Re: May I recommend rsyslog?
It wasn't uncommon to syslog to a remote machine which would print. We were doing that in the 1980s in telcos
15086 publicly visible posts • joined 8 Feb 2008
"The Titanic's problem was that it had a double bottom but not double bilges."
Bilges weren't so much of a problem so much as the "waterproof compartments" not being sealed at the top.
As each one filled up, the bulkheads were overtopped and water fillled the adjacent compartment. If they'd been sealed it would have been swamped but stayed afloat.
That's quite apart from the deficient crush structure White Star adopted vs P&O, in order to have the big open spaces internally and a reckless captain who treated the liners like speedboats (he'd already badly damaged Olympia off of Liverpool causing it to be dry-docked for months)
TItanic's best chance of survival would have been a head on collision with an iceberg - it would have killed the entire crew of ~200 stokers asleep in the forward compartments but the ship would have survived just fine. On the other hand we wouldn't have SOS and all the marine rescue stuff that resulted form the loss.
"I don't think that's what Scott was saying-"
Scott's hypothesis is that the fuelling lines broke loose - and he showed slowmo footage of the thing blasting itself into the ground when it vented methane out the top after pushing the 20 tonne mass adaptor sitting on top off at about 100km/h (which happened when the entire stack "jumped" thanks to the explosion under the base)
ie: This was a setup/ground crew error, not a rocket/vessel failure.
"It is also building lots of nuclear power stations, which is why it now seeks to build them for others"
most importantly, China's investing $$SHITLOADS into molten salt systems (far safer than steam bombs) and aims to get those into production sooner rather than later.
If they get it working, it solves the "nuclear waste problem", makes the entire uranium enrichment system clearly military in scope and drops the price of the fuel by a factor of 200 whilst improving safety by a factor of at least 1000 and thermal efficiency of the generation plant by 50%
What do you do when your light water plants are shown up for being the crude, dangerous rube-goldberg contraptions that they are?
The Datahoarder community did document it. Chris made no secret of the discussions in that or that I brought it to him in the first place
Other "journalists" have been less open about it, and there's been a lot of flat out plagarising of Chris's reports without attribution, or crediting him for the story without looking at what he wrote.
A lot of the astroturfing pushing WD smells less of Stockholm Syndrome and more like having the same grammar and language constructs as the WD PR and manglement twats I was dealing with before I went to Chris.
"I'm curious as to why the rebuild actually fails."
Because the drives have a firmware bug and when given sustained non-sequential writes (ZFS RAID rebuilds aren't sequential. RAID5/RAID6 may not be either) after 40 to 90 minutes they start throwing bogus internal errors "Sector ID not found" - you can see this in "smartctl -x" views of the drive internal extended log) and throw a HARD write error back to the host computer
The only way to stop this happening is to idle the drive after 40 minutes activity for 1-2 hours to allow it to write out its CMR cache area (about 100GB) before you can resume the rebuild.
"I would like to see some proper data on how this affects the performance of the drives in their intended use"
I kicked this off with Chris because the replacement 4TB REDs I got ti replacing aging ones in my home NAS went from 18 hour RAIDZ3 rebuilds on the old drives to failing to complete after 21 DAYS and WD were blowing off complaints.
RAIDZ3 rebuilds were _still_ 18 hours on the old REDs, and on Toshiba N300s + Seagate IronWolfs purchased to crosscheck. The "New" SMR REDs were the same price as the CMR REDs they replaced but more importantly the CMR drives were simply no longer available anyway. The N300s and Ironwolfs were slightly cheaper than the SMR REDs
That's when I started digging, found out about the shingling and realised they had bad firmware - both problems WD were denying but were fairly well documented in tech circles. THEN I found out about the SMR Barracudas and realised "ghost" networking problems my group had been chasing for over a year weren't caused by networking issues - we verified it was down to the SMR drives choking the desktop computers they were installed on.
WD and Seagate were empathically _refusing_ to tell media or customers whether drives were SMR or CMR. It was only when Chris confronted them with retailer lists (skinflint.co.uk) that they relented and admitted it.
"But when a drive failed, only about one third of the time was it preceded by SMART monitoring errors."
Conversely: IF you know what you're looking for, monitoring SMART can tell you a drive is having trouble long before it even issues a SMART warning
"But the competitors are SSD drives."
There's SSD and there's SSD.
SK Hynix have shipped some utterly shitty NVME drives (less than 140MB write speeds and 300MB/s read speeds) which HP have bundled into their desktops and sold as premium product - but because they refuse to publish specs on what they put in their products they got away with it.
HP Europe response to complaints has boiled down to "We got your money. Tough luck!"
Our response was "You blew any chance of future sales and jeopardised tens of millions of dollars of nationwide future contracts on the supply framework we use. Oh dear, how sad"
(Disclosure: I'm the one who brought this mess to Chris' attention along with various tech mailing lists and notified ixSystems about the F/W bugs. Chris verified the issue, got admissions from WDC & SGT which they'd refused to give to consumers plus a voluntary statement from Toshiba about their use of SMR in consumer products - I'd assumed Tosh had been more ethical in disclosure up to that point)
ixSystems have withdrawn their recommendation for the SMR drives - and they verified my report of FW bugs in RED SMRs.
The reason they gave that recommendation was that WD RED CMR drives have been hellaciously reliable. My old REDs have run 8-9 years 24*7 without missing a beat whilst other drives tended to only last 5-6 years. WD built a reputation and trashed it in a matter of 18 months.
As for the Class Action - this COULD be a submarined way of WD heading off consumer protection "death of one million paper cuts" in US courts. By certifying a class in USA courts it means anyone filing in lower courts (small claims) is likely to find their claim pulled into the class action.
Don't forget: WD, Seagate and Toshiba have been explicitly marketing SMR drives as suitable for home and SOHO use. That means the "Get out of jail free" card that was used against misleading advertising litigation in the past ("these are business devices and businesses must do their homework") is ripped up and thrown away. WD compounded the crime by selling the things explicitly as suitable for RAID use when what they meant was "Mirroring", then gaslighting consumers who complained (as did Seagate) and then issued press statements gaslighting the public (Say kids, Can you say "Exemplary Damages"? I knew you could.)
In this case the class action is targetting WD. I doubt it will be the only one and I'm doubtful consumer protection agencies around the world will sit on their hands - particularly when they look at the combined actions of the three companies
If the rollout of DM-SMR in consumer drives had been "coincidental", then at least one of the makers would have disclosed it on the spec sheets and/or would have released SMRs at a considerably lower price point than their CMR product and/or would have pointed out what the others were doing.
If it looks like a cartel behaviour duck , quacks like a cartel behaviour duck and walks like a cartel behaviour duck, it's highly unlikely to be a coyote.
Icon, because Micron have parked SSD tanks on HDD maker lawns.
You can buy their 5210 ION "cold storage" (meaning 0.2-0.8DWPD, something akin to archival SMR HDDs, not "powered off") SSDs for twice the price of Enterprise HDDs of the same capacity (or about 3 times the price of a RED/IronWolf/N300 - and the quoted endurance of 180TB/year for RED or IronWolf or N300 NAS drives is a lot LESS than 0.2-0.8DWPD)
Not bad for an ENTERPRISE drive that draws ~2W, has power failure protection and a 5 year warranty (the HDDS referenced above are all 3 year warranty)
RIght now on Insight, UK IONs are listed at £308/4TB($380), £580/8TB($680) +tax - and the way SSDs are still falling in price they'll rpobably hit parity with the Enterprise SMR drives by the end of 2020
"Sure they just change the socket 90% of the time just to force us into that extra upgrade and not because some new chip model actually requires it"
Intel: yes
AMD: no
Even to the point where the latest AMD APUs use the same AM4 socket but can't be run on older boards due to bios constraints (the older APUs can only address 16MB in real mode) and vice versa. There are some kludged workarounds coming to allow it on enthusiast boards but they involve serious hackery of the underlaying codebase to make it all fit
Intel would change the socket and force the issue
"I lost count of the number of people who were very surprised at this asking how we travelled so far in a day."
I lived in New Zealand - which is only about the same size as the UK. Driving 100 miles each way for a decent coffeehouse and night out wasn't considered unusual. Even in a country that size it was perfectly possible to live places where the nearest neighbour was 20-30+ miles away.
Then again, consider it can take 2-3 hours to drive the 19-20 miles _across_ London, vs under 2 hours to travel the ~100miles between the ring motorways connecting London and Birmingham or that it's 90-120 minutes on the commuter network from the outer London suburbs to St Pancras Station, vs a fraction over 2 hours to Paris from St Pancras).
Where I live now, I can _see_ Canary wharf on a clear day. It takes 2 hours at best to get there (3-5 on a bad day). Brighton Beach is 3 times further away but I can be there in 45 minutes. You eventually start thinking of things in terms of travel time rather than distance
It's a common failing of organisations outside the UK to assume all distances are equal when planning appointments for technical staff - which can lead to some "interesting" conversations with manglement who really don't understand that "it's only 5 miles, how can it possibly take 90 minutes??"
"those of us who don't fully comprehend Australian-style distances"
Australia is about the size of the US lower 48 - or would completely cover western Europe right up to the Russian borders: https://www.mylifeelsewhere.com/country-size-comparison/united-kingdom/australia or https://www.mylifeelsewhere.com/country-size-comparison/united-states/australia
Darwin to Sydney is about the same distance as London to Moscow (or London to Tel Aviv, or Seattle to Houston)
China is about the same size as Australia: https://www.mylifeelsewhere.com/country-size-comparison/united-kingdom/china or https://www.mylifeelsewhere.com/country-size-comparison/united-states/china
Plugging in a few equatorial countries (try Ghana vs United Kingdom) can give an understanding just how distorted the standard Mercator projection makes our view of the world. It's a navigator's map, not an accurate sizing one.
"suggesting there was an initial wave in UK/Europe last autumn..."
1918 seems to have started around Kansas - already deadly when it struck there and travelled to USA military camps - and went to Europe. It changed and came back to the USA. Nobody tries to argue that what left the USA was "Wave 1"
In the case of what's been found in Europe/USA, it's clear it was out and circulating sooner than people thought - but that could have been a less virulent form too. My GP is already suggesting the "extreme flu" I had in late December-early January (and still suffering after effects from) was Covid, although I'm personally doubtful and tend towards H1N1. Without swabs at the time or accurate antibody tests it's anyone's guess. (Amongst other things I lost smell entirely for a couple of weeks and went completely deaf for 3 days/impaired for 5 weeks)
"You had a cumulative known total of 321."
That was a cumulative total of people who had died in hospital after confirmed diagnosis, when there were no fucking diagnostic kits available.
You can warp stats around in any number of ways if you restrict the way you express your datasets
The best overall proxy for covid19 death rate is the excess population deaths over the 5 year average deaths for this time of year and that's painting a VERY bleak picture.
It was only when the FInancial Times started weighing on how appallingly bad things are that the rest of the media started sitting up, stopped reportng the govt "official" figures as if they were gospel and started making noises to say "hospital deaths" or other footnotes to say the numbers were incomplete
https://www.ft.com/content/40fc8904-febf-4a66-8d1c-ea3e48bbc034 - currently pointing out that ONS stats show 50,979 deaths vs the govt's claimed 32,065 - the curves showing at the bottom of the page are nice, but they make it clear that this is the FIRST wave of the disease and there is worse to come(*)
At the time the UK government was saying "22,000 deaths" the figure was already PAST 45,000 and when they said 32,000 (after "Adjusting for care homes") they're still missing out around 1/3 of the recorded deaths above the baseline.
An awful lot of people are dying at home, or in care homes _without_ the benefit of a covid diagnosis and medics are afraid to put "possibly covid" on the death certificate without a test result as the health authority manglement have been making various noises about funding, etc. They know they're going to be facing corporate manslaughter investigations and the better they can make numbers look, the harder it is for prosecutors
This is going to be the Potters Bar event for public health investigations and politicians will be held to account. There have been too many deaths for the to deflect to 3rd parties.
Opening up too soon will result in things taking off like an exploding boiler. Wuhan, Singapore, South Korea(**), Germany and other places are already finding this out the hard way
(*) Every pandemic in the last 150 years has had multiple waves with the second being significantly larger than the first (even the last Ebola outbreak). looking at https://www.cdc.gov/flu/pandemic-resources/1918-commemoration/three-waves.htm should be eyeopening
Some countries actually TEACH this stuff as part of history due to the impact it had. Samoa lost almost 1/3 of its entire population in 12 days in 1918 - and the British governer behaved almost exactly like Boris - refused to allow quarantines, then to accept responsibility, running away to New Zealand as fast as he could - leaving his assistant to deal with hundreds of grieving families - then back to England when investigators tried to detain him for a commission of enquiry. Most Pacific Island countries saw 20-25% death rates - that's why they locked down tighter than a snail's arse when this one came along.
(**)Seoul: 1 carrier, one epic pub crawl to celebrate the night of the easing of lockdown. 30 confirmed infections already, another 7200 people exposed. It only takes a couple of carriers to restart the blaze
Really? http://offloop.net/covid19/ (which I assume you're looking at) shows "reported hospital confirmed covid deaths", not "excess deaths above baseline" for the UK. That big jump in the curve was when it was adjusted to "reported hospital and carehome confirmed convid deaths" for the week.
In reality it's up near Belgium and Belgium has 1/6 the population, mostly geriatic. Offloop's admin is aware of the ONS API but it runs a week behind the government announcements ( I asked him to plot UK excess deaths as a separate line)
and 32,065 "hospital deaths" is significantly different to 52,000+ "excess deaths above baseline"
Unless there are some REALLY stealthy ninja assassin squad carving their way across the UK countryside it's a sure bet that Covid19 is the culprit for most-if-not-all of the line above the "normal for this time of year" figure.
(especially since "influenza" "road crash" and other "accidental" causes have taken a general dip - the first for general reporting reasons, but the other two because they genuinely have dropped - it's even more pronounced in South Africa where the rise in Covid deaths has seemingly been outweighed by the drop in the murder rate.)
"Population density seems to be a bigger factor."
Not really.
Packing hundreds of confirmed infected seniors off to care homes and telling those care homes that they will lose ALL funding immediately if they don't take them (yes, really) is on par with trebucheting infected corpses over the wall into a beseiged town
It's no wonder care homes across the UK are seeing staggering death tolls with most seeing between 1/2 and 2/3 of their residents die in the last month - one site reported losing 85% of their residents in a single week.
The words "corporate manslaughter" are appearing more and more often in connection with covid deaths
"Competent countries do actually have pandemic plans in place"
Whereas the UK fared so badly in BOTH of the last 2 pandemic planning exercises (the germane point being they were under the same government) that UK dot gov kicked the results under the carpet and ignored it whilst whistling loudly if anyone bought the subject up and talked about the exercises before those.
"the convenience factor is being able to instantly adjust the lighting or heating as you walk in or out of a room "
Get it _right_ and the lighting will adjust as you walk in/out of a room, heating will sort itself out and NOT come on in rooms with the windows open, etc.
It's a bit more fiddly to setup and put the sensors/bridges where needed but if you tie it all together it means you have a system which works and you don't need to touch anything - and if you really want you can even setup some old fashioned "light switches" (433MHz transmitters) wherever you find convenient rather than where the sparky decided to shove 'em 40 years ago.
It's been pondered - and the general concensus is that lasering from on-orbit is impractical and even more problematic than lasering from the ground.
Laser Brooms are perfectly feasible NOW. The problem is that they can be used to bring down other countries' satellites as well as debris and technically the debris that's up there is the responsibiilty of whoever made the mess
Unless/until there's an international treaty and an international organisation tasked with doing the job that has FULL transparency/oversight from all parties, any one party deploying one could be construed as an Act of War by other spacefaring nations
Good luck getting the USA and China in the same room to agree on this, let alone China and India or India and Pakistan
"It's instructive to compare the history of the James Webb "
The problem with spacecraft is that launchers are sodding expensive.
Which means you only get one launch and one spacecraft, (and maybe a backup if the first goes bang)
Which means you heavily invest in making sure your spacecraft works
And also means that you add every bell and whistle you can think of to your spacecraft, because you only have one launch
Which puts up the launch mass and complexity
Which means you need to invest even more heavily in making sure your spacecraft works
Which means you build hundreds - if not thousands - of test articles
Which means in the end, your spacecraft is valued in $billions, but only has an actual materials cost of a few million - and you have a backup in case it didn't fly - and if you wanted to build several of them you'd be able to without even making much dent in the project budget
Except that doesn't happen
Because space launches are primarily a dick^H^H^H^Hflag-waving exercise and the science is in a distant second place by comparison to the "Gee, aren't we great!" behaviour when the thing is launched.
And space launches are like that because they're so sodding expensive.
Cheaper launches means simpler payloads are more easily justified, and really cheap launches means you can even afford to send up a dud occasionally without it being a career-ending move.
And it might even mean that development moves fast enough that a mission might actually take less than someone's entire career from first concept to actual use in space (seriously - taking 20-25 years from inception to launch is par for the course. There are several spacecraft sitting around that were completed and have been waiting to be launched for 20+ years, thanks to "funding issues" that cropped up along the way.)
James Webb is a good example of old-school thinking. It's big but hideously complicated because it was _too big_ for any existing launcher when first designed. The fact that better/bigger launchers now exist hasn't caused a redesign, but a redesigned JW might be flyable sooner because the single most problematic part of it is the sunshade and those problems are vastly reduced if you can have more launch mass and/or payload volume to work with. (The foldout mirror wings is another issue but not as problematic as the sunshade)
With something the size of starship, at the price of starship you could launch something like JW in two parts (instrument and sunshade/spacecraft) mate them in orbit and send them out to L5 with a tug sent up on a 3rd launch and STILL have change leftover from the cost of the original project.
"The problem with this case is that the launch vehicle was empty of fuel once the satellite was in GTO."
They're _supposed_ to be put into GTO so that the booster ends up atmosphere-grazing - that way if everything goes pear shaped before separation then it will all come down quickly.
The russians fucked this up in more ways than one. Exploding fuel tanks are a known issue on orbiting leftovers and why protocol on western launchers/boosters which can't be left in atmosphere-grazing orbits is to open all valves and vent EVERYTHING at the end of the mission
"From the above, it looks more like he committed a crime, ran, informed on accomplices to avoid jail-time and then proceeded to cover it up."
Yup, and at that point you have to start considering what else might be covered up.
He's running a company in an area which _requires_ transparency and as such not only demonstrates his own unsuitability for the job, the company (plus any more he might "start", join or have worked for) is irretrievably tainted by association to the point that goodwill(*) has instantly been destroyed and the bottom line badly damaged by the disclosure.
(*) goodwill as an accounting term for an intangible asset - it's usually the single largest intangible a company has on its books and the easiest one to blow
> How long before a person can be forgiven for past actions? How long before they are rehabilitated?
When you work in that kind of field you need to acknowledge ALL your past if you're going to tell your backstory. The fact that he selectively edited out is enough to make one wonder what else he hasn't owned up to.
> Not that we needed a reboot once it was stable - we deliberately left it running at the end of the project, just to clock a full year uptime :).
I had a linux box on a customer site (mail gateway/leasedline router before the days of DSL ) that clocked just over 3 years of uptime.....
A cleaner unplugged it in order to use her vacuum cleaner. Despite the DO NOT SWITCH OFF label on the socket and DO NOT DISCONNECT on the plug
> [users firing up dhcp servers] I also found that if you have an assistant, its better all around if you have your assistant disable the port (and the 2nd port they move their dhcp server to, and the 3rd port
I found it's better to have a switch smart enough to notice unauthorised DHCP servers and shut the port down all by itself. Also smart enough to notice (l)users who've decided to manually configure their system - with the IP address of the network gateway - and do the same thing, only in that case it STAYS disabled until we "have a chat" with them. Ditto if a port suddenly starts seeing multiple MACs (unauthorised switch, DoS attack, rogue VMs and in one case unauthorised wireless bridging)
Dickheads who wander around a lab causing every port to get shut down for security violations tend to find themselves rather unpopular with everybody else trying to work.
"LABEL THY PRODUCTION SERVERS. Even if it means you print a label out of paper and cellotape it to the front of the server."
THIS is why I insist on labellers being in the server room with supplies of appropiate tape.
If the labelmaking kit is right there, there's no excuse not to do it (Brother TZ labels work out at about 5p each, so not exactly bank-breaking.)
>>A big organisation should have had more than 1 DHCP server
> But not on the same subnet, otherwise arguments happen.
Why would you be assigning random IPs to MACs that just happen to appear on your network? Security FAIL right there.
If you have IPs assigned in dhcpcd.conf then everything works - and rogue boxes don't get an IP address.
Of course if you have 802.1x, they don't even get onto the vlan in the first place
Back in the early 1990s I did just that.
The organisation in question had setup email as an experiment (uucp on a 14k4 modem), with the understanding that my assistance was on a best effort basis and might take a few days to sort out as they have their own people who should be able to deal with things
Fast forward 6 months and I'm on holiday on the other side of the world. A paniced phone call comes in "email's not working, it's critically urgent. we need you here now!"
They weren't impressed when I said that as I was on holiday - which they'd been told both before I left and when they phoned the office insisting on speaking to me - I could be there in 2 days and it would cost them $3000 in callout fees plus whatever the airline charged for a walkup return flight, which I estimated would be not less than $12000 for a business-class seat, given they'd want me fresh enough to work on the issue as soon as I arrived, having crossed 12 timezones and flown for 26 hours.
Funnily enough, they declined. When I got back I was told they'd found one of the secretarial staff had emailed 200+ copies of "the dancing baby" clip (at 4MB apiece) to friends offsite. 14k4 was about 1MB/hour. Those emails got flushed, the link got upgraded and people got a lesson on how experimental systems can become critically urgent when useful