP failure in a data center in I want to say the past 15 years. (Shit ISP I think the outage was less than 30min). Obviously this happens more often in the world I just speak to data centers I have been a part of.

Doing maintenance on power or layer 1-3 networking seems to be a well understood thing. Maintenance on cloud has orders of magnitude more complexity and more prone to failure as a result.

Which is why we have things like maintenance windows and notifications. Doesn't work as well when the provider is "agile" and is constantly fucking with it.

When cloud can offer a 100% uptime SLA like Internap does for it's network services with guarantees on latency and performance(and a track record to prove it) we'll be in a good spot. But I am not holding my breath.

2
0
Nate Amsden
Silver badge

Re: This is why the "cloud" is a terrible idea

Someone should remind evernote.

4
0

Did you know iOS 10, macOS Sierra has a problem with crappy VPNs? You do now

Nate Amsden
Silver badge

WEEKS

that's plenty of time to prepare, really. I don't use apple and have never used PPTP but seems weak of them to do.

Sort of like firefox saying "oh this is not secure https so I won't let you connect no matter you know what you are doing or not".

But from what I've read at least apple allows a somewhat easy rollback to earlier OS? Such an option doesn't seem to exist for android (maybe it does after rooting or something). I've used android 5 on one pf my note 3s for 4 months now and am happy to stick to 4.4 on my main phone. I keep wifi off so ATT can't sneak in an upgrade when I'm not looking.

1
4

Aruba OS8 lands, with APIs so non-NetAdmins can do NetAdmin jobs

Nate Amsden
Silver badge

up vote from me

I've worked with maybe 150 to 200 devs over past 16 years(mostly in SaaS). Need 1 hand to count those that knew networking to any real degree.

1
0

Gimps with big Dell balls hijack HPE's partner shindig

Nate Amsden
Silver badge

Re: Analytics?

You should be happy to know that HP no longer does analytics having sold their software group which includes atonomy and vertica.

1
0

JDK 9 release delayed another four months

Nate Amsden
Silver badge

Re: Is Oracle really serious about Java?

I was just going to say how can anyone think oracle is not serious about java when so much of their own software runs on it.

Yeah i read the stories earlier this year ahout them trying to be more proprietary and then changing course but regardless unless they are re writing their apps in another language java isn't going anywhere.

0
0

VMware Workstation's installer loads danger-junk, so patch it ASAP

Nate Amsden
Silver badge

workstation is mature

Been using vmware since 0.9 on linux back in 99. Vmware workstation has been mature enough for me for a decade or more. Only reason i upgrade (last one was from 10 to 12 black friday saw the deal here on el reg) is for newer OS support (going from 10 to 12 did nothing for me but it seemed cheap at the time and i believe i still qualified for upgrade pricing at the time. I think one more major version and that wouldn't of been the case). With Linux kernels breaking drivers so frequently not like i can keep running workstation v4 forever.

I suppose this is what vmware gets for axing the workstation team and I assume outsourcing it or something ? There was news on that here in the past year or so.

I also suppose some people will confuse these security issues with vsphere hypervisor when they should not of course(they will just see the name vmware).

5
1

Hypervisor security ero-Xen: How guest VMs can hijack host servers

Nate Amsden
Silver badge

Re: not sure about enterprise but

let me know when one is published for vsphere.. I have looked in the past and the only thing I have found has been related to vmware tools running on windows I think specifically for HGFS in particular (and that may of been workstation only since there's no point of HGFS on esxi)

Some folks like to say, you never know what is undisclosed for security in vmware, same can be said for Xen or any other product as well.

Having the source by no means solves anywhere close to all problems. There's been some nasty bugs released in open source stuff that took well over a decade to find.

I'm not complaining either way(when I see the Xen stuff like this I find it funny), but am a happy vmware customer(enterprise+ 5.5, and none of their fancy automation stuff which seems to be breaking a lot)

0
3

IBM lifts lid, unleashes Linux-based x86 killer on unsuspecting world

Nate Amsden
Silver badge

HP beat em already?

IBM is comparing to HP DL360 when Apollo is the big data solution from HP. I'm not in this space but a quick search came up with

"The HPE ProLiant XL270d Gen9 Accelerator Tray provides up to 56 Tflops of single precision performance per server with eight NVIDIA® Tesla M40, and up to 15 Tflops of double precision performance with the NVIDIA® Tesla K80 GPUs and two Intel® Xeon® E5-2600 v4 processors in a 2U server."

https://www.hpe.com/h20195/v2/GetDocument.aspx?docname=c05069270&doctype=quickspecs&doclang=EN_US&searchquery=&cc=us&lc=en

(technically it is a 4U enclosure that looks to house 2 x 2U servers)

I am not sure if they can fit 8 K80s in a 2U package (vs 4 in IBM?) or if it is 8 M40s and a lesser number of K80s. Nvidia says K80 does up to 2.9Tflops of double precision so I would assume they have roughly 8 K80s in a 2U package.

Though HP is using PCIe I think and not NVlink.

1
0

Apple killed OS X today and binned its $10,000 BlingWatch too

Nate Amsden
Silver badge

why would a new version of android help tablet sales?

I'd expect the vast majority of the market to not give a shit what version of Android is on their tablet as long as it can run whatever apps or games they want it to run and it has the hardware they need. Unless of course there is a significant change in interface which causes users to revolt like windows 8.x etc.

I bought a toshiba android tablet what seems like almost 2 years ago now, it was a pretty good deal at the time, nvidia tegra something, bunch of ram, bunch of storage with SD card.. though it's sat on a shelf for 95% of it's life I don't know what to use it for. Anything that needs mobile gets done on one of my two Galaxy Note 3s (the 2nd one is not currently on any phone plan and has no personal data on it so in the event I need to install something like a chat app that wants insane permissions I do it on that device there is nothing to steal there.

I was going to try to use the tablet as an interface to one of my Tivos via tivo stream now that there is an Android app out that is compatible (my tivo stream has otherwise sat in it's box for the past 3-4 years). Though after 30 minutes of fighting it I gave up, the app crashed constantly and when it didn't crash itself the traffic it generated somehow caused my shitty wireless AP to crash and reboot as well. So the tablet went back on the shelf. Generally my wifi serves only light web browsing(or very occasional youtube etc), apparently it can't handle whatever tivo is dishing out.

My HP touchpads get more use than my android tablet (the touchpads sit in their "touchstone" charging stands and act soley as digital picture frames).

I'm sure some people have uses for tablets, if I had a really small phone then perhaps a tablet would be good but the 5.7" screen is good enough for me to even watch full length movies while traveling (helps to have ~140GB of usable storage on the device as well). Back when I had a HP Pre3 with it's what was it 3.5" screen with low resolution (relative to my 1080p phone now) the tablets got much more usage as web browsing devices.

3
0

My Dell merger wish list

Nate Amsden
Silver badge

as a former dell customer

(Sonicwall customer, not sure if that division is spun out from Dell or not yet), I have no wish list for Dell or EMC myself. I do like Sonicwall though it has worked well for me over the past 5 years. Our usage of it is mainly for site to site VPN but it's been very reliable, fast, very easy to manage, and trouble free for the most part. Their SSL VPN solution wasn't flexible enough to do what I needed it to do, so never got to deploy that.

My wishlist is fulfilled by HP. (Servers+storage+FC storage networking anyway(HP also provides us with front line vmware support), I haven't bought into any of the converged stuff yet, maybe will get blades some day, I had a fully populated Dell blade chassis 6 years ago at another company for vmware (free ESXi only the cheap bastards), it worked ok, though the lack of integration with the management between the chassis and the iDRACs on the blades was very surprising. I'm sure Dell has improved on it since then.

As someone else mentioned I do hope Vmware gets their quality back up again, while the problem hasn't impacted any of the systems I run(still run vsphere ent+ 5.5 and vcenter on windows, deploying a new vcenter soon going with windows again there too, even though I am a linux guy). At some point I may upgrade to newer things from vmware and it would be nice to continue the track record I have had with them with good success across ESX/ESXi, vmware workstation(on linux), and before that VMware server, and GSX before it was renamed to vmware server.

Maybe will be an EMC Isilon customer soon too (one of their smallest boxes, need something highly available and high level of support for maybe 2-4TB of critical data, and HP doesn't have a good enough NAS in that space), been using ZFS to-date(a couple different solutions) and it does the job but is a very hacky solution. Basically set it up and don't touch it afterwards. Tried HP StoreEasy which is HP's answer in this space but it sucked really really badly, even had the failure of a single volume (due to internal file system problem with dedupe on the windows storage server) take out the entire cluster instead of just taking down one volume. Turned off clustering after that.

I don't use HP for ethernet networking either (Extreme networks is my switching company of choice), though if I do go the HP blade route at some point I'm sure there will be some HP networking in the chassis.

0
0

Linus Torvalds won't apply 'sh*t-for-brains stupid patch'

Nate Amsden
Silver badge

Re: He's right. Again.

"You must have a short memory. What generally happened was that the unstable branch got dragged out too long and distros/maintainers would then try to backport required changes to the stable kernel resulting in TWO unstable branches. "

It may be short, perhaps people just tend to remember bad experiences more than good ones. I have fond memories of the 2.2.x kernel, for some reason 2.2.19 sticks out as a kernel I ran for a looooong time, and 2.0.36 as well. I remember some special security patch I would apply to my kernels back then as well(forgot the name of it).

But what sticks out more is the memories from about a decade ago(or more) of having to hack together kickstart disks with newer hardware (that CAME WITH DRIVER DISKS FOR REDHAT), but those drivers were not compatible with whatever kernel breed was in the kickstart kernel (the kernels were almost identical, not talking 2.2 to 2.4). Spend so many hours doing that, especially for Intel (or was it Broadcom, or both? I do specifically recall e1000e driver as problematic at the time) network cards (these were HP DL3x0 G3, G4, G5 and G6 servers). Also for one or more SATA controllers. So, it sort of came down to extracting the kickstart disks to find what kernel they are using, find the source for it, and the configuration. Build the kernel so I know it builds, and then compile the drivers against that kernel, re-insert the drivers back into the kickstart data files and try to boot the box(we booted over PXE at the time), and hope it works. I think it was the only time in my life I had to work with CPIO was with those driver modules/disks/files.

That stuck out so much that two years ago when I deployed the first bare metal servers in my data center in 6 years I really feared facing that situation again. Almost all my physical servers run Vmware and the drivers there have been solid as a rock for me at least for the past decade.

I much prefer back ported stuff myself. Most recently on my brand new laptop (Lenovo P50) I installed Linux Mint 17, ran for about a month or so and it was working great. Then I went to travel and that was the first time I tried to use wifi. The kernel with mint originally is 3.13 (or at least that is what my other Laptop with mint has on it right now, in any case a 3.x kernel).

It did not support the intel wifi chip in my laptop. OK so I go around hunting for a driver, totally ready and willing to compile the driver for my kernel. I come across Intel's open source website with their drivers that specifically says something like kernel 4.2 required. WTF ?

ok so I go hunting for a 4.2 kernel, and find that FORTUNATELY at this point the Mint people have included "unofficial" 4.2 kernel as an optional package in their repos (yay). So I install that and wireless starts working (along with the SD card slot which didn't work before). After about a day or so the system freezes and perhaps the caps lock light is flashing (kernel panic). I reboot, and it freezes again maybe a week or so later (not happy that I am on a 3 month trip and this starts happening).

Fortunately even more again Mint folks have a 4.4 kernel in their package repos as well and that resolved the issue, whatever it was that was causing the panics or lockups with 4.2.

Though now (I think even with 4.2) I ran into a problem where the system would just go nuts, so I put in a cron that runs every minute that runs

echo 0 > /sys/kernel/mm/transparent_hugepage/khugepaged/defrag

echo never > /sys/kernel/mm/transparent_hugepage/defrag

echo never > /sys/kernel/mm/transparent_hugepage/enabled

I forgot the background on the issue but that has since resolved it. Just running it once after bootup did not seem to be enough.

I have run across driver issues in linux for a long time of course, but this was my first experience where the driver wouldn't actually work (or so was advertised) unless you were on (what I would consider) a very bleeding edge kernel. This coming from Intel which is a really big company. I would not entirely expect a driver developed on 4.x to work on say 2.6 (though it would be nice), it should of worked on 3.x at the very least if compiled from source.

Now that things seem to be working again hopefully i don't have to touch the kernel for another 3-5 years.

4
3
Nate Amsden
Silver badge

Re: He's right. Again.

As a linux user since 1996(desktop and server), I can say drivers on linux have been almost nothing but hack jobs the whole time. Linux driver ABI is always changing and breaking shit.

I gave up on hoping they would stablize to some degree about 10 or 11 years ago but is still sad to see. Just have to look at how many dozens if not hundreds of kernel combinations for example that vmware tools distributes drivers to see how bad it is. Looking myself I see two hundred and ninety three different kernel drivers. (And yes of course there is source code too).

I think the issue has hurt android as well making it more difficult for manufacturers to upgrade drivers for newer kernels.

Linus won't fix it. I got over it a long time ago. But kind of funny to see him complain about that kind of thing when drivers on linux have been broken forever.

The linux kernel devs like to say just release the source. Yeah like that magically solves all problems.

I stopped paying close attention to the kernel when they abandoned the "stable" and "testing" branches, what was it the 2.4 days ? Before that it was say 2.0.x for stable, and 2.1.x for testing/dev, then 2.2 was stable, etc..

bring on the downvotes, totally expected that from the fan boys.

43
36

EMC/Dell deal respectfully caps the minicomputer age

Nate Amsden
Silver badge

why does it matter

Where they are located? I'm glad HP did not merge with EMC for sure, but I don't see why anyone at HP would give a shit that EMC wants/wanted to stay in Mass, and how that could possibly be a deal killer.

2
0

Delta computer outage costs $100m

Nate Amsden
Silver badge

$100M

Would probably fund a very good disaster recovery project

0
0

Extra Bacon? Yes please, even though the Cisco bug of this name is bad for you

Nate Amsden
Silver badge

reboot doesn't mean patched?

At my org they(IT) had to reboot their ASAs about once a month otherwise a memory leak would cause them to stop passing some kinds of traffic and remote management would break needing someone on site to physically cut power to the unit. They never got budget to get redundant pairs too. Maybe the issue had another resolution but I try to stay out of corp IT's business.

They moved off the last ASA maybe a year ago.

At my end of the org we never use Cisco so didn't have that issue. Haven't had the misfortune of managing a Cisco device directly since 2008.

Other than here the last time I saw ASAs was 2 jobs ago and there they had a similar issue but wasn't a memory leak (and they did have pairs). The dumb network engineer set tcp state timeouts to something like 1 week and the state table would fill up maybe once every 1 to 3 months (faster after I started at company ) . They couldn't figure it out for over a year even with cisco support(their workaround was reboot both units remotely via PDU at the same time. They didn't understand why both were affected until I explained to them the state table is replicated between units). Took me about 1 minute. Told network engineer to change the timeout(he feared people would complain about dropped tcp sessions ) and started monitoring state table size via SNMP. State table size never got above maybe 10 or 20k entries for the next roughly 18 months I was there for (the unit max was 500k entries if I recall right). I also don't recall anyone complaining about the firewall dropping their idle connections.

1
0

Brave idea: Ex Mozilla man punts Bitcoin adblocking browser

Nate Amsden
Silver badge

count me in

as someone who would gladly pay a subscription fee to 'el reg. I don't use a "ad blocker" (I consider an ad blocker something to be fairly automatic at detection). But I do use a plugin that allows me to right click on an object and remove it from the page/site/domain (works most of the time not all the time). I also control cookies that come in as well.

I do use an ad blocker on mobile though in part for performance also in part because I have no way to control cookies in mobile.

I don't use the internet for much(major non work sites I don't need two full hands to count how many sites I regularly visit), but el reg is a place I do spend a lot of time on(and have been for what seems like at least 15 years now).

1
0

Mangstor, Mellanox flash rig crowned 'fastest in the lab'... for RAID-0

Nate Amsden
Silver badge

too bad

They don't seem to give any comparison points other than

"The NX6325 topped out at 38,673 TPS, while the previous record holder for a modern flash shared-storage solution came in at "only" 17,316 TPS."

So difficult to know just how fast it is. They give no indication(that I can see I didn't read word for word just browsed through the paragraphs of the site) of IOPS, or throughput actually achieved(they quote the specs of the system though).

Also seems the memory cache they used was way too low, 24GB ? 800GB DB and if you really care about performance memory is cheap I would bump it to at least 128GB. I imagine they set it low to stress the storage more, though I imagine not difficult to run a couple of tests with different configurations.

at my org 99% of OLTP MySQL DBs activity comes out of the buffer cache, main use for storage is for writing log files.

I remember our first DBA got hired and we built out first data center, I hooked a MySQL VM up to our small 3PAR array and he ran his benchmarks (sysbench I think ?? I don't recall). He was blown away by how fast everything was. I was blown away by how little disk I/O was actually being generated(maybe it was in the sub 200 IOPS range or something tiny). He was used to running very large databases(hundreds of them) at a very very large company, so I assumed he knew what he was doing (I sure as hell don't know how that benchmark works). In any case everything was fine when we went live that was all that mattered at the end of the day.

0
0

Enterprise storage is a stagnant – and slightly smelly – pond

Nate Amsden
Silver badge

servers sold with storage

I assume that means servers sold with internal storage on their RAID cards or whatever and not say buying a dozen HP servers with another HP 3PAR storage array on the same quote or something ?

If the former, then I wonder if there is any cut off as to the amount of storage required to be considered for being classified as that (e.g. I would not count a server with local disks just for boot drive with intentions of connecting to external storage to be a server "with storage", not everyone boots their servers from their SAN(I do).

And hell if I'm going to trust some piece of shit $10 USB stick(or SD card) to run a hypervisor on an otherwise $25-30k+(hardware+software) piece of equipment.

0
0

Is VMware starting to mean 'legacy'? Down and out in Las Vegas

Nate Amsden
Silver badge

why use vmware

The author of the article said why use vmware when you can just use containers. My specific use case across about 1,000 VMs is probably 700 of which are single points of failure (mostly in test environments), if one VM goes down in an environment that has 5 or 6 VMs quite possible the entire environment is mostly unusable until the VM returns(which in my case really only happens in the rare event a host fails and HA kicks in and moves the VM in seconds). The overhead of having everything redundant in a test environment is overkill (at least when you are dealing with so many environments).

There are other places where single points of failure are built into the app(s), and even in a world of high stability it is sometimes needed to take underlying hosts offline to do maintenance on them. So being able to live migrate the systems off a host is important. Last I heard containers lack live migration.

Not that my org doesn't use containers, they are used in production (LXC containers that look more like VMs than they do containers as they run many system level services like postfix, splunk, ssh and each of their own IP). Looking to significantly expand this use case very soon. I wrote more in depth on this last year in one of my last blog posts (http://www.techopsguys.com/2015/09/04/containers-my-experiences-good-and-bad/).

In an ideal world perhaps everything is automated and everything is redundant, and everything is shiny.

I'd bet 98% of the organizations in the world don't live in that world and won't for a long time yet.

2
0

VMware fixes 'split brain' caused by 'stubbed toe' of botched NSX update

Nate Amsden
Silver badge

Re: Sorry but

Yes i don't see going past vsphere 5.5 until it goes end of support. I did the same with 4.1.

Maybe by then the major bugs will be fixed.

I file maybe 1 vmware support ticket per YEAR on average(through HP).

1
0
Nate Amsden
Silver badge

Re: Sorry but

I've been a vmware customer for almost 17 years now (1999), never had a vmware patch or product kill my data centers, only have had a handful of PSODs (hardware failures) in the past 10 years of vSphere usage.

Though I have stayed away from the fancier datacenter automation toolsets(like NSX), as I don't need them.

I run only about 1000 VMs though.

2
0
Nate Amsden
Silver badge

the NSX glitch not a sign of troubles in engineering

but the adoption of agile much more a of a red flag. break early, break often is bad for something you want to be very stable.

But at least they've said they are on 4 month release cycles, so customers who care more about stability can stay 2-3 releases behind the curve for more stability.

2
0

Holy friggin' Dell! $67bn EMC mega-gobble to complete on Sept 7

Nate Amsden
Silver badge

so,

what does DVMT stand for ? Dell something something technologies..

0
0

Redmond reveals Hyper-V 2016 beats vSphere's RAM and CPU count

Nate Amsden
Silver badge

wonder how they tested more than 12TB

I see Dell has a 12TB system, HP has a 6TB system (expected their 8 socket box to go higher but it tops out at 4TB). Only x86 system I could think of that may go higher is SGI UV (though officially apparently only supports a couple variants of linux).

All of my vSphere systems are 384GB, not too excited about going far beyond that still right now. Though I'm not running something like SAP HANA either.

0
0

DCIG mid-market array guide: Why we left those companies out

Nate Amsden
Silver badge

HPE was in the report

From what I saw in the previous el reg article(not interested to register to read the original report from DCIG), 3PAR is the go-to market system for HPE, and there are no current 3PAR systems that scale to less than 500TB as of about late September 2015(announced earlier but not shipping till later in September I believe).

1
0

The TPC-C/SPC-1 storage benchmarks are screwed. You know what we need?

Nate Amsden
Silver badge

Re: Nothing beats real applications - BUT

Oh, don't even get me started on load testing :)

In my 13 or so years working at various SaaS-type companies I have never, EVER seen a valid load test against an application(across many companies). Every test has always fallen far short of real production load.

My favorite test was at an advertising company I was at, the front end app was basically pixel tracking, so very light weight, and no external dependencies beyond the single server that the app lived on. The "performance" testing we did there was basically shift enough production load onto a single server until it could no longer keep up, and the load balancer automatically routed requests to other systems when the system under test went down. At the peak we managed to get about 3,200 HTTP transactions a second from a server(disk I/O bound), I assume by now(that was many years ago) they probably use SSDs or something to help with that.

Worked really well, but obviously very dangerous to do in a more sophisticated environment with shared components like databases or caching servers.

My first SaaS job over a decade ago it was semi routine to have to double production application capacity after a major release (and of course we didn't plan for it, so we would order HP servers and have them shipped overnight). At some point the need to double stopped but it was there for quite a while.

0
0
Nate Amsden
Silver badge

As a 3PAR customer for the past 10 years I say I do want another synthetic test, the more data the better, especially if it has full disclosure requirements like the SPC seems to.

POCs are very time consuming to do, many vendors won't do them at all (NetApp refusing a POC 10 years ago is why I became a 3PAR customer, and NetApp again said 5 years ago it would be very difficult do do a POC(not that we could do one anyway but I was curious what the situation was with them and POCs at that time)). I'm sure it's difficult for other vendors too(though the startups generally don't have an issue with POC since they want customers so badly).

If you got your list down to 1 or 2 systems to test then maybe it is realistic to try to do real workloads. But I go back again to my 3PAR experience with 3PAR saying several customers wound up feeling they were "forced" to buy the system because it was in production(and working well). For me when I did my first 3PAR evaluation (small E200) I made sure it wasn't even in the same data center as our production stuff, I didn't want to get stuck in that situation.

When the company I am at now did our initial data center specs(5 years ago), we literally had zero equipment. No racks, no servers, no storage. Nothing to do a POC on(everything was "cloud" based at that time). So impossible to do a POC (in any reasonable time frame anyway).

Synthetic tests are far from perfect, even though some people claim "SPC-1 is "broken"" there's no other alternatives at this point.

If there's a dozen systems or more to choose from POCs from everyone is just crazy.

Fortunately for me my choices these days are pretty simple, 3PAR is good enough, fast enough, cost effective enough, and most importantly mature enough for everything I need (well except File storage, looking to do a small real Isilon POC soon, after testing their Isilon SD edge and having it implode in my tests due to the software architecture they have(their hardware works around some of that with metadata acceleration not available in the software product, something EMC staff seemed to know nothing about when they pitched SD Edge as being basically identical to what runs on bare metal). Wish HP had good file services...and data compression for SSDs(sort of feel like I should give up asking what the status of that is at this point after 4+ years of asking for that feature).

1
0

Google's brand new OS could replace Android

Nate Amsden
Silver badge

Re: Open Sourced or Binary Blob?

I guess by your statement you have no idea what kind of latency hit you'd take trying to share CPU cycles over a network especially for something that needs real time like VR.

9
2

Thailand waters down alien-tracking plan

Nate Amsden
Silver badge

Re: Registration in most countries already

It is the case in thailand, I just got back from a 3 month trip there a couple of weeks ago, had to hand over passport for SIM card.

Even though I signed up for the high end international roaming package with my regular phone calls were still 35 cents a minute.

0
0

They kept that quiet: Toshiba unsheathes FlashBlade rival

Nate Amsden
Silver badge

hardware only?

Article makes it sound like this isn't really a flashblade competitor(not complete at least) if there is no software to go along with the hardware.

Perhaps they are hoping some of the software defined storage players to pick this up and turn it into something more usable.

0
0

Forums