906 posts • joined 19 Jun 2007
the real question perhaps
is when will vvol-like tech reach the other hypervisors..
is it really enterprise grade
if it's also shipped as a laptop drive? Or perhaps another way of asking is flash still so unreliable that you need "enterprise grade" in a laptop drive to make it reliable.
that vmware continues to try(and fail) to diversify away from it's core competency. If vmware actually spent over a billion on buying this mobile device management company that was a big waste, much better use of that money should go to making their various cloud things (even) better (thus far their higher end management things do not interest me anyway).
(vmware customer for 15 years)
Re: 100M records? 75M photos
big oracle DB, one company I was at a long time ago their largest OLTP oracle instance at the time I left (I'm sure it grew a bunch after that) was about 60TB. Oracle told them at the time it was by far the largest single OLTP in the world(next biggest was apparently Amazon at well under 10TB for a single instance). It consumed that much space due to bad application design not because they were doing trillions of transactions.
Re: rarely update my home rooters
ssh is exposed that is it, my previous openbsd install I think was limited to keys only my current one is not will change that now :)
rarely update my home rooters
yeah I called em rooters, sounds funny.
My current home internet gateway is a Soekris box running OpenBSD 5.4 i386. Before the CF card failed roughly 222 days ago (current uptime) it was running I want to say OpenBSD 4.8 (maybe 4.4?) for however many years that was out for(never applied any updates). System is pretty locked down running pf (the only reason I use OpenBSD otherwise I do not like it at all). I login to the openbsd box maybe 3 times a year seems like. Haven't had to adjust the firewall rules in years.
I have a very old netgear WRT54G or whatever they are called running a version of DD-WRT (???) from about 6 years ago. It works, don't feel a need to change it (never crashed, never caused a problem). It's locked down as well as it can be I think. It's on a segmented network that routes through the Openbsd box before it can get to my main LAN or internet. I don't even login to the UI to this thing more than 2-3 times a year.
99% of my traffic runs on wired ethernet.
I have another wifi access point a netgear powerline extender thing-a-ma-bob. it's directly connected to my main network and has the max encryption it supports (forgot which offhand) and mac filtering (far from perfect but anything helps) enabled (WRT54g has both as well).
my laptop and work desktop for that matter run ubuntu 10.04 LTS (end of life was a year ago I believe). At some point I will buy a SSD for my laptop and install Linux mint (whichever one feels closest to Gnome 2 I forget which off hand). Don't plan to upgrade my work desktop unless something fails.
Never once to my knowledge anyway has any of the systems I personally manage either at home or work been successfully hacked in the 18 years or so I've been doing this, so I feel pretty confident that the stuff I do is adequate.
nice to see
another voice of reason. Too many people think Netflix does no wrong.
(myself I killed my account when they had the first price hike because I realized how little I was actually using it at the time - have not seen any reason to re-subscribe to it or any other streaming media service)
firefox ESR updated too
after that ugly as sin UI change in a recent firefox I converted all my systems to ESR. I heard a couple folks say the new firefox UI looks more like chrome (I've never used chrome). I suppose if I wanted Chrome's UI I would use that.
I remember a LONG time ago I was at a company that used BEA Weblogic, and we had a lot of problems with JMS. One of our top engineers engaged BEA and asked them if they had any outstanding issues with that subsystem, they said no.
About 3 months later we had about 36 hours of downtime, due to JMS in weblogic. We "worked around it" by deleting all of the data in the JMS queues(last resort). After about 2-3 weeks of investigation the teams determined not only was it a known bug, but BEA had a fix for it all along and did not tell us even when we explicitly asked. It was their policy at the time not to disclose bugs to customers unless the customer was experiencing that specific problem.
The customer of ours that was down for 36 hours was one of the largest telcos in the world also a big customer of BEAs directly. BEA changed their support policies very quickly(at least for us) after that.
It was an interesting experience to be in my then boss' office with about a dozen senior software engineers and architects and all of them shrugging "I have no idea why it is broken, it shouldn't be doing this".
I talk to people since and they say "oh my god we've had a 2 hour outage! the horror!" pfft. Wake me up when you've been up for 24 hours straight and still don't know what the problem is.
It was easy for me to drop comcast
Not that I was not a happy customer, I would be a customer again if they were in my new area.
"I need to cancel my service because I am moving and you don't have service in my new area"
"Oh, are you sure you want to cancel what is your new address?"
(tell them my new address)
"Oh we don't serve that area, sorry to see you go!"
Took all of about 2 minutes.
My new cable provider to me is about the same as Comcast service(overall probably less service but those services that they lack are ones I don't use), performance and cost wise, though it is a small fry by comparison only serving the city I am in (apparently has ~42k people).
My bill is about $170/mo with most premium/HD channels and standard broadband (~16Mbps down / ~2Mbps up). It serves me well I just wish I had better upload speeds(I have a server at a colo with unlimited bandwidth and TBs of space about 15ms away). Even their top tier 100mbps service has only 5Mbps upload.
I'd be happy to pay Comcast $110/mo for their business class service with 50/10. There are no plans with 10Mbps on my cable provider for any price(on the website anyway).
AT&T Uverse keeps pestering me at least twice a month every month to sign up, but my Tivos don't work with them, so that is a big drawback, they don't offer anything special over my local cable company anyway (internet speeds may actually be worse with IPTV I am not certain).
I was a Comcast (and before that AT&T Broadband) customer for about 10 years or so, never really had a problem. I did stream netflix for about 6 months or so before I lost interest in their lack of content and haven't been a customer since their first price hike (mainly because I realized I wasn't using the service). Never used any of the "on demand" services from comcast, nor voice, nothing but tv(via tivo) and internet.
90% market share
and from the looks of it not much to show for it. Yeah I know they just license the designs but still, seems puny. Samsung could probably buy the whole company with the spare change in the couch in the office of their CEO
good data is valuable
I'd rather have no data than bad data. Some people think it's better to have SOMETHING (even if it's wrong) than nothing, not me.
who cares about power?
With the cost of these systems(both hardware and software) the cost to power them has got to be a rounding error by comparison in most data centers anyway.
The consolidation factor alone should pay for power costs probably 10x over.
vmware dedupes memory
Though it's not very effective for linux VMs in my experience, windows VMs seem to get massive deduplication benefits
11:30:19am up 21 days 18:09, 486 worlds; MEM overcommit avg: 0.42, 0.42, 0.42
PMEM /MB: 393179 total: 800 cos, 3164 vmk, 198343 other, 190871 free
VMKMEM/MB: 389783 managed: 23387 minfree, 123791 rsvd, 265992 ursvd, high state
COSMEM/MB: 91 free: 1521 swap_t, 1521 swap_f: 0.00 r/s, 0.00 w/s
NUMA /MB: 96512 (60234), 97916 (68875), 97916 (56153), 97916 (21283)
PSHARE/MB: 5145 shared, 821 common: 4324 saving
SWAP /MB: 0 curr, 0 rclmtgt: 0.00 r/s, 0.00 w/s
ZIP /MB: 0 zipped, 0 saved
MEMCTL/MB: 0 curr, 0 target, 345823 max
I think that equates to about a 3% savings for transparent page sharing? (ESX 4.1U3 host)
I really can't think of anyone who would consider Hyper-V if they weren't already primarily a Microsoft shop.
Really? 0-60 in 4.4 seconds? Seems kinda slow.
Get it under 3.5 seconds and maybe you can call it a super car. Get it under 3 and you've really got a super car.
People who drive super cars don't give a shit about gas mileage etc. Make it go fast.
vmotion doesn't move storage now
vmotion doesn't move any data on shared storage systems now. vvols won't change that.
all depends on access requirements
sending data to an external cloud provider may look cheap, but if it means you can only push and pull data at a few megabytes/second because of the latency to the remote provider a lot of folks will likely keep stuff on site for performance, maybe encrypt+ship stuff to external cloud provider for stuff with really low access rates.
An example I give people is the main data center for my org is in Atlanta with a 1 gigabit uplink to a tier 1 ISP. Transfer rates to S3's east region for a single stream connection taps out at about 5 megabytes/second(current test is tapped out at 3MB/sec). It seems S3's providers are filtering way upstream as my traceroute to them dies after a few hops and less than 1 millisecond. I want to say before this filtering it was about 15-20ms away.
almost the biggest
HP of course has a 1.9TB 2.5" SSD on 3PAR (technically 1.6TB from the factory I suppose but 1.9TB usable by the customer due to adaptive sparing before any data reduction functions).
Let me do the math here..
6 x 20TB = 120TB raw
Assuming these new X-bricks are the same physical size as the previous units your looking at a full rack for 120TB of raw flash.
6:1 data reduction on 120TB means around 720TB "effective"
Maybe I missed a digit somewhere because I'm not seeing a petabyte here. (their data sheet tries to claim petabyte by saying if you take a lot of snapshots and stuff you can see a petabyte)
Seems like an improvement but still a pretty inefficient system. Would be curious why they limit themselves to only 20TB of raw flash per X-brick when that brick requires 6U of rack space. Seems like there is a pretty severe limitation in the software and/or hardware that they have. What if the workload is not de-dupe friendly? You have to buy a bunch more X-bricks because you need more raw capacity.
Can XtremIO be scaled out on the fly yet? I read or heard somewhere somewhat recently I think that you could not add more bricks to an XtremeIO system online, you had to buy the footprint up front(or perhaps take a big downtime for data migrations etc). Not sure if that was ever the case or if it was if it is fixed now or what. If it's not fixed yet I'm sure it'll be fixed at some point.
brings new meaning to
Re: works well for me
on that note per the article saying itunes not supported, you could hook up tv mobli which can import itunes and access the data through DLNA.
I am not affiliated with tv mobli, but the product has worked very well for me and is pretty cheap and easy to setup.
works well for me
I have used a WD TV since last fall exclusively with TV Mobili running on a linux box (DLNA), works very well. It doesn't play flash video (I have a few downloaded music videos that I need to convert from flash), but other than that I have no complaints.
TV Mobli says I have 32GB of music(3k files) and about 2T of video(2900 files), and thanks perhaps to it's indexing browsing from the WD TV is instantaneous. Haven't had a need to try NFS or other data access methods.
StorSimple was acquired by Microsoft a couple years ago. HP has StoreEasy as their brand name for their windows-based NAS appliances/gateways(not the sort of thing I think that would be competing in this space). StorSimple is for sure not a competitor to any of these products they are more of a file serving engine with a cloud back end, not transactional.
Maybe you were thinking of StoreVirtual (as in VSA) though you did mention Lefthand (maybe you were referring to that as the hardware appliances).
I suppose it's news like this is why I haven't read vmware related blogs in a few years now, none of it excites me even remotely. This does though
I'll be happy when I get my 4-node 3PAR 7450 though.
Re: netapp is scale out
HP's website is reeeeeeeeeeeeeeally slow maybe they need to reboot their moonshot servers or something.
But what I was going to say is the 3PAR 7450 goes up to 460TB of raw flash, and the scale up 8 controller 3PAR 10800 goes up to 3PB of raw flash (1,500 x 1.9TB drives which can be ordered starting in October from what I read from Sandisk).
So comparing 36TB is not really fair. Perhaps hence NetApp working on flashray.
netapp is scale out
The 8080 EX is a scale out 2-controller system.
"Technical specifications for the NetApp FAS8000 Series unified scale-out storage."
(there is a hint in there somewhere)
Max flash capacity is 36TB (in a 2 controller system).
3PAR is scale up, and they would like to consider scale out (with peer motion) though I think the scale out part isn't so elegant as the netapp scale out.
I'm going to be ordering a 4-node 7450 pretty soon with the new 1.9TB drives, and the minimum configuration is about 27TB raw flash (16x 1.9(in my case)TB SSDs 8 per controller pair).
reminds me of
(then) Xiotech's "cash for disk clunkers"
a day trip?
looks like Chessington is only 16 miles from what google maps seems to consider down town London. Doesn't seem far at all.
Re: Don't knock water
HP's new Apollo 8000 can run on water temps as high as 86 degrees F (@80kW/rack), which I thought was pretty crazy, maybe it is typical though.
can they get it from their preferred vendor?
Which OEMs use Permabit? Doesn't seem like a good question to ask unless/until we know which OEMs are using this fancy Permabit technology.
what does that mean
Is the EQL tech in this system basically invisible to the end user? Does it do the same sort of "scale out" over iSCSI that EQL does? (I assume not).
Re: No influence?
Perhaps you misread the article, this has nothing to do with vmware's own appliances/etc that use SuSE. This is a program that allowed customers to use SuSE for general purpose stuff(e.g. install via ISO image and get a regular linux install)
some may not even know
that they have a NTP server running.
Me for example I was not aware that the IPMI interface of my supermicro server at a co-lo had a NTP server running (I knew it had a NTP client). My ISP notified me a few months ago that the IPMI interface participated in a DDOS attack and I shut off the NTP client (and thus server apparently which surprised me).
It is a standalone server, so there is no firewall or anything protecting it. It is a personal server, not a business thing.
Then last week I kicked my IPMI interface offline by upgrading the firmware to fix that security problem (a problem I could not even tell if impacted me the advisory was too vague and there was no changelog information in the firmware update). So some day I'll have to drive out on site again and re-ip the interface. Love that supermicro..... (not for business)
you read way too much into this. Dell is just saying they aren't as competitive as they want to be storage wise so are getting another product to offer. They already partnered with some other company that provides a converged offering(sorry forgot the name, smaller player).
VMware partnering with supermicro would be a mistake, you don't buy premium software at several thousand dollars a socket and plop it on shit hardware,or if you do your an idiot, fork over a bit more and get some quality stuff, if your going to be consolidating a lot of VMs onto fewer systems anyways. Critical features like a working remote management, Advanced ECC or Chipkill to me are vital for any dense VM platform. I upgraded the remote management on one supermicro system I have in a colo last friday, guess what I still can't ping it, I have to go out on site and fix it manually, at least it's no longer vulnerable to that security exploit since it is no longer on the network (it was on the network because I have exactly one server at the colo for personal stuff, it is not for business use).
supermicro is ok for throwaway stuff, or perhaps really large scale with tons of custom management software on top. Vmware isn't that, probably never will be. The cost of their management layers keeps on going up, making the underlying hardware seem cheap by comparison.
What perhaps is a more interesting angle to this whole thing with Dell is how might it hurt their own storage sales even more (they have already been on a decline recently). Dell hasn't yet shown it's able to compete.
Plenty of people throw rocks at Dell, especially when it comes to storage, or blades for that matter. From what I've read this new "deal" is not much more than a reseller arrangement, and perhaps Dell providing some tier 1 support or something, so it really doesn't change anything.
I don't read most blogs so I don't know what goes on out there. I stopped reading most vmware blogs a couple of years ago.
Nutanix sounds like they have an interesting solution, not something I would consider for my mission critical stuff, but if I was in a larger organization with more sprawl I could see myself trying it out for some edge cases. As-is I'm at a smaller company and the $ value is high, so that means something more industrial grade.
if your not a fan
then you should install this plugin if you haven't already, makes reading about cloud enjoyable.
To quote the last paragraph of your article
"The value of the service trumps the risks – real and imagined – of it being in my butt."
chrome(I use firefox):
(I have used that for the past 8-9 months now, I wish I could add other words easily to be replaced but haven't figured out how)
stuff like this
is why I almost never use public wifi hotspots, I just use the mifi on my phone instead. Even at hotels (another reason is performance is usually universally bad) I usually stick to mifi.
my most recent trip was about 10 days and according to my phone I used 1.8GB of data(5GB monthly committed rate).
Re: Good for Nutanix, Bad for Dell
In speaking with the HP storage and cloud folks at Discover they firmly believe that the Storevirtual platform is hyper converged. It may not be as pretty perhaps as these startups, but they have sold a lot of these systems(likely more than Nutanix and Simplivity combined a few times over). Their vision includes seamless replication and peer motion between that and 3PAR as well, though that doesn't exist yet today.
I'd say whether or not it is hyper converged is somewhat debatable (same goes for VSAN), but at the most basic level it seems to be (storage+servers running on the same hypervisor w/o dedicated external storage).
Don't know what Dell has these days I haven't been following them closely, but it seems like they continue to struggle to execute on the storage front, unfortunately they do seem to have some decent tech via acquisitions(outside of equallogic).
who would buy this?
Outside perhaps of Violin's existing niche customer base? The company's future is obviously not certain, this is brand new tech not battle tested, the cost is high etc etc.. I'm sure there are use cases where their stuff out shines everyone else but myself I don't see any way they can become truly competitive in the more mainstream market segments, they are just too far behind. On the same note I am not sure who would want to buy the company either.
to put quarters in the SDN controller, gotta keep the meter fed.
they have regulated it, and it's banned.. doesn't sound like this is a new ban. I still won't go to SFO, don't want to get attacked by hipsters riding their segways or whatever. I avoid that place at almost all costs.
""Our software," they will say, "has been designed from the get-go to use flash and be aware that it wears out with repeated writing, unlike disk. It minimises the number of writes by coalescing them and deduplication to get rid of redundant data.""
I think HP's new "unconditional 5 year warranty" on their 480GB, 960GB, and 1.9TB SSDs says they too have the software that is built for flash, with the adaptive sparing, and adaptive write cache specifically (among others too). That 5 year warranty exceeds that of the manufacturer of the SSDs themselves (Sandisk). Not to mention 99.9999% guarantee as well.
Pure is dead, they got nothin. Two years ago they looked interesting, now they are dead. I know a few folks (ex-HP) over at Pure too.
Some of the other players like Nimble(hybrid disk/ssd) will be able to nibble around at the low end for a while yet. Tintri is dead when VVOLs hit the scene.
Re: Another reason not to have every tech company in the same small area
who cares? apparently a lot of folks since more and more are coming all the time it seems like.
The job market is pretty crazy here for sure at least for tech, but for the most part I just see a bunch of companies I have no interest in working for. Having to shoo away 1-3 recruiters/week. It's gotten to the point where I don't even bother replying on linkedin anymore(eventually I do usually it's a week or three later though)
I have a 1U server at a colo for my own personal stuff, and I could not tell if it was impacted or not the docs were not clear. So I decided to upgrade anyway just in case since it hadn't seen an IPMI update in about 3 years now.
This server is directly on the internet because well it's the only system there, I have contemplated putting a Soekris in front of it though my IP space is limited.
Supermicro sucks for not having changelogs on pretty much anything to start with.
Next off in their instructions they say in big red letters
"NOTE !!! Uncheck preserve configuration box during flashing (very important step for FW to work properly). All settings will be reset to default."
I was going from version 2.x to version 3.x if I recall right.
later on in the documentation it says
"1.8 Click < OK > System will reboot after upgrade complete. The web page will redirect to the login page automatically."
So that implies the web page will still work after the upgrade.
So I thought - this is Supermicro, so I'm thinking when they say "reset to default" that means what they say. Which means once this thing reboots there is no more connectivity to IPMI and I have to go on site to fix it.
But I thought, maybe.. just maybe..they preserve the IP address info and perhaps authentication info.
I clicked upgrade, it upgraded, and rebooted, and well that was about 3 hours ago and the ipmi is not responding to pings even.
Fortunately the data center is only about a 45 minute drive away w/o traffic and I am not in any urgent rush to get it fixed I can fix it on the weekend.
But just goes to show, you get what you don't pay for..
I can live with it for my own personal stuff but wouldn't use this in a business of course (I have used supermicro off and on for the past 11 years, so this experience is par for the course for me).
Async streaming replication and file /object stores on 3par nodes announced a year ago if not earlier(that is when I saw them speak publically on it anyway)
Im another el reg article I pointed to videos hp posted of that event, which was sveral hours long. Not hard to find but I am on my phone so kinda pita.
Real time compression coming too if you want to write about a rumor on that(see above earl).
hp sells a lot of sandisk
not just the new stuff for 7450
90 iops is high?
Confused I am.
Havent tried to verify because public cloud is a waste of my time but just seems odd in the world where people quote 10s of thousands of iops on a ssd that this gives you the same number as a 7200 rpm disk.
Maybe thats what it takes to get predictable performance in a public cloud.
they already are
A couple of weeks ago people at my org asked me to block large swaths of Africa due to abusers on the web site. One of those ranges was 126.96.36.199/8 (well that is what it came down to there were so many 154.x subnets that these people were coming from). Our business does not serve anyone in Africa and the developers were going to fix the bugs that the people on those IPs were exploiting in a few days anyway. So I asked our CDN to block that range (amongst others).
status: ALLOCATED PA
source: AFRINIC # Filtered
parent: 188.8.131.52 - 184.108.40.206
Not long after some ISP in Canada started reporting issues of users not being able to hit our site, and they were in the 154.x range too (I didn't get a specific IP).
I asked the CDN to unblock the 154.x range a couple of days later(kept the other more restrictive blocks). But found it sort of interesting that those IPs ended up in Canada. In the grand scheme of things it was better to keep them blocked(until the issue was fixed) then continue to let those folks exploit that issue. Everyone at my org understood that blocking IPs was not a sustainable defense since there are infinite proxies and clouds etc.
Meant 22tb per U
The 45 number came from my friend jack Daniels
3par 7450 blows em all away...45tb raw per U
Over 400T raw flash and over a PB usable and 6 nines reliability gauranteed in a single system.
At under $2/gb usable
Seems pretty reasonable to me if they are in fact taking what VCE says is needed to support such a workload and saying "this is how we run the same workload"
the burden is on VCE to come back and say HEY you can do it for a fraction with xtreme io if that is the case or some other config.
Re: Cloud - or just a big hosting operation?
Just? Verizon bought terremark years ago for their cloud. They also announced partnership with amd seamicro for compute and block storage.
All this stuff is here is object storage a small part of the total cloud offering
- Asteroid's SHOCK DINO KILLING SPREE just bad luck - boffins
- Just TWO climate committee MPs contradict IPCC: The two with SCIENCE degrees
- BEST BATTERY EVER: All lithium, all the time, plus a dash of carbon nano-stuff
- Stick a 4K in them: Super high-res TVs are DONE
- Review You didn't get the MeMO? Asus Pad 7 Android tab is ... not bad