Posts by Nate Amsden

Re: So, how many credit cards do I need now?

suggest having at least one credit card account where you can generate virtual credit cards. For me that is Bank of America (Shop safe is the product, I use it all the time, though it does require flash to interact with). I also have other credit cards but it seems that particular capability is far from universal.

I think my (real) credit cards have been compromised 1 time(MAYBE 2) in the past 3 years at this point. I did have one ShopSafe card compromised, which is odd because only 1 vendor ever got the number(hotel reservation system). Because the original vendor did not charge the number, it was still "open" to be used(the moment it is charged it is locked to that vendor). About 2-3 months later a strange charge showed up from another website that I had never used, it was especially weird because there was only that one charge - normally I would see multiple fraud charges in a short time period. After some investigation I tracked it to the specific virtual credit card I used to reserve the hotel room. The vendor that had charged my card with the fraudulent transaction refunded the money. I sent a message to the hotel chain with the details but never heard back. Bank of America saw no need to cancel my main card since it was only shop safe that was compromised (maybe 7-8 years ago their reps/fraud system wasn't sophisticated enough and they would insist canceling my main card when shop safe was compromised even though there was no need, now they know better).

Few years ago I had another shop safe card fraud attempt (that was blocked). I used that card to pay my cable tv subscription, I forgot how I got notified of the charge, but once again the only company in the world that number was given to was the cable company, so the breach happened with them or with their processor. They were very apologetic and offered to pay for credit protection(local cable company not a big brand name). I told them don't worry about it there is no harm done.

so in general for me at least credit card security(whether it is chip and sign or swipe) really hasn't been much of a bother for me in many many years. I would say before 2010 my card(s) would get compromised on at least an annual basis, and it was more of a bother.

Re: Forgot a couple

Cisco and Dell are special, el reg gave them dedicated articles

http://www.theregister.co.uk/2017/07/11/ciscos_fifth_ucs_server_generation_surfaces/

http://www.theregister.co.uk/2017/07/11/dell_14g_servers/

possibly those vendors announced their gear (again) ahead(few hours early?) of the CPU announcement, not sure

Re: just be prepared to wait if you want SSDs

yeah I suppose I could try to get some vendor to make me a custom build with the CPU, was quite surprised that the online store HPE(sorry) has did not have that as an option for the DL20.

Re: This has been going on with Win10 for a couple years, Win8 too if I recall

Not that I need to sync with this but isn't the point of not using NTFS so the SD card is portable to other systems whether a camera, or something ??

Re: 2 vlans in same subnet is a bad thing

I suppose if you run your switches only layer 2 then having overlapping IP spaces isn't an issue(since the switch isn't tracking IPs) -- I have run pretty much exclusively layer 3 for the past 13 years.

Re: I'm sure CloudFlare serves a purpose. Somehow.

even at it's most basic level cloudflare is a CDN. And the goal of the CDN is to cache assets close to the client for faster performance. The security add on stuff I'm sure is nice for the clients that need it, my experience says most do not, but if it's cheap and/or easy to setup then there may be little harm in just doing it.

The only attacks I have witnessed myself in the past 24 years of doing internet server stuff -- attacks where something like CloudFlare's services would of helped were attacks directed at other customers on shared services. e.g. the dyn DNS attack last year- as a customer we were not the target but were collateral damage. Also last year our primary upstream ISP came under a ~200Gbps attack for about 3 weeks(off and on as the attacker shifted attack vectors), which caused quite a bit of harm to us (the attacker was going after someone else on the provider, not us). about 18 months ago our upstream ISP got hit again with a big attack that was mitigated in a few hours(that time it was one of the game companies - EA or something that was the target - there was lots of news articles about it at the time).

Having fancy DDoS protection when you aren't the target doesn't help, when the shared pipe(s) are saturated by attacks on other customers.

I have never used Cloudflare as a customer, so have no idea how well they work -- though they are attacked a lot - I suppose the upside is they are generally better prepared (the CDN we use says the largest attack they have gotten didn't go much more than 2-3% of capacity last I talked to them), but also means they are a much bigger target -- I recall on more than one occasion pretty major cloudflare outages due to attacks(at least one article on el reg).

If you don't have the need for edge CDN caching, and you don't run a site that is likely to attract attackers then you generally don't need something like CloudFlare. I'm sure the biggest sites have a combination of edge defenses as well as core defenses. But that is overkill for 99.999% of sites out there.

Re: Walmart understands the basics

Yeah. Quite shocking to me netflix is among them. Though 99% of netflix's bandwidth is served through their CDNs(many of which of course are on site at ISPs).

(Haven't been netflix customer since their first price hike, they lacked content I was interested in and I've read it's only gotten worse as they shift to in house content (from what I've seen not a single one of which is interesting to me.)) But same goes for hbo, showtime etc etc not 1 show of interest. (I did really love 'Strike back ' on cinemax but tgat series is over now.

I miss showtime's 'sci friday' from the 90s - combo stargate sg1 and the outer limits. Really miss stargate franchise too though did not enjoy the original movie.

Re: 5%

Less than at least 25% and I can't tell a difference (I probably could in a game though I don't really play games).

On the same note I can tell zero difference between 6gbps sata and 10gbps nvme (both samsung pro) on my quad core skylake lenovo p50 laptop. (Outside of benchmarks anyway).

My last phone upgrade was galaxy note 3 (still using it), quad core 2.3ghz probably at least 350% faster than the hp pre3 it replaced (single core 1.4ghz) and 600% more ram, and 36X more storage(currently 256gb sd+32gb vs 8gb on pre3, though when I originally bought note 3 it had 96gb total).

I have looked at benchmarks for newer phones and have not noticed anywhere close to a similar leap in specs. Note 3 works fine so no interest to change it.

Maybe when quad core 5ghz comes round with 12G of ram and 1tb of storage(with comparable battery life) I'll be willing to drop android 4.4. Not holding my breath though.

Re: Server prices rising

perhaps unit shipments of servers are down (I don't know off hand), but memory chips per server is up. I know all of my servers have 24x16GB memory modules a piece (including the DL380 Gen9s I just ordered yesterday).

Looking back in history, the DL380 G5 for example (about 10-11 years ago) seemed to top out at 8 memory sockets. G6/G7 increased that to 12 sockets, though AMD G7 could go to 24 sockets(I have a dozen of those left).

My G5s back in the day I think were 8 cores in two sockets with 32GB. Now 44 cores in two sockets with 384GB( 384 has been my standard since 2012).

Re: disappointed

Not sure what you mean by 1.x Ghz Xeon, in my OP both Xeons that I referenced were 2Ghz+.

If AMD could make a 32-core 2Ghz chip @ 130W, and charge 20% more than Intel's high end part(s) I would buy it without hesitation. Their wattage numbers are just crazy, even their 16 core parts are 155W (up to 170W ?? based on the article - which is higher than Intel's 22 core Xeons). This just doesn't seem like a good enough attempt to break back into the data center market.

AMD seems aiming squarely at the 1 socket market. I can certainly see some use cases for such systems, am not sure how much of the market they make up. Last time around (Opteron 6000) AMD argued you don't need 8 socket systems, when their chips can do 4 socket real cheap with lots of cores. Now they say you don't need 4 socket systems(which are a small part of the market I'm sure), but having a pair of 150-180W CPUs in a two socket system just seems crazy.

I want AMD to do well.. but they abandoned the data center market for the past 4+ years, it will take a lot to come back, and these wattage numbers make me think it is not enough.

Re: RMA the controllers?

As article stated the controllers were fine. The cables/connectors were more the issue (well really the issue was how it was physically setup). HP obviously did a poor job the people who installed the system were not knowledgeable enough. It's not THAT complicated.

When my 7450 was installed a 3rd party did the actual installation (distributor), even though I bought installation services. Anyway that guy cabled the system wrong. It is just a total of 8U with 4 enclosures. But he did it wrong. The system was working, but it was me that noticed at the CLI one of the enclosures was not right and one of the SAS ports was not used on the controller.. So I had them come back and fix it(fix was zero impact to operations though the array was not in use yet anyway). I believe even before the fix full array availability was there, just that shelf was configured to be in a loop config with another shelf instead of directly attached to the controller (higher performance ).

Customer unwilling to spend money on a proper backup system doesn't help either (my companies have not done so either).

At the end of the day this problem had absolutely nothing to do with 3par tech and was an entirely implimentation specific. Still HPs fault of course since they set it up. I heard that the specific person responsible for the cabling stuff is not with HP anymore.

Re: Vendor Lockin

HP has had a policy of no support no firmware updates for years now..though at least with proliants they don't do anything to really enforce it (e.g. you can download the Proliant service pack cd with a support contract from one system and use it on any other systems you want).

For me it would be nice if they fixed this in Gen10, which causes my Gen9 systems to hang on boot unless I am really careful(yes I boot from SAN):

http://h20564.www2.hpe.com/hpsc/doc/public/display?docId=emr_na-c04822613&sp4ts.oid=7500984

Re: Does Oracle

I see about 120 mysql servers in my network, though they are all percona builds, not the native oracle one. Though typically percona picks up these kinds of changes into their system.

Re: Shelf failure?

At one point there was a report of someone trying to move shelves around while the array was online(stretching cables etc). So maybe no actual shelf failure but perhaps an induced failure from doing stupid things.

Good that HP accepted responsibility for it. Also obvious lesson don't store your backups on the primary storage array, obviously budget issues caused that. The original report made it sound like multiple full array failures(not impossible just unlikely).

I got the real story as to what happened(as well as what happened in the UK) but will stick to what is public. Short story short i have complete confidence in my 3par arrays.

(3par customer since 2006)

Re: Never again

No they won't. At least not ATT. The upgrade notice explicitly says wifi is required. And i have stopped the upgrade(downloading) on several occasions by disabling wifi.

Not certain about Google. But am certain the likes of amazon, azure and even facebook make huge use of enterprise storage arrays internally. I'd wager google does too. Certainly not everywhere but I'd wager they have 10s of PBs of storage on enterprise systems.

The impression I got from reading the article is they might use the same hardware as VSAN ready nodes, because they fill a similar purpose.

Obviously it would run different software on top.

Re: It's called the SPC1

So, assuming you can't bring a dozen different array vendors in house to test them with your workloads, what other options are there other than widely used benchmarks. Shit some storage companies don't even do evaluations (NetApp being one, from at least 2006 until 2011 they(multiple people at NetApp explicitly told me they do not do evals -- maybe that has changed for solidfire etc). NetApp refusing my requests for an evaluation back in 2006 is why I ended up on 3PAR.

I totally agree SPC-1 is far from perfect, as are most(all?) benchmarks, but it's a lot better than nothing. The disclosures in the benchmarks like SPC-1 are very informative (e.g. showing that the workload in question is running entirely out of cache, as Infinidat likes to show, or if high availability is disabled on the array(as I have seen on one or more 2nd/3rd tier vendor arrays in the past). At least the benchmark gives a common workload that people can use for comparison at some level. I would like to see SPC-1 revised though. I haven't paid attention to it in a while(haven't paid attention to storage in general for a while my arrays just happily run along).

Bringing everything in to evaluate with real workloads is just not realistic for 99% of customers. I know there is a company out there that makes a high end storage load simulator, forgot their name, but even that seems likely to be only used by very very large customers.

I remember getting a presentation on Hitachi's AMS2000 series back in 2010 I think it was, just before the arrays were coming out(the 2300 came out 2 or 3 months after the presentation), and they had slides showing their systems could do 1 million iops or something like that.. Certainly sounded impressive, but it wasn't until I explicitly asked them that they admitted that was from cache(and obviously the AMS2k series had a tiny cache relative to something like Infinidat). I could absolutely, positively see the VP I had at the time making a purchasing decision with data like that slide in mind(granted he was an idiot, but there are lot of those out there, if I had to bet I would say the majority are). So I'd happily take SPC-1 over a marketing slide like that any day of the week.

I remember another presentation(maybe 2012), it was from X-IO (at the time they were still called Xiotech), they did tout SPC-1 numbers, they showed themselves as #1 for that particular metric, and then they showed 2 or 3 other vendors (they just named them "Vendor A" "Vendor B", did not name the real names). I think the metric was IOPS/disk or something, or IOPS/GB or TB. I recognized the 3PAR result at the time, and asked them, and they confirmed yes the #2 system was 3PAR, I remember my boss(Director, not the VP) saying "oh, looks like we made a pretty good purchasing choice then", thought it was funny. We brought Xiotech when they came up with a trade in program where they give you free storage and you give them old storage. We had about 150TB of old storage from BlueArc (they used LSI storage at the time), so we thought hey maybe we can give this shit to Xiotech and they give us Xiotech boxes for free, but in the end they had no interest in that old storage we had so they went away. Ended up having to pay someone to take the old stuff and toss it out, I had a friend who worked at a recycling company for IT stuff, he initially expressed great interest in those racks of gear, but after I gave him more in depth information I guess they decided there was no value in it after all so they didn't even want to take it to recycle (their thing was give them your old IT shit and you get a tax write off).

Re: What's missing is...

ARM remains as stillborn in datacenters as intel is on mobile.

For my newest home server I got a xeon e3-1240L v5. Quad core 8 thread 2ghz 25W TDP. With 16g of ECC ram(soon 32gb) 6 HD 2 SSD and low power (19W) nvidia video card, and 4 80mm fans, the whole system(70W) uses less power than the triple core amd cpu alone that my home server was using before (90W).

I just love this cpu. But holy shit is it hard to find. Took 2 months to get one. Seems very strange for a current gen xeon to not be sold by anybody. After waiting a month from one vendor to buy a 1235L v5 for about 300, i found the 1240L v5 from dell for 500. Still took dell a good 3 weeks to deliver it. I left the other order for the 1235 on for an extra 2 weeks the vendor kept saying the distributor will have it in stock next week.. thinking I would use it to build another box. But I canceled it in the end.

It rips in encoding blu ray too, easily twice as fast as my dual socket 6 core (12 total cores) opteron workstation. Mainly because the video encoder doesn't scale to 12 cpus on blu ray for whatever reason.

Main thing is i wanted a chip/system that could easily survive the 100 degree summmer that is coming.. my amd system lasted fine last year but the chassis is just too tight. I don't want to risk another hot year on the old system(4 or 5 yrs old). New system has great filtered airflow front to back. And the cpu cooler is built for a 90w("normal" chip.

Server side i am interested to see how AMD's new 64? core chip comes out. My work servers run 22 core xeons today and I'd love to have more cores (still have about a dozen dual socket 16 core(ea) opterons in use) vsphere loves cpu cores.

Re: Article title is misleading

I agree should say something like "Do you use squirrelmail with sendmail.."

on that note, as a squirrelmail user for 17 years now(even though I use roundcube today I still have SM installed for some family members who use it, last used SM in an office environment probably 2002), even back in the days when I did use sendmail I have always had squirrelmail just use smtp to localhost to send email. Not sure what the advantage ever might of been to using a local binary instead of smtp. I certainly never got any complaints.

Re: Reboot

This doesn't make sense to me. Upgrading the kernel(without reboot) doesn't remove the old kernel in my experience. It doesn't change anything. I have never ever in 20 years seen a problem from upgrading and not rebooting like you describe (symbol issues and stuff ) on linux.

On my mint systems I don't upgrade kernels often maybe once a year. My last upgrade(very recently) had problems with sound (maybe those pulse audio problems I read so much about, fortunately there was an even slightly newer kernel that I was able to upgrade to which resolved the sound problem). I don't plan to change kernels again this year on my main system, too much risk. Two kernel updates before my laptop would panic once every few days. The kernel before that wifi and SD card didn't work. So am very weary of changing something that is working as a result of that (laptop is lenovo p50 which has high linux compatibility)

The security issues aren't nearly as bad(as in risk you will be attacked) as some try to make them out to be. Or maybe I've just been lucky running internet facing linux hosts since 1996.