RSS
Sorry the inconvenience of the server crash Friday afternoon, which involved absolutely nothing nefarious.
As you’ve noticed, the website has now been restored from backup, and while things are being tested and fixed, commenting has been temporarily disabled.
Unless unexpected problems appear, commenting should be restored shortly, but we would greatly appreciate if your comments addressed the substantive issues of the various articles on the website rather than too much discussion of the unfortunate server crash.
UPDATE: Unfortunately, another server problem forced once again us to re-install our website from last night’s backup. My sincerest apologies for all the comments that were lost over the last 18 hours or so.
UPDATE #2: Most of the problems reported, including comment-ordering, new-comment-tinting, and comment-edit-window errors have *finally* been fixed. The Search system is also working again, though with some problems. Feel free to alert us to any remaining bugs on the new server.
English
Arabic
Chinese (Simplified)
Dutch
French
German
Hebrew
Hindi
Italian
Japanese
Korean
Persian
Polish
Portuguese
Russian
Spanish
Swedish
Turkish
Vietnamese
Some content that went up shortly before the crash seems to have disappeared (e.g. the latest post on Anatoly Karlin’s blog, a post about Catalonia with about 20 readers’ comments)…I assume those posts are gone for good?
Unfortunately, yes, though perhaps Anatoly might have kept a local draft of his post and can republish it.
One reason the website was off-line so long was that the server people kept assuring us they’d probably be able to get it running again shortly, and it wasn’t until early this morning, 15 hours after the crash, that they told us the data was probably unrecoverable. Our backups only run daily, and until then I’d been hoping we wouldn’t have to revert to the previous backup and abandon all the comments and posts.
Frankly, the whole thing is simply astonishing to me. We leased a top-of-the-line cloud-server configuration from America’s leading webhosting company, and I just assumed that nothing like this could possibly ever happen. My mistake.
The search option/engine seems to be broken for the comments histories/archives after the server crash, Mr. Unz.
Examples:
Search word “all”
https://www.unz.com/comments/all/?s=all&searchsubmit=Search&ptype=all&commentsearch=only&commenter=FKA+Max
Search word “a”
https://www.unz.com/comments/all/?s=a&searchsubmit=Search&ptype=all&commentsearch=only&commenter=FKA+Max
Not related to the server crash, just a micro cosmetic improvement suggestion. On the categories bar on top of the Unz Review front/home page under “Columnists” there is a category/option titled “Your Selected Tweeters”, but it does not seem to list anybody or lead/link anywhere. Should or is “Your Selected Tweeters” perhaps supposed to be linked into this page https://www.unz.com/tweeters/ ?
I am glad the backup saved most of the content on the Unz Review, but unfortunately there were some really perceptive and insightful comments by some other commenters yesterday that, it looks like it to me, were lost/erased in the server crash.
I was lucky, only 2 comments of mine were lost, and I was able to remember what I had commented and was able to repost one of them already.
Despite this setback, I hope, everything will go smoothly from here on out.
As always, thank you very much, Mr. Unz, for everything you do (for free!) for all of us.
Yes, I’m well aware of that. We’re still in the process of rebuilding/restoring various subsystems, so there’s absolutely no need to provide long lists of non-functional elements until we’ve finished that process in another day or two.
I take the opportunity to note a bug that I first came across on my UR phone app two or three months ago and have now confirmed is not confined to the phone app version. It is this:
When I seek to email a comment I am required to enter the Capcha details correctly. Until a few months ago they presented themselves legibly and immediately so it was easy (especially when I discovered it was not case sensitive). Now the Capcha details are not visible – apparently off to the left hand side but not draggable or otherwise obtainable. So….
That is a vauable function now missing.
{ and I just assumed that nothing like this could possibly ever happen}
Sue the bastards.
(some of the best literary work in the English language in the form comments @unz.com was lost on that unfortunate day – forever)
Fortunately it was more of a shitpost than a real post so not much of value was lost (comments might be another matter). 🙂
Another small point. The time for editing seems to be truncated since the server crash and the reported time available to be inaccurate.
Unfortunately, we ran into new and totally unexpected server problems again this morning, and were forced to ONCE AGAIN restore our website system from last night’s backup.
I sincerely apologize for the many comments that were permanently lost from the last 18 hours or so.
Needless to say, I am very, VERY unhappy with the server company currently hosting our website.
I sincerely apologize for the many comments that were permanently lost from the last 18 hours or so.
Well, I’m sure you are sincere and no one blames you Mr Unz (again thanks for the site) — but bloody fucking hell! — it is annoying.
How about getting your own dedicated server, then installing software for backups every 15 minutes or hour?
https://www.ovh.com/us/dedicated-servers/storage/
There is a guy who specializes in scaling wordpress. I will break his name up: Pa tri ck G a r m an
One doubts whether anything in the way of deathless wisdom will have been lost.
Since it was lost in the second server crash, and Ron’s reply is even more pertinent, I will paste in my earlier comment and Ron’s reply (saved in my email because my comment was replied to).
***
Q: Did they give you an explanation of why this happened in an unrecoverable way and what steps they are taking so it won’t happen again?
A: Not really. Our cloud server crashed, so after trying various things, we asked the server people to just reboot it, at which time their discovered that the problem was that their parental server had crashed during its regular backup, and they needed to first reboot that. Within an hour or so they got it up again, and then they assumed they could also quickly reboot our “container” within that portion of the cloud, but repeated efforts failed. Finally, the next morning they sent up a simple note saying that the container’s data could not be recovered and we would have to reinstall everything.
Since the problem happened Friday afternoon, by the time the seriousness of the situation had become apparent, most of their tech people had left for the weekend. It’s possible that on Monday, they’ll figure out how to restore the container, but we obviously couldn’t wait that long.
Our regular website isn’t very large, maybe a few tens of GB total, and probably only a couple of GB in files change each week, which is all an incremental backup would see. I’d guess that most of the other containers in that cloud have far fewer changed files each week, so their regular incremental backup might only be 10 GB or so.
As it happens, I’ve been working on a major extension of this website, and as a final stage in that process had transferred 900 GB of new files to the server just the day before the crash. So either there was an astonishing coincidence, or the 900 GB of new files just choked their parent server during its incremental backup. That’s my guess at least.
Frankly, 900 GB of files really isn’t all that much these days, and it’s a very negative indicator if their cloud servers couldn’t handle it.
***
So now I can repeat the question.
Did they give you an explanation of why this happened in an unrecoverable way and what steps they are taking so it won’t happen again?
An additional question you can ask them is “how many other customers were affected by this crash?” — if it is a significant fraction of their users you can expect they will take steps to prevent recurrences.
Actually, as a mentioned yesterday in a (now vanished) comment of mine, we already do have a top-of-the-line cloud server configuration from the leading webhost company in the U.S., which is why I’ve found these problems so outrageous. This is just unacceptable.
However, 15-minute backups is just ridiculous since this is a small webzine, not a financial-transactions company. Overnight backups should be fine, and we’ve never previously had these problems.
How much time and money have you invested in this project Ron? I suspect it’s a pretty substantial amount of each. And I struggle to think of many analogues other than maybe John Tanton, although I suppose many of your columnists have also made similar sacrifices relative to their means. In any case, we are all in your debt including the vast multitudes of sheeple who haven’t figured it out yet.
RE cloud computing generally, isn’t it pitched as N+1 base with options for 2N+1? I’ve still never understood the corporate interest however as a loss such as yours would be devastating to a profit seeking venture. Offsite redundant backups make sense, but to put your livelihood on somebody else’s hardware seems crazy. For you, it makes sense since it’s all a time/money outlay and your time is your most precious commodity, but if I were a corporate IT guy I’d be shopping for co-lo space or investigating costs of rebuilding that old IBM 400 computer room into a modern DC.
“Once is happenstance. Twice is coincidence. The third time it’s enemy action.”
The UR is at twice.
I probably echo most folks here. No one blames you, Ron, and we’re glad to have UR, but, yeah, the server failure does screw up our readin’ ‘n’ commentin’ mojo.
Publish the photo of you bound and gagged by 7 maked ANTIFA thugs. The world NEEDS to see this!
My take if I may.
Know something about it.
Wrote a comment before, got lost, so, again in a nutshell.
I…suspect…..it is some upgrade/migration done within the host.
A project gone bad. They sometimes do.
The reality of Data Centres is NOT what people think they are.
That “Cloud” thing is, at the end of the day, digital electronics at the bottom. From then on it’s firmware, operating systems (hypervisors and SAN solutions) housing customer operating systems, housing business applications. All interconneted again by digital electronics and on top of it all that OSI model. Etc…….
After bragging how I know about it, the point was/is , it’s a complex system.
The complex system is managed as any business. Cutting costs first and foremost.
From (redundant) hardware, to skipping best practices in configurations, to, the most important, SUPPORT.
The support has been outsourced to India, China…Russia….whatever.
Good techs but with terrible communications/business skills.
And they, themselves, work as slave labor effectively. 24/7, 12 hour shifts, on call etc.
Constantly downsized themselves. Constantly overworked. Prone to make mistakes.
Constant shortcuts about processes, procedures….documentation. TESTING backups…testing business continuity/disaster recovery shortcuts.
Etc…
On top of it projects.
The most important element COST. Not best practices re reliability.
On top of it, WHO is the customer.
Small customers are at the end of that support. Important aren’t.
Etc…etc….
That’s the “Cloud” for you.
Not what you read in papers.
What engineers see at 3 in the morning trying to put it back to service.
The best is to have it on your own premises.
Yeah…right….too expensive. Hardware, software…licencing…..employees to support all that.
The drive is to “give it to somebody to take care of all that”.
Has been going for some time. The name of the game in IT.
Well….that’s why you get this.
And, you SHALL have problems.
The only way to deal with this paradigm is: accept that your Web site/business will go down and deal with it in preplanned and organized way.
For example, for unz.com, simple message from the provider in plain language.
And understanding that it could take a couple of days to have it back.
Simple.
Ah, yes, this one IS a rant. Apologize.
said:
“Publish the photo of you bound and gagged by 7 maked ANTIFA thugs. The world NEEDS to see this!”
ADL / SPLC thugs is more like it.
Dear Mr. Unz,
Thank you for running a great website to help disseminate news and views that are being suppressed or downplayed in the MSM, often for nefarious reasons.
By keeping the flame burning, you are doing a great service not just for actual readers, but for our larger polity. Even a speck of sunlight can dispel a an expanse of darkness.
On a technical note, having set up websites myself, I am in awe of your skill and dedication in setting up and running this system.
Actually, the true story of the second “crash” this morning is far, far worse than that, so I might as well tell it…
As all of you know, over the weekend, we’d been feverishly working to get the website installed and restored on a new server, after they told us Saturday morning that the data on the old one was unrecoverable. Finally, we succeeded, and late Sunday night almost everything was working fine. Pheh!
So this morning I was actually feeling pretty good about things, and thought we’d come through the crisis reasonably well. And since it was Monday morning, the regular server tech people came back, apologized for all the problems, and said that since our new server was working fine, whether they could just wipe the old, unrecoverable one. We very foolishly said Yes.
Naturally, the server people got a little mixed up and wiped the NEW server rather than the OLD old. Aaarrgh!!! So we had to spent several hours this morning fully reinstalling and reconfiguring everything for a second time on the wiped new server. And apologizing to everyone for a second batch of vaporized comments…
Someone upthread suggested that we set up our automatic backup to run every fifteen minutes rather than once per night. At first I thought that was ridiculous, but maybe an hourly backup wouldn’t be such a bad idea…
I’m seeing something that may or may not be of interest.
The edit-comment timeline of five minutes has been booting me out before five minutes. It’s as though the countdown timer runs out prematurely.
I have some sort of inexplicable cognitive barrier to seeing typos or miscoding in my comments till publishing the comment, so rely heavily on the five-minute edit window.
In the past perhaps two days I’ve lost not only edits, and the edit window (with as much as 4:35 remaining), but entire comments, though the latter may be more related to the larger comment problem.
FYI, and thank you for your work in making this space available to countless people you will never know.
These server techs: Home-grown, imported, or offshore?
Mr. Unz – thanks for your time and work and patience going into unz.com. I feel greatly indebted.
As so often – The Stones resonate in my head – “Fool to Cry”.
This sequence made me even laugh:
“(…) unrecoverable one. We very foolishly said Yes.
Naturally, the server people got a little mixed up and wiped the NEW server rather than the OLD old. ”
(Especially the word “naturally” I found quite – convincing… (= a well chosen expression)).
Something doesn’t seem right about that. One does not want to fall victim to conspiratorial thinking, but why on God’s green earth would the server tech people, who had just inexplicably destroyed your website, ask you if they could wipe anything of yours off their system? Where did they get the chutzpah? Wouldn’t you think that they of all people would know that data is their bread butter? It’s really the only commodity they trade in, and they should be guarding every last bit and byte of it with same ferocity with which ants guard their larvae. If these losers had any competence or professionalism at all, they would have done exactly the opposite. They would have said, “Mr. Unz, we’re terribly sorry that we don’t know how to recover your data, but we’re going to keep that old server just as it is, because the data is in there somewhere and we’re going to find out what happened to it or know the reason why.” This is like your mechanic ruining your ’57 Corvette and then saying, “Here, why don’t you just take this nice Oldsmobile off the lot. By the way, can I hang on to the Corvette so I can part it out?”
Absolutely unprofessional pieces of human garbage, these server people are.
Are we sure that it was not something nefarious? Some of the blog entries have gone against powerful interests and upset some very influential people. Could there be sophisticated sabotage activity against the website?
No one blames you, Mr. Unz, we are thankful for the site.
On the other hand, I cannot help but thinking about ever more elaborate conspiracy theories regarding these crashes. Am I totally alone in this?
Isn’t it ironic – – the classic Die vier Grundformen der Angst (=The four basic forms of fear) states that people suffer from a) Schizophrenia, b) Depression c) Paranoia d) Sadism.
This means – nobody ever walks alone – at least as long he suffers from one of those plagues.
I have used pair.com hosting since 1996. They may not have the most snazzy user interfaces and are not the cheapest, but they are a very serious company.
The fact that I’m paranoid is no evidence against the existence of a vast conspiracy working against me.
Ron’s explanation is the most reasonable. One server failure compounded by inattention and miscommunication. Not uncommon.
Keeping a server secure is a job for a specialist. You can do it yourself, but you won’t get much else done. Better to hire someone who takes care of many sites to invest his time in keeping up with the latest attacks.
I find myself wondering whether the server is linux or windows based.
Mr. Unz,
Some commenters mention possible bad actors. You support a website that discusses non-PC themes and dissident opinions. Put those two ideas together and expect that bad actors will act to frustrate your enterprise.
Another commenter mentioned sunshine.
Thank you for keeping a ray of sunshine.
I would have liked to push the agree-button – but – big BUT – – it din’t work – – – so!!…
That’s the prevalent viewpoint.
You only need to trust that “someone”.
And that “someone” is bunch of guys. Some of them, often, from abroad (India, China, Russia, Brazil), having, fromt their countries, real time access to the hardware/software.
At the same time, the virtual server is running on shared platform with other customers/players.Etc.
Keeping all that secure from INSIDE is, how to put it, a challenge.
I could argue that the best security would be simply putting all that on a physical server in own garage.
Having it all installed by a professional and then, later on, simply have him (a guy you personally know) do maintenance/support.
The only issue is Internet pipe, but, I am positive it can be done, especially in US.
All the rest, easy……
Bottom line: you outsource security you have no security. Especially today with tech turnover in I.T.
Forgive me, but
simply doesn’t go with
One or another is just not correct.
Leading webhosting companies have 24/7 support. No…way…that “server down” waits till Monday. Just now way.
Server goes down, within a minute tier 1 gets notified->they instantly do their checks/attempts to fix it-.within certain timeframe they escalate to tier 2 (often in the same office)->they, most of the time are capable of restoring the server, if they cant’ within certain timeframe they escalate to tier3. Tier 3, even from home, start working, Should they fail to restore the server, within a cetain timeframe they escalate to “incident manager”. He/she instantly wakes up/gets a team together. Now we are talking abut what’s called “severity 1” case-the team works until the case is resolved.
That’s the process in nutshell.
My point is that the work starts as soon as “server down” is seen and does not stop until the server is up and running again. People can get into working team and leave it, but the process starts now and lasts till all is O.K.
There is no waiting over the weekend.
For serious business that is.
Now, if a system is supported by a single engineer who is on a fishing trip over the weekend it’s another matter……..
If this happened to the web site in Soviet Union during Stalin people responsible for hosting this website would be already at Lubyanka writing as we speak the lists of names of people who made them do it.
Exactly. It’s really no big deal to setup two machines, one mirroring another in near real time.
reiner Tor is not alone, utu. Sabotage of this very intelligent Web site is a possibility I don’t want to think about.
Haha….you can, actually, find on the Web a very good instruction how to set the thing as:
3 second hand PCs, running VmWare VCenter with, as you say, a proper Web server mirror on top of it.
A decent UPS, fibre Internet and you are set.
The backup on another PC FreeNAS/OpenFiler/whatever solution.
All in garage, all under your own control.
Admit, really cheap, but, looking at that “server people not working over the weekend” looks more reliable.
Tells a lot…..
reiner Tor is not alone I know. Everybody, get in touch with your inner Chekist.
“You fucked up…you trusted us”
Video Link
{Exactly. It’s really no big deal to setup two machines, one mirroring another in near real time.}
Right.
Ron Unz is this technologically deficient dolt who just happened to come up with one of the best web sites and commenting software in the business, Entirely by accident.
Clearly it would never occur to him to setup his own server setup with two machines, one mirroring another in near real time. Good thing he’s got all these tech savvy posters giving him detailed advice about servers’n all.
Otherwise he’d be lost in the tech jungle.
Riiiiiiiiiight.
Well……this triggered ….giving opinion/advice.
Bold mine (my favorite quote here).
I get people being…….supportive….here (King and the court thing and such)
But being dismissive to posts by people actually knowing about all that feels…….stupid. Not helping the King, you know.
I mean, a couple of us here who know a bit about this could’ve kept our mouths shut and just watch what we’ve been watching, in real life, for quite some time. Like…”hehe….users……look at them…”.
It’s easy to stop offering any opinion/advice. We spent quite some time, effort and money getting that type of expertise; offering a bit of it for free…does it make sense actually?
Oh, come on. Not just “one of the best web sites and commenting software in the business” but simply the best – by far.
And of course Ron is perfectly capable of setting up his own server. Would likely need help here and there but ultimately this stuff is not a brain surgery. He just believed GoDaddy’s hype that letting others do it is by far easier, more reliable and more secure. Everyone, however smart, makes mistakes.
Like others, I am appalled by the story of techs going out for their weekend break, leaving a customer with a dead server they were supposed to keep running. Unbelievable.
Look, security matters. The simple point remains: physically owning the server is more secure than keeping it in the cloud and allowing scores of people access to it.
Leftovers from Team Schumer and Wasserman Schultz’s DNC. BIG stars on Capitol Hill, dontcha know?
pathalogizing political stances is a lefty game (“totalitarian personality” etc)
what’s worse, we pathalogize what used to be simply vice and/or character defects. now everyone just says they have “issues.”
The right lags the left in terms of credentialism but I think they can win easily in terms of mental health (just look at tumblr.)
Btw right now In America, half the population thinks they live under a Russian Occupational Government. The Right needs to troll the f*** out of them.
Based on
and my previous posts above I’ll give my advice/opinion again. The first got lost on the second server crash.
Call it….professional attitude.
Here I go:
Ron finds a trustworthy, experienced SMB/SME engineer.
An experienced, competent, mature engineer with considerable experience in designing, implementing and supporting small to medium businesses/offices.
The most important quality, though is the TRUST. The engineer should approach the job as a friend helping a friend.
I am positive that Ron can find such person with ease.
That person then, simply, does his job and makes a proposal.
A proper document.
Key points in the document: physical and logical architecture, cost, OPTIONS , expectations, and, the most important: processes and procedures with exact time frames. From disaster recovery, through business continuity to maintenance and monitoring.
Important to understand is: with more available expertise price goes down.
The starting point of work for that engineer is just one thing: for how long the site can be down.
I…suspect….that this site can be down for 3 days.
If…if that is the case, though, honestly….I’d stay where I am if I were Ron.
Simple.
Regardless of shown attitude/expertise the current provider will bring the site back within 3 days.
The only thing is a bit more of communication to “visitors” as, say “the site is experiencing technical difficulties and we are working on it”.
Now….if that’s not 3 days, well…..let’s say that with a proper setup the site should be back within 12 hours (don’t nitpick on this, a lot of variables here).
Now, even with the best “home setup” there is a case of “power down”. Even with a generator for truly paranoid in cases like that (power) network boxes in streets would most likely have issues….anyway…
And, of course, physical sabotage, arson and such. Taking care of that is next level. Can be done, sure, but adds cost/complexity.
So, again, taking those “3 days” and the rest of considerations….my advice would be: keep where you are (because the rest of providers will be the same regardless what they say…).
Just, maybe, have stronger and clearer lines of communications with provider (support techs-account manager).
Do you plan to fix certificate problems?
https://www.unz.com uses an invalid security certificate. The certificate is not trusted because the issuer certificate is unknown. The server might not be sending the appropriate intermediate certificates. An additional root certificate may need to be imported.
Error code: SEC_ERROR_UNKNOWN_ISSUER
I know. This is a reasonable thought.
Is your long quote from Theodore Dalrymple (aka Dr Anthony Daniels)?
Edit timer has a bug.
It works fine for the first edit, but if I try to re-edit (2nd edit), the post immediately kicks to moderation, even though the timer could still have 2,3,4 minutes.
Perhaps it’s because I once owned a 1960’s Jaguar but I’d go with two servers side by side with perhaps twice a day backup from the live one and a couple really good UPS systems.
“Bottom line: you outsource security you have no security.”
That’s what makes me laugh about the current trend amongst the rich for building bug-out bunkers for TEOTWAWKI. If I were the contractor building the place I’d make damn sure I was in there before the client.
Gates, I hear has one beneath every one of his homes.
The word Back-door doesn’t just apply to software..
Aren’t clouds supposed to be independent of any single point of failure, Ron?
Smells like someone’s overselling a regular dedicated server.
I think MySQL can do distributed databases. So the really reliable option would be to have a bunch of lightweight and identical DNS load balanced frontends that talk to an independent network of 2-3 backend DBs that store and synchronise all the dynamic content.
It’d be right pain in the arse for Ron tho 🙂
“maked”: masked and naked?
How come we can only edit posts twice before a message pops up that informs us we can no longer make edits and the editing time completely disappears? Hopefully this is a glitch that gets fixed.
I second this.
Mr. Unz,
you and your team are probably still repairing the site after the two server crashes and as you said you are already well aware of this problem, but since it is such a crucial function of the website, I just wanted to bring it to your attention one more time, since this function seems to continue to be unavailable, and the problem might have been forgotten about due to the stress, chaos and confusion caused by the server crashes:
Thank you very much.
Is there a bug with the handling of new comments? The numbers given on the home page seem right, but when I open the post there are more comments in blue. It is almost like my comments read status got frozen at a particular time earlier today.
Ditto.
It’s now November 5th and the site still isn’t working correctly. I still see replies to comments appearing earlier in the thread than the comments to which they are replying, like some sort of splintered time paradox. Also, just like res mentioned above, many comments in the threads are still highlighted as “new” even though I’ve read those comments already. It seems like the time-stamping of comments, on which all else depends, is all messed up.
And I’m still mighty pissed off about two rather lengthy comments of mine getting completely destroyed by the server fiasco. Is it not bad enough that I have to deal with the trolls, ignoramuses, and Steve Sailer’s “whimming”? I now have to worry about server meltdowns as well? It’s getting harder and harder to justify spending the hours it takes to write something decent, since there is such a significant risk of it never appearing at all, given everything else I have to do.
Thinking about this some more. Is it possible that there is different handling of time zones in two different modules of the software? That could explain both the unread comment behavior and the out of sequence comments. I can’t be sure, but it feels like the unread threshold for blue highlighting is not static but is moving. Perhaps an offset of GMT vs. my time zone?
I don’t know enough about the software to be sure, but I am guessing it tracks both comment time and approval time as well as previous page open time in order to distinguish new and old comments?
Is there a difference in behavior between edited comments and the originals? Why would there be a (substantial) difference between the home page number of new comments and the number of new comments when looking at a single post?
The problem with unread comments is making it difficult to follow long threads.
It is early in the morning 7 November. Everything is working normally as far as I can tell (I don’t do edits after I post). Some people see posts they have already read marked as new, but that is a normal hiccup from the restore from backup. You only have to leave the site and come back to have the new posts flag set to the current date.
Thank you Ron for providing this forum. Please forgive a few rude postings from people who expect something for nothing.
The search is working again!
Thank you very much, Mr. Unz.
Sorry, now it seems to be down again.
Same problem. Many more blue-background comments than the (correct) “New” figure.
I wonder whether it might be due to sometimes using the local time, sometimes universal time, when deciding what’s new and what’s not.
Yes, we’ve been absolutely 100% aware of the various date conventions problems for several days now.
The problem seems to be that the new server we installed the website on which we installed the website runs a somewhat different version of Linux/Apache/PHP than the previous one, and there appear to be some date-convention incompatibilities with the version of WordPress we use and/or our very heavily customized and complex code.
There’s absolutely no need for people to leave endless, redundant comments alerting us to something we certainly know and are trying to fix.
Since most of the problems resulting from this particular issue are merely inconvenient rather than fatal, we devoted our first efforts to getting the website generally operational before concentrating on this remaining bugs. I suppose we should have considered just entirely shutting down the entire website on the new server for a week or two until we finally resolved all these issues, but I suspect that would have generated even more unhappiness among the many people complaining above.
Frankly, since we have enough current problems still to fix, further comments complaining about this and other bugs serve absolutely no useful purpose except to irritate us. Once we get these problems fixed, we would welcome commenters bringing additional ones to our attention.
Was a report, not a complaint. But OK, got it, don’t report bugs that have been previously reported.
What I do, as a matter of course when it comes to comments of any length that I intend to post on this or any other site, is to compose them in a text-editor (or word-processor) and then, when finished, copy and paste my completed text into the comment field. This allows me to have a copy of all of the comments I submit, which can be highly useful in a number of cases– not least of which is the one under discussion here: a server crash that results in loss of comments. With the entire text of your comment saved locally (and, ideally, backed-up), you can simply copy and paste again and then re-submit it.
Similarly, if a comment of mine is rejected by a moderator or gets stuck in iSteve comment limbo for an inordinate of amount of time, I can modify and then re-submit the comment.
As for the other bugs you describe, I appreciate and share your frustration with them. I would remind you, however, that Mr. Unz provides this site for us to use and enjoy completely free-of-charge, not even burdening us with advertising. I would suggest, with that in mind, that you try to adopt a more gracious and forbearing attitude.
1. Two days’ worth of vanished comments isn’t great, but hardly significant in the big picture.
2. Easier alternative to Dissident’s recommendation: Just install an extension like Comment Save and/or Lazarus Form Recovery on Chrome (or their equivalents in Firefox).
{ I would remind you, however, that Mr. Unz provides this site for us to use and enjoy completely free-of-charge, not even burdening us with advertising. I would suggest, with that in mind, that you try to adopt a more gracious and forbearing attitude.}
Very well said.
I don’t understand the attitude of some users of UNZ.com.
It’s one of the cleanest, fastest, most elegant web sites out there.
If not THE.
Its commenting system has no equal that I have seen.
And I visit lots of sites.
Most sites have stopped taking comments altogether, because it’s too much work and headache for the site provider (…..to moderate and such), and then they have to deal with people complaining.
As you said, Ron provides this superb web site for people to offer their opinions (…and vent, and bitch,….) completely free of charge, _plus_ uncluttered by ads that slow down the web pages to a crawl. Like pretty much every site out there.
Like other people I lost a lot of comments: so what?
This not my life.
I visit the site, read what interests me, offer my opinions on some things, have occasional “exchanges of opinion” with some posters, then I’m done. Tomorrow is another day.
If my comments caused offense I apologize. As a user it can be hard to know how well understood the taxonomy of bugs is and what manifestations are known.
If multiple reports are that much of an issue it might be helpful to make a list (viewable by users, I suspect such a list already exists) of currently outstanding bugs and their manifestations then note when each is fixed.
Thanks for all of your work on the site, Ron!
Thanks Ron.
Keep your suggestions to yourself. You don’t know anything about my “attitude.” If you did, you would realize that your unsolicited advise was apropos of nothing whatsoever.
You don’t know anything about my “attitude.”
Well, I certainly do now. Thanks for that enlightenment.
I think I’ve finally managed to fix the various time/date problems, mostly related to the comments, and also we’ve gotten the Search system working again, though it’s still exhibiting some episodic anomalies that we need to figure out. That main time/date problem was responsible for most of the complaints above, including comment-ordering, new-comment-tinting, and comment-edit-window problems. So all of you should now feel free to let me know about any remaining bugs you’ve noticed on the new server.
Otherwise, thinking it over, I decided to switch from daily external backups of website data to hourly ones, which seems to be working fine. And we’re also currently looking at switching the hosting to an Amazon server, since they’re generally used by most of the largest corporations.
Given the horrible experience of the last two weeks days, I might as well explain some of the underlying software dynamics…
Almost everything any of you see or use on this website consists entirely of code that I wrote myself, and I can therefore usually fix problems pretty quickly. However, the “backend”—the underlying database and storage system—is based on WordPress, a very popular open-source webzine/content-management system. Unfortunately, WordPress contains a huge mass of complex, interwoven code, and WordPress problems often take a great deal of time and effort for me to untangle. It doesn’t help that many aspects of the WordPress software architecture are horribly designed for this sort of thing, having been originally developed for individual bloggers.
For example, the PHP/Linux configuration settings of the new server somehow conflicted with some of the WordPress settings, and that’s what caused the time/date problems, which probably took me something like 15 hours(!) to finally track down and fix.
Naturally, at times I’ve considered just getting rid of all the WordPress code, which would make many things much easier for me. However, WordPress actually does have a pretty good backend editing interface system allowing writers to easily produce articles, embed images, and that sort of thing, and it would take a tremendous effort for me to produce something as user-friendly for other people. Even more importantly, since WordPress is pretty much the industry-standard open-source website design system, a huge number of third-parties have produced very useful “plugin modules” for various features (e.g. Social Networking, Google Analytics, page caching) and our website uses quite a number of them. Recreating those specialized sub-systems myself would be a very difficult task.
So just think of WordPress as something like the old Microsoft Windows system of the 1990s—a gigantic, complex mess of code, difficult to work with, but providing an software ecosystem with such a huge number of useful third-party packages that it can’t be easily abandoned.
Well, as far as I can tell, you’re just some random nut who for some reason has chosen to hang around my website, spouting off on this and that.
Since you’ve now become so totally dissatisfied with its operation, I assume you’ve posted your last comment here. Bye-bye…
I am very satisfied on the other hand and consider the event as a minor hiccup that no user of the site should get too upset about and instead be grateful that it was fixed that quickly. On the other hand I find “whimming” by Sailer and blocking by Jayman and Karlin upsetting. I want them to keep the right to do it, however I would propose to have their censorship acknowledged by leaving a blank comment w/o a text. This was practiced by newspapers in Europe and still is by International NYT when the censorship did not allow them to print a particular article. A blank column would be left like here:
http://www.india.com/news/world/pakistani-edition-of-new-york-times-posts-blank-page-on-gay-kiss-in-china-cites-censorship-905719/
Statistics on banned commenters and rejected comments should be proudly displayed for each article. How many times Sailer exercise his whim in a given article?
Furthermore the euphemism “moderation” should be replaced with the word censorship on Sailer, Jayman and Karlin pages. So instead of
Leave a Reply – Comments are moderated by iSteve, at whim.
have this
Warning, the censorship zone. Leave a reply on your own risk – Comments are moderated by iSteve, at whim.
On the other hand bloggers who do not use censorship should have a banner : Free speech zone.
The banned commenters should be informed that they are banned the moment they type the first letter or the list of banned commenters should be made public.
Also a separate open thread could be made available to gather all (within a reason) rejected and banned comments that did not pass the censorship.
I am not saying that this must be or should be done. I am just describing the ideal world so we have more awareness about the world in which we live.
Thank you very much, Mr. Unz!
Apologies if any of my comments above were distracting and/or irritating during this stressful time.
Your analogy should be not to columns but to reader letters in newspapers. “Moderation” is the accepted term for this.
Not bad ideas.
Anyway Sailer hardly ever blocks things in my experience though approval is sometimes slow.
Mr. Unz or his mods do this for most of the site as you may have noticed.
Anyway Sailer hardly ever blocks things in my experience though approval is sometimes slow.
How do you know it is hardly ever? You know the fate of your comments only.
Mr. Unz or his mods do this for most of the site as you may have noticed.
No, I did not notice. By blank comments I mean that the comment is there but w/o a text. So you will know that X commented the article or responded to the comment of Y on day D and at time H but the text of the comment was censored for some reason.
By keeping blank comments the ordering of comment would be simplified and there would be no need for renumbering them which happens on iSteve which results in flagging many old comments as new but when you go to read them they are the comments you have already read.
Ron:
Good luck with WordPress. You are not the only one who has problems with it and screw those people who complain. They obviously never tried to run a site. WP is the mostly widely used software for blogs etc. And yes there are many plug ins but the more plug ins you use as far as I am concerned the greater the probability you’ll be hacked. I’ve built tons of sites and I try no longer to use WordPress.
Most of the major hacks that occur on sites are the ones that use WP. The joke is not if WP sites will get hacked but when. WP open source is both good and bad as you probably know. The real problem is the server. Most WP attacks occur through servers especially Apache. If there is any weakness anywhere on the server the exploiters seem to find it. You should be careful. You are probably a target for Antifa.
This is a great site and I appreciate the time you put into it. You should probably look into some type of back up mirror or redundancy that no one can access but you on a completely different system. One other point the intelligence agencies thank you for using the cloud. What I’ve been told about the cloud from people who set them up makes me scared to even think about using them.
I agree, this is a great site. Though I don’t like the absence of ads that much, because that means that Ron only provides the site as a hobby. It’s a great sacrifice on his part, and I’m constantly worried when he’ll get fed up with all this (or be pressured into stopping this) and gives it up. If it was a profit-making enterprise, then I’d feel safer.
I did say in my experience, did I not? Occasionally people on the rest of the site complain about Sailer but they too seem generally to complain about comments being approved slowly or out of order rather than deleted.
Here’s an example: https://www.unz.com/jderbyshire/i-dont-hate-muslims-im-prudent-about-them-theres-a-difference/#comment-2070302 .
Thanks. Now I know what you meant.
That would also mean that if the site was not making profits (which, let’s be honest, is hardly realistic with the content here) then he’d be even more incentivized to let it go.
Incidentally, that’s what recently happened with DNAinfo/Gothamist/Shanghaaist – the billionaire behind them got fed up them and abruptly shut them all down.
That’s an interesting perspective, but I think it’s probably mistaken. The key point you may be missing is that advertising would generate no significant revenue for a website such as this one.
Back several years ago when I served as Publisher of The American Conservative, I thought their advertising policies made absolutely no sense. They allocated something like 30% of the best “real estate” on their Home page to ads, with comparable space on all their other pages as well. But the ads only generated something like 5% of their operating expenses, so the trade-off seemed ridiculous. I argued with them a little but finally gave up. And since that time, I think ad rates have collapsed because Facebook has unlimited cheap and targeted inventory and has therefore captured so much of the market.
Why do all the political webzines require huge annual financial subsidies from their donors or owners? Because advertising doesn’t pay their bills, and I think the notion of an opinion website becoming a “profit-making enterprise” seems very unlikely. The only question is how large the annual operating loss would be.
Therefore, my strategy has been different. Instead of vainly seeking to maximize revenues, I think it makes much more sense to try to minimize expenses, and the highly-automated software I’ve written does just that. The bloggers write whatever they want on their own blogs, and the total time it takes me to fully “publish” any of the other articles or columns is only about 5m each. Also, we generally “moderate” comments with a very light hand.
The combination of all these factors means that despite the huge volume of articles, columns, posts, and comments we publish every, the whole thing only absorbs about 10% of my own time, which is very fortunate since I’m busy with my software work. Obviously, the situation has been very different during the last week or two because of the crash and recovery, but I’m hoping that’s mostly over.
As a result, I’m pretty sure that most somewhat comparable webzines require vastly more man-hours and dollars to operate, and these huge resource requirements lead to all sorts of other constraints on their coverage and activities.
Regarding financing, I suppose I could solicit donations or hold fund-raising drives, but that strikes me as unseemly and the sums would surely be much too small to be worth the necessary record-keeping. However, if any of you are especially eager to financially support this website, just go ahead and donate to some of the bloggers or other writers who allow that option, while saying nice things about the overall webzine.
That shuttered DNAInfo media group apparently had 15M visits per month compared to our small webzine, which averages around 1M or a little less.
However, just out of curiosity I checked, and they apparently had 115(!) staff employees to produce their content and run their various web publications. That total is quite a bit larger than ours, and I’d suspect that the payroll ratio was absolutely enormous. So overall, I’m very pleased with our cost-effectiveness.
Things are much easier when you’re not toeing a particular ideological line, and just presenting “interesting and controversial” perspectives, frequently republished from elsewhere.
From the beginning, I always envisioned this website as primarily being a content-distribution channel and commenting platform, which allows for much greater operational efficiency.
{…..and they apparently had 115(!) staff employees }
According to NYT, Joe Ricketts shut down the sites after staff voted to unionize.
He denies that was the reason, but of course he would.
Probably Ricketts was already tired of throwing money into the maw and the unionizing vote finally did it for him.
[DNAinfo and Gothamist Are Shut Down After Vote to Unionize]
https://www.nytimes.com/2017/11/02/nyregion/dnainfo-gothamist-shutting-down.html
Actually, I might as well provide a little more background about why I originally launched this website, and why I’m therefore likely to keep it going in the future. Many of you may already know this history, but others may not.
During the 2000s, I was totally preoccupied with my software work, but from about 2006 onward, I began financially supporting a wide variety of “controversial” websites, writers, and thinkers. The bulk of my funding went to The American Conservative, whose budget was about $1M per year, which seemed like a great deal of money, especially since most of it was coming straight out of my own pocket. I thought TAC’s operations were rather inefficient, but I was too busy with my own work to get actively involved and try to change things.
Around 2011, my big software project wound down, and over the next couple of years, I started to do quite a lot of my own writing, including a number of major articles that were very well received. Naturally, I published almost everything in TAC, my own magazine, and that was extremely convenient.
Then, in 2013 I was suddenly purged for “incorrectness,” and faced the dilemma of lacking any easy outlet for my own articles. I’m not a blogger and I typically produce long, fairly infrequent pieces, so there was no way my writing could maintain ongoing interest on its own website.
But then it struck me that so many interesting writers, including those I’d been supporting for many years, had been gradually purged from various publications. Thus, there was probably now a critical mass of such individuals available to populate a new webzine, and the ongoing content they generated would attract a natural potential audience for my own future articles. So I started putting things together, including some new ideas I’d had about content presentation and operational design, and after getting sidetracked several months into a political campaign, the website was fully ready by mid-2014.
However, things often turn out differently than expected. Over the last four years, I’ve been so preoccupied with various software projects, as well as some political campaigns, that there have only been a few months during that entire period in which I had time for doing any serious writing. But although the original purpose of this website hasn’t really been fulfilled, I’ve been very pleased with how well it’s done otherwise, and the financial expense has been almost negligible compared to many other things I’ve supported. Also, I do think I’m now getting very close to being able to return to my own research and writing, though I’ve been saying exactly the same thing for most of the last few years.
Then, in 2013 I was suddenly purged for “incorrectness,”
What were they thinking? Getting rid of their benefactor?
Ron, thanks for the backgrounder. I raised a question about successorship recently in a casual comment, and I think you sort of answered it here.
BTW-UR is a great resource. The rhetorical public space you’ve opened up here simply doesn’t exist in my area, with a few very rare exceptions.
Perhaps a dumb question from a coding know-nothing, but is your commenting system sellable? Talk about revenue. It’s the best one I’ve ever come across.
Ron,
Thank you for your contribution and dedication. Your website has been very informative and continues to provide a real look behind the news and at many under-publicized international perspectives.
Best regards
Mr. Unz, please PLEASE keep this ad-free scheme, you’ve no idea how off-putting other ad-heavy sites can be. Especially the tracking-moving ones. I think it must be because I enjoy reading books, and I can’t be the only one here who feels this way. I wouldn’t be surprised if this promotes more thoughtful comments. It is anyway a somewhat quixotic enterprise, so the expense, well, in for a penny, in for a pound!
Thanks for the explanation here and in your subsequent comment.
Yes, a site without ads is vastly preferable to one full of them, so if it brings no benefits at all, then I do agree it’s much better that way.
It’s interesting that this hobby-site of yours (for the purpose of publishing your few and far between – though generally high quality – articles) is perhaps the best political website in the world, both in terms of content and in terms of technical things like the commenting system (which is by far the best commenting system ever). Well done, Sir!
Thanks again, Mr. Unz!
Eternally grateful to you for your sterling efforts in providing a counterpoint to all the co-opted, Narrative-pushing, news-suppressing, comment censoring MSM sites out there.
What is particularly worrying is the vicious and thorough censoring of perfectly “respectable” political viewpoints and thoughts, e.g. in the area of immigration, not only by the usual suspects (e.g. NYT), but through a broad range of publications and radio/television outlets.
It is now clear that a wide range of media websites are spending serious money on HUMAN censors whose job it is to suppress ruthlessly any public awareness that the Republican/Democrat preference for mass ILLEGAL immigration EXISTS AS A POLICY ISSUE that can and should be rationally discussed and addressed soberly and objectively.
Instead, we see ideological shrieks, smokescreens and virtue-signaling, all carefully arranged as panem et circenses to distract not just the multitudes but many individuals with at least an inchoate interest in truth and objectivity. Meanwhile, disinformation and trivia flourish unrestricted. The Swamp lives.
I don’t know Ron, about the ad thing; I for one am really looking forward to being able to purchase an ap for Asian girls that signals to white guys within a one-mile radius that they are available, or a hand-held device that plays country music, really, really loudly when one comes within sight distance of a group of angry knee-grows, or along that line, maybe hollowed-out replicas of old books that contain one’s money and valuables.
Not a bug. The thumbnail picture for Mr. Thompson’s article “Small Incident in London, Not Many Dead” seems to be broken, as can be seen here (it is the 6th ranked article on that page): https://www.unz.com/author/james-thompson/?ItemSorts=comment_count
I’ve noticed over the last several days that the Search system would be sporadically unavailable, but for the last 24+ hours it seems to have been stable.
Also, Mr. Unz, do you feel the TAC controversy was accurately portrayed in the following article? I am always somewhat skeptical of the National Review, but they seem to have giving you a relatively fair hearing in the piece.
The American Conservative, Unfused?
There’s a rift between the publisher and the editor, over editorial decisions and finances.
http://www.nationalreview.com/article/354893/american-conservative-unfused-betsy-woodruff
Thank you very much.
According to the Alexa siteinfo the Unz Review is currently, globally, ranked higher than TAC:
https://www.alexa.com/siteinfo/unz.com 32,160th global rank
vs.
https://www.alexa.com/siteinfo/theamericanconservative.com 35,208th global rank
There seems to have been a noticeable drop in their web traffic/readership around the time they fired Mr. Giraldi.
they got money from their new overlords and wanted Mr. Unz gone.
OT but related: Today I notice that both Bing and Google search engines no longer put UNZ.COM at or even near the top when I search for “UNZ”. Things like UNZ.ORG come up, as well as other things related to Ron Unz, but the Unz Review is invisible or very hard to find in the results.
This is a big and sudden change. I have routinely entered “UNZ” into either Google or Bing as a lazy and quick way of getting to UNZ.COM, and the Unz Review has always been the top result, but today the results have changed completely.
One suspects nefarious activity on the part of those who run Google and Bing. However, for both services to do the same thing at the same time seems less likely. Who knows why this is happening?
First result for me. Crash-related?
Possible Bug:
I embedded 2 jpgs from Wiki in my comment, but they disappeared: there is just white space in the published comment in their place.
Comment #413 Avery says: November 9, 2017 at 8:45 pm GMT • 200 Words.
https://www.unz.com/tsaker/do-you-think-his-assessment-is-accurate/#comments
Back to normal now. This morning, Unz Review was invisible on Google and Bing, here at least.
re: https://www.unz.com/tsaker/do-you-think-his-assessment-is-accurate/#comment-2071452
Avery,
I have had the same problem before, but it is not a bug, you just uploaded the wrong version of the image.
You uploaded the “en.wikipedia.org” version of the image:
but in order for the image to show up you need to upload the “upload.wikimedia.org” version of the image:
The former is technically a web page, I believe, and the latter is an actual image file, if I am not mistaken. I am not an expert, I just found this out through trial and error.
In order to find the latter, correct version of the photo google “battle of stalingrad dead bodies wikipedia” and search under “Images”, then click on the photo and choose the “View image” option, and you will get the version/URL of the image that uploads properly here on the Unz Review: https://www.google.com/search?q=battle+of+stalingrad+dead+bodies+wikipedia&source=lnms&tbm=isch&sa=X&ved=0ahUKEwivnaWmlL_XAhWhqlQKHQB7CSQQ_AUICigB&biw=1024&bih=635#imgrc=kqNW3SmjNhnaxM:
Hope that helps.
P.s.: Typo in my above comment: … they seem to have *given* you … – https://www.unz.com/announcement/server-crash/#comment-2074360
Thanks for the tip [FKA Max].
Good to know.
Avery.
OT: Am thinking vaguely of an “Unzian Society” for my area. Bimonthly or quarterly dinners, good fellowship, brief talks by locals on articles that interest them, and so on. Seems to me Unz Review is doing a great job restoring free speech as it once existed before speech was channeled, suppressed, programmed, “coded”, bureaucratized, distorted, borked, or whatever in hell happened to it.
Fantastic idea. Would love to do the same in my area. The only issue is numbers, are there sufficient Unzers in our respective areas to make this a feasible proposition?
Another issue we’d probably have to contend with is keeping the infiltrators out.
A few minutes ago the page failed to load for me. “failed to reach server at https://www.unz.com” or suchlike.
(I’m using firefox on ubuntu if that matters).
geokat62, thanks. Try “selling” Unz Review to people you think are looking for fresh intellectual mojo. That’ll let you know if there’s a market for an Unz Society in your area. Freely acknowledge the obvious here. Some of the posts and comments that Ron hosts here go into some pretty dark places.
Just say that, before introducing someone to a good article by Anatoly or an acerbic observation by Steve. Let your prospect decide for himself whether Unzie is his cuppa. Light touch and a smile, buddy. Good luck.
Another issue we’d probably have to contend with is keeping the infiltrators out.
Maybe we could have a secret handshake or password as a security feature.
Great idea… the password(s) I’d suggest is “Thank God I’m not a PEAbrain.”
Considering the vicious hatred of women expressed by commenters at Unz every day, while any sort of defense of women will not pass moderation, you could call your club G.R.O.S.S. (Get Rid Of Slimy GirlS), and your password could be “Our Hero Rodney Alcala” — I wouldn’t be surprised at all if he were posting here.
Whoever, am I right in thinking you’re not interested in being the keynote speaker?:)
He’ll be home cataloging his Calvin & Hobbes books 😉
Reporting bug:
The commenting software stopped numbering comments on the latest Godfree Roberts piece after comment #303:
https://www.unz.com/article/mao-reconsidered-part-two-whose-famine/#comment-2086024
Just noticed:
On the same article the sidebar is (re-)moved from its original place/position at the top of the page and now appears towards the bottom of the page underneath the commenting box and the “Past Classics” section: https://www.unz.com/article/mao-reconsidered-part-two-whose-famine/#respond
Fixed.
Thank you very much, Mr. Unz!
Word of advice, whatever handshake you select make sure it involves a firm-wrist. I mean, you are “conservatives”, people are going to jump to conculsions anyway…
you are “conservatives”,
I am going to let this little piece of name calling slide because I give special consideration to any commenter that can come up with something as hilarious as Knee-grow.
If I leave a comment-reply from a commenter’s page it gets attached incorrectly as a reply to a more recent comment.