I'm appalled at the way some people here receive an honest postmortem of a human fuck-up. The top 3 comments, as I write this, can be summarized as "no, it's your fault and you're stupid for making the fault".
This is not good! We don't want to scare people into writing less of these. We want to encourage people to write more of them. An MBA style "due to a human error, we lost a day of your data, we're tremendously sorry, we're doing everything in our power yadayada" isn't going to help anybody.
Yes, there's all kinds of things they could have done to prevent this from happening. Yes, some of the things they did (not) do were clearly mistakes that a seasoned DBA or sysadmin would not make. Possibly they aren't seasoned DBAs or sysadmins. Or they are but they still made a mistake.
This stuff happens. It sucks, but it still does. Get over yourselves and wish these people some luck.
The software sector needs a bit of aviation safety culture: 50 years ago the conclusion "pilot error" as the main cause was virtually banned from accident investigation. The new mindset is that any system or procedure where a single human error can cause an incident is a broken system. So the blame isn't on the human pressing the button, the problem is the button or procedure design being unsuitable. The result was a huge improvement in safety across the whole industry.
In software there is still a certain arrogance of quickly calling the user (or other software professional) stupid, thinking it can't happen to you. But in reality given enough time, everyone makes at least one stupid mistake, it's how humans work.
It is not only that but also realizing that there is never a single cause to an accident or incident.
Even when it was a suicidal pilot flying the plane into a mountain on purpose. Someone had to supervise him (there are two crew members in the cockpit for a reason), someone gave him a medical, there is automation in the cockpit that could have at least caused an alarm, etc.
So even when the accident is ultimately caused by a pilot's actions, there is always a chain of events where if any of the segments were broken the accident wouldn't have happened.
While we can't prevent a bonkers pilot from crashing a plane, we could perhaps prevent a bonkers crew member from flying the plane in the first place.
Aka the Swiss cheese model. You don't want to let the holes to align.
This approach is widely used in accident investigations and not only in aviation. Most industrial accidents are investigated like this, trying to understand the entire chain of events in order that processes could be improved and the problem prevented in the future.
Oh and there is one more key part in aviation that isn't elsewhere. The goal of an accident or incident investigation IS NOT TO APPORTION BLAME. It is to learn from it. That's why pilots in airlines with a healthy safety culture are encouraged to report problems, unsafe practices, etc. and this is used to fix the process instead of firing people. Once you start to play the blame game, people won't report problems - and you are flying blind into a disaster sooner or later.
That reminds me of the time during the rise of the PC when windows would do something wrong, from a confusing interface all the way up to a blue screen of death.
What happened is that users started blaming themselves for what was going wrong, or start thinking they needed a new PC because problems would become more frequent.
From the perspective of a software guy, it was obvious that windows was the culprit but people would assign blame elsewhere and frequently point the finger at themselves.
so yes - an FAA investigation would end up unraveling the nonsense and point to windows.
That said, aviation level of safety is reliable and dependable and few single points of failure and... there are no private kit jets darnit!
There is a continuum from nothing changes & everything works to everything changes & nothing works. You have to choose the appropriate place on the dial for the task. Sounds like this is a one-man band.
It's good to have a post mortem. But this was not actually a post mortem. They still don't know how it could happen. Essentially, how can they write "We’re too tired to figure it out right now." and right after attempt to answer "What have we learned? Why won’t this happen again?" Well obviously you have not learned the key lesson yet since you don't know what it is! And how can you even dream of claiming to guarantee that it won't happen again before you know the root cause?
Get some sleep, do a thorough investigation, and the results of that are the post mortem that we would like published and where you learn from.
Publishing some premature thoughts without actual insight is not helping anybody. It will just invite the hate that you are seeing in this thread.
> I'm appalled at the way some people here receive an honest postmortem of a human fuck-up. The top 3 comments, as I write this, can be summarized as "no, it's your fault and you're stupid for making the fault".
It seems that people annoyed mostly by "complexity gremlins". They are so annoyed, that they miss previous sentence "we’re too tired to figure it out right now." Guys fucked up their system, they restored it the best they could, they tried to figure out what happened, but failed. So they decided to do PR right now, to explain what they know, and to continue the investigation later.
But people see just "complexity gremlins". The lesson learned is do not try any humor in a postmortem. Be as serious, grave, and dull as you can.
For me, this is an example of DevOps being carried too far.
What is to stop developers for checking into Github "drop database; drop table; alter index; create table; create database; alter permission;"? They are automating environment builds and so that is more efficient right? In my career, I have seen a Fortune 100 company's core system down and out for a week because of hubris like this. In large companies, data flows downstream from a core system. When you have to restore from backup, that cascades into restores in all the child systems.
Similarly, I once had to convince a Microsoft Evangelist who was hired into my company, not to redeploy our production database, every-time we had a production deployment. He was a pure developer and did not see any problems of dropping the database, recreating the database, and re-inserting all the data. I argued that a) this would take 10+ hours b) the production database has data going back many years and that the schema/keys/rules/triggers have evolved during that time-- meaning that many of the inserts would fail because they didn't meet the current schema. He was unconvinced but luckily my bosses overruled him.
My bosses were business types and understood accounting. In accounting, once you "post" a transaction to the ledger that becomes permanent. If you need to correct that transaction, then you create a new one that "credits" or corrects the entry. You don't take out the eraser.
Culturally speaking we like to pat people on their back when they do something stupid and comfort them. But most of the time this isn’t productive because it doesn’t instil the requisite fear required when working out what decision to make.
What happens is we have growing complacency and disassociation from consequences.
Do you press the button on something potentially destructive because your are confident it is ok through analysis, good design and testing or confidence it is ok through trite complacency?
The industry is mostly the latter and it has to stop. And the first thing is calling bad processes, bad software and stupidity out for what it is.
Honestly these guys did good but most will try and hide this sort of fuck up or explain it away with weasel words.
You should have zero fear instilled when pressing any button. The system or process has failed if a single person pressing a button can bring something down unintended. Fix the system/process, don’t “instill fear” onto the person, it’s toxic, plus now you have to make sure any new person on boarded has “fear instilled”, and that’s just hugely unproductive
> Computers are just too complex and there are days when the complexity gremlins win.
I'm sorry for your data loss, but this is a false and dangerous conclusion to make. You can avoid this problem.
There are good suggestions in this thread, but I suggest you use Postgres's permission system to REVOKE DROP action on production except for a very special user that can only be logged in by a human, never a script.
And NEVER run your scripts or application servers as a superuser. This is a dangerous antipattern embraced by many and ORM and library. Grant CREATE and DROP to non-super users.
You have to put a lot of thought into protecting and backing up production databases, and backups are not good enough without regular testing of recovery.
I have been running Postgres in production supporting $millions in business for years. Here's how it's set up. These days I use RDS in AWS, but the same is doable anywhere.
First, the primary server is configured to send write ahead logs (WAL) to a secondary server. What this means is that before a transaction completes on the master, she slave has written it too. This is a hot spare in case something happens to the master.
Secondly, WAL logs will happily contain a DROP DATABASE in them, they're just the transaction log, and don't prevent bad mistakes, so I also send the WAL logs to backup storage via WAL-E. In the tale of horror in the linked article, I'd be able to recover the DB by restoring from the last backup, and applying the WAL delta. If the WAL contains a "drop database", then some manual intervention is required to only play them back up to the statement before that drop.
Third is a question of access control for developers. Absolutely nobody should have write credentials for a prod DB except for the prod services. If a developer needs to work with data to develop something, I have all these wonderful DB backups lying around, so I bring up a new DB from the backups, giving the developer a sandbox to play in, and also testing my recovery procedure, double-win. Now, there are emergencies where this rule is broken, but it's an anomalous situation handled on a case by case basis, and I only let people who know what they're doing touch that live prod DB.
> after a couple of glasses of red wine, we deleted the production database by accident
> It’s tempting to blame the disaster on the couple of glasses of red wine. However, the function that wiped the database was written whilst sober.
It was _written_ then, but you're still admitting to the world that your employees do work on production systems after they've been drinking. Since they were working so late, one might think this was emergency work, but it says "doing some late evening coding". I think this really highlights the need to separate work time from leisure time.
No. Your systems and processes should protect you from doing something stupid, because we’ve all done stupid things. Most stupid things are done whilst sober.
In this case there were like 10 relatively easy things that could have prevented this. Your ability to mentally compile and evaluate your code before you hit enter is not a reliable way to protect your production systems.
Coding after drinking is probably not a good idea of course, but “think better” is not the right takeaway from this.
The whole piece has a slightly annoying flippant tone to it. We were drunk! Computers just.. do this stuff sometimes! Better to sound contrite and boring in such a situation IMO.
Also I agree with other comments: doing some work after a glass or two should be fine because you should have other defences in place. “Not being drunk” shouldn’t be the only protection you have against disaster.
I am not a drinker myself (drink 1-3 times a year), but in the past I have coded while slightly buzzed on a few occasions. I could not believe the level of focus I had. I never investigated it further, but I'm pretty sure the effects of alcohol on our coding abilities is not nearly as bad as it affects our motor skills. Imo, fatigue is far worse.
I had a narrow escape once doing something fancy with migrations.
We had several MySQL string columns as long text type in our database but they should have been varchar(255) or so. So I was assigned to convert these columns to their appropriate size.
Being the good developer I was, I decided to download a snapshot of the prod database locally and checked the maximum string length we had for each column via a script. Using this script it made a migration query that would alter column types to match their maximum used length keeping the minimum length as varchar (255).
I tested that migration and everything looked good, it passed code review and was run on prod. Soon after we start getting complaints from users that their old email texts have been truncated. I then realize the stupidity of the whole thing, the local dump of production database always wiped out many columns clean for privacy like the email body column. So the script thought it had max length of 0 and decided to convert the column to varchar(255).
I realize the whole thing may look incredibly stupid, that's only because the naming for db columns was in a foreign european language so I didn't know even know the semantics of each column.
Thankfully my seniors managed to restore that column and took the responsibility themselves since they had passed the review.
We still did fix those unusually large columns but this time by simple duplicate alter queries for each of those columns instead of using fancy scripts.
I think a valuable lesson was learned that day to not rely on hacky scripts just to reduce some duplicate code.
I now prefer clarity and explicitness when writing such scripts instead of trying to be too clever and automating everything.
Just my 2 cents. I run a small software business that involves a few moderately-sized databases.
The day I moved from a fully managed hosting to a Linux VPS, I have crontabbed a script like this to run several times a day:
for db in `mysql [...] | grep [...]`
mysqldump [...] > $db.sql
git commit -a -m "Automatic backup"
git push [backup server #1]
git push [backup server #2]
git push [backup server #3]
The remote git repos are configured with denyNonFastForwards and denyDeletes, so regardless of what happens to the server, I have a full history of what happened to the databases, and can reliably go back in time.
I also have a single-entry-point script that turns a blank Linux VM into a production/staging server. If your business is more than a hobby project and you're not doing something similar, you are sitting on a ticking time bomb.
Anyone reading the above: please don't do this. Git is not made for database backups, use a real backup solution like WAL archiving or dump it into restic/borg. Your git repo will balloon at an astronomical rate, and I can't imagine why anyone would diff database backups like this.
Happens to all of us. Once I required logs from the server. The log file was a few gigs and still in use. so I carefully duplicated it, grepped just the lines I needed into another file and downloaded the smaller file.
During this operation, the server ran out of memory—presumably because of all the files I'd created—and before I know it I'd managed to crash 3 services and corrupted the database—which was also on this host—on my first day. All while everyone else in the company was asleep :)
Over the next few hours, I brought the site back online by piecing commands together from the `.bash_history` file.
Same happened to me many years ago. QA dropped the prod db. It's been many years but if I recall, I believe in the dropdown menu of the MongoDB browser, exit & drop database were next to each other...Spent a whole night replaying the oplog.
No one owned up to it, but had a pretty good idea who it was.
>Note that host is hardcoded to localhost. This means it should never connect to any machine other than the developer machine. We’re too tired to figure it out right now. The gremlins won this time.
Obviously, somehow the script ran on the database host.
some practices I've followed in the past to keep this kind of thing from happening:
* A script that deletes all the data can never be deployed to production.
* scripts that alter the DB rename tables/columns rather than dropping them (you write a matching rollback script ), for at least one schema upgrade cycle. you can always restore from backups, but this can make rollbacks quick when you spot a problem at deployment time.
* the number of people with access to the database in prod is severely restricted. I suppose this is obvious, so I'm curious how the particular chain of events in TFA happened.
I have a little metadata table in production that has a field that says “this is a production database”. The delete-everything script reads that flag via a SQL query that will error out of it’s set in the same transaction as the deletion. To prevent the flag from getting cleared in production, the production software stack will refuse to run if the “production” flag is not set.
One aspect that can help with this is separate roles/accounts for dangerous privileges.
I.e. if Alice is your senior DBA who would have full access to everything including deleting the main production database, then it does not mean that the user 'alice' should have the permission to execute 'drop database production' - if that needs to be done, she can temporarily escalate the permissions to do that (e.g. a separate account, or separate role added to the account and removed afterwards, etc).
Arguably, if your DB structure changes generally are deployed with some automated tools, then the everyday permissions of senior DBA/developer accounts in the production environment(s) should be read-only for diagnostics. If you need a structural change, make a migration and deploy it properly; if you need an urgent ad-hoc fix to data for some reason (which you hopefully shouldn't need to do very often), then do that temporary privilege elevation thing; perhaps it's just "symbolic" but it can't be done accidentally.
Quote: "Note that host is hardcoded to localhost. This means it should never connect to any machine other than the developer machine. Also: of course we use different passwords and users for development and production. We’re too tired to figure it out right now.
The gremlins won this time."
No they didn't. Instead one of your gremlins ran this function directly on the production machine. This isn't rocket science, just the common sense conclusion. Now it would be a good time to check those auditing logs / access logs you're suppose to have them enabled on said production machine.
> Instead one of your gremlins ran this function directly on the production machine.
Exactly my first hypothesis too. But then keepthescore claims,
> of course we use different passwords and users for development and production.
How would this hypothesis explain that?
John Watson: "I deduce that someone changed the dev config source so that it uses the production config values."
Sherlock Holmes: "My dear Watson, while that is sensible, it seems to me that the balance of probability leans towards their production instance also having the development access credentials."
Just my way of saying, I think this case isn't as shut and closed as most comments (including parent) imply. I personally find the /etc/host mapping a likelier hypothesis but even that can't explain how different credentials failed to prevent this. Without more details coming from a proper investigation, we are just piling assumptions on top of assumptions. We are making bricks, without enough clay, as Holmes would say.
If managed DBs on DigitalOcean are anything like those on AWS, you can not run commands directly on them since SSH is prohibited. EDIT: there's also the deal with different credentials for dev and prod envs.
That it happened meant that there were many things wrong with the architecture, and summing up the problem to “these things happen” is irresponsible, most importantly your response to a critical failure needs to be in the mindset of figuring out how you would have prevented the error without knowing it was going to happen and doing so in several redundant ways.
Fixing the specific bug does almost nothing for your future reliability.
I'm not sure I follow your point - I think you'll find the same attitude towards complexity by operators of military rockets and nuclear power plants. If you look at postmortems/root-cause analyses/accident reports from those fields, you'll generally find that a dozen things are going wrong at any given time, they're just not going wrong enough to cause problems.
I feel that computers make it easier for this danger to be more indirect, however. The examples you give are physical, and even the youngest of child would likely recognise they are not regular items. A production database, meanwhile, is visually identical to a test database, if measures are not made to make it distinct. Adults though we may be, we're human, and humans can make really daft mistakes without the right context to avoid them
One explanation for the author feeling that way is that the system is has too much automation. Being in a situation where you take on more responsibilities of the system at a shallower level leads to less industry expertise. This, as it turns out, places the security of the system in a precarious position.
These is pretty common, as devs tool belts have grown longer over time.
I think at some point we will stop automating or reverse some of the automation.
Are you sure it was the production database that was affected?
If you are not sure how a hard coded script that was targeting localhost affected a production database, how do you know you were even viewing the production database as the one dropped?
Maybe you were simply connected to the wrong database server?
I’ve done that many times - where I had an initial “oh no“ moment and then realized I was just looking at the wrong thing, and everything was ok.
I’ve also accidentally deployed a client website with the wrong connection string and it was quite confusing.
In an even more extreme case: I had been deploying a serverless stack to the entirely wrong aws account - I thought I was using an aws named profile and I was actually using the default (which changed when I got a new desktop system). I.e. aws cli uses —profile flag, but serverless cli uses —aws-profile flag. (Thankfully this all happened during development.)
I now have deleted default profiles from my aws config.
The lack of the seriousness/professionalism of the postmortem seemed odd to me too. So, okay, what is this site?
> KeepTheScore is an online software for scorekeeping. Create your own scoreboard for up to 150 players and start tracking points. It's mostly free and requires no user account.
> Sat Sep 5, 2020, Running Keepthescore.co costs around 171 USD each month, whilst the revenue is close to zero (we do make a little money by building custom scoreboards now and then). This is an unsustainable situation which needs to be fixed – we hope this is understandable! To put it another way: Keepthescore.co needs to start making money to continue to exist.
This post is embarrassing. "yeah we were drinking and accidentally nuked the prod DB. Not sure why. Shit happens!" Who would read this and think they should trust this company? Any number of protections could have been taken to prevent this and production access in any state other than fully alert and attentive shouldn't happen unless it is absolutely necessary for emergency reasons
Yeah why should I treat anything this company does with any level of seriousness? Why should anyone?
It's lucky it's just some online scoreboard because I'm sure as shit this stuff has happened before with more critical systems and it scares the hell out of me that engineers are fine blaming "gremlins" instead of taking responsibility for their own incompetence.
Somewhat similar story from many years ago. Was in ~/somedirectory, wanted to clear the contents, ran `rm -rf *`. Turns out somewhere in between I had done a `cd ..`, but I thought I was still in the child directory. Fastest Ctrl+C ever once I saw some permission errors, but most of the home directory was wiped in that second or two.
Didn't have a backup of it unfortunately, though thankfully there wasn't anything too critical in there. Mostly just lost a bunch of utility scripts and dotfiles. I feel like it's beneficial in the long run for everyone to make a mistake like this once early on in their career.
Ah man, these things happen. One of our developers - very new to elastic - was asked to modify some indexes. Folks were a bit too busy to help or heading out on holiday. One stack overflow answer later... delete and recreate it... and she was off to the races. When the test was tried, it looked like things still worked. A quick script did the same to stage and prod, in both data centers. Turns out that is not a great way to go about it. It deleted the documents. We got lucky, as we still had not killed off the system we were migrating off of and it only took three days of turn and burn to get the data back on the system.
So many lessons learned that day. I trust her with the master keys at this point, as nobody is more careful with production than her now. :)
At my job, the company computers are configured to send “localhost” to the local company DNS servers, of which they happily reply with the IP address of the last machine that’s gotten a DHCP lease with the hostname “localhost”. Which happens often. Needless to say, our IT dept isn’t the best.
Yeah, imagine if a bridge engineer said the same thing: "You aren't an engineer until your bridge collapses. Congrats!" I am starting to hate tech culture. Nobody cares about correctness and discipline. Mention "math" and everybody spreads like cockroaches.
Definitely, the likelihood of these things happens goes way up alongside the amount of concurrent tasks, meetings, or other forms of distraction. I’ve seen my fair share of production administration as a side task.
A likely culprit. Having worked on a bunch of early-stage products where best practices are a distant future dream, I’ve developed a few “seatbelt” habits I use to avoid these kinds of things. One of them is to always use a random high-number local port if I’m tunneling to a production service.
Another is to change my terminal theme to a red background before connecting to anything in production...never want to click the ‘psql ...’ tab, run “truncate table app_user cascade” and realize afterwards it was a lingering connection to production...
Things like these happen, and we should be compassionate towards them.
Often small changes to the structure drastically reduce probability of stuff like this happening.
Eg. we use docker to setup test and dev databases and seed from (processed) dumps. When we need to clean our database, we simply put down the docker container. Ie. we do not need to implement destructive database cleanup eliminating structure that could potentially fail.
Having policies about not accessing production database directly (and allow the extra time for building tooling around that policy), good preview / staging environments, etc. All fail eliminating structure.
I wouldn’t want to be on the wrong side of a lawsuit, defending drunk employees working on production data. What outrageous recklessness. And how imprudent to admit this to the public. Some things are best kept to yourself. No one needs to know that.
If you built good systems that have reliable backups or rollback mechanisms, dev should be able to have a beer on a friday and do a deployment without worrying about the fire they just caused.
I'd rather a culture where people admit to their mistakes than one where they try hide or get whipped for owning up to them.
We're people after all and some of us like a glass of wine and unwind while still performing our duties as engineers. After all, it's not like we're in charge of life support or critical systems which absolutely cannot fail.
We have something similar with AWS Cognito. If a user signs up but doesn't go through with the verification process, there's no setting to say "remove them after X days". So we have to run a batch job.
If I screw up one parameter, instead of deleting only unconfirmed users, I could delete all users. I have two redundant checks, first when the query is run to get the unconfirmed users, and then again checking the user's confirmed status before deleting them. And then I check one more time further down in the code for good measure. Not because I think the result will be different, but just in case one of the lines of code is altered somehow.
I put BIG LOUD comments everywhere of course. But it still terrifies me.
localhost is an abstraction, it's a non-routable-outside-your-machine network...except it's not. It's nothing more than normal TCP traffic except with a message to the OS and other programs that whatever is on that local computer network, you don't want it routed outside the local computer.
There's absolutely nothing stopping anything with access to localhost from routing it anywhere that process wants. Does not even take a malicious actor, all kinds of legit programs expose localhost. It's really not something you should use for anything except as a signal to other well-behaving programs that you are using the network stack as a machine-local IPC bus.
A very similar thing happened to living social during their brightest years, but the replication and the backups had failed too.
The oldest valid backup was about a week old.
It took the whole company offline for 2 days. It took a team of 10 people and some extra consultants to come up with a half backed version of the latest database based on elastic cache instances, redis caches and other "not meant to be a backup" services.
It was insane walking in an office that had hundreds of employees and see they all gone while we rebuild this cobbled together db.
At one point someone called it "good enough" and they basically had to honor the customer word if they had purchased something and it wasn't there.
It was a mess.
It was on all major news, and it was really bad press. In the end, they actually had a massive bump in their sales afterwards.
Everyone went to checkout their own purchases and ended up buying something else, and the news was like free ads.
Hmm, there seems to be some holes in their system. A database might go down for any reason.
I also have daily backups, but I write logs (locally and regularly copy from the production server) for all database actions to disk for the purpose of checking through them if something goes wrong, or having the option to replay them encase something like this happens. SO you have your database backups as "save points" and the logs to replay all the "actions" for that day.
I once replaced a bunch of customer photos with a picture of Spock as part of a test on my first week on the job.. The dB admin had just overwritten a sales force dev dB from production and a previous developer had hardcoded the IP address of production in the code of a script somewhere..
I like to use the lifecycle feature for suuuper core things that will never be deleted (VPC, r53 zone, etc) and eventually when I start targeting multiple DCs w/ lots of infra I'll eventually move to many state roots (or use tools like Terragrunt, which make things mildly scary again).
I am guessing this part. Even though the host is hardcoded as "localhost" , when you do a ssh port-forwarding, the localhost might actually be the real production.
e.g sudo ssh user@myserverip -L 3333:localhost:3306
In context, statement 1 regards proactively eliminating all bugs and risks. Statement 2 regards understanding the root cause of this particular incident and reactively fixing it so it won’t happen again.
Acknowledging statement 1 doesn’t mean giving up—it simply means being clear and realistic about the nature and scale of the problem we’re facing when we try to build complex software systems. In the face of that we can give up, or we can just do the best we can, and it sounds more like these people are doing the latter.
I like to think that, by addressing the known bugs as they pop up, over time you can box the complexity gremlins into tighter and more predictable spaces. Though as long as humans are building these systems, that box will always be there, and the predictability of those gremlins' behavior will only go so far.
> Note that host is hardcoded to localhost. This means it should never connect to any machine other than the developer machine.
Just to help with the postmortem:
1) “localhost” is just a loopback to whatever machine you’re on
2) the user and pw are pulled from config
So someone was running this from the production server or had the production DB mapped to localhost and ran it with a production config for some reason (working with prod data maybe). The hard coding to localhost will only ensure that it works for the machine it’s called on - in this case the prod server.
Things you might do to avoid this in the future include a wide spread of things, the main recommendations I’d have are:
This is what I came to say. Postgres usually allows a connection string that contains all the details as the database name postgres://username:password@host:port/dbname and I think it will take priority over separately specified host, depending on the client.
If you keep configuration in the environment (/etc/default/app-name) rather than in the application package, it's nearly impossible to make this mistake (especially with proper firewall rules). You can even package your config as a deb and keep it encrypted version control.
A question to the DBA experts from a developer: is there a way in MySQL and Postgres to configure a log specifically for destructive SQL queries so that it's easier to investigate a situation like this? I.e. to log most queries except for usual SELECT/INSERTs.
Also, @oppositelock pointed out that WAL would contain the destructive query too. How does one remove a single query from a WAL for replay or how does one correctly use WAL to recover after a 23-hour old backup was restored?
Finally, how does one work on the WAL level with managed DBs on AWS or DO if they block SSH access?
While I'm very sympathetic to "we accidentally nuked our prod DB" because, let's admit it: we've all been there at some point, I'm also a bit baffled here because I don't think that the problem lies with too much wine, Postgres permissions or scripts on localhosts but the fact that recreating a database by FIRST dropping all tables and THEN recreating them is like deliberately inviting those gremlins in.
But, as I said, that happens and blaming doesn't fix anything, so, for the future:
1. make a temporary backup of your database
2. create tables with new data
3. drop old tables
I totally sympathize with you and yours, I've made sphincter-clenching mistakes a handful of times during my 20 years of experience.
This is an abject lesson that understanding human psychology is actually a huge part of good architecture. Automating everything you do in production with a script that is QA tested prior to use is the best way to avoid catastrophic production outages.
It does take a bit longer to get from start to finish, and younger devs often try to ignore it, but it is worth putting a layer of oversight between your employees and your source of revenue.
I did the same thing once by accident and thankfully only lost 1 hour of data. The single lowest point of my career and thinking about the details of that day makes my stomach sink even now as I type this.
I ran a rake db schema dump command of some kind and instead of it returning the schema of the database, it decided to just completely nuke my entire database. Instantly. It's very easy to fuck up so cover your ass gents, and backup often and run your restores periodically to make you can actually do them in case of an emergency.
In one of my first jobs I deleted the last 30 days of our production data.
Shit happens. You learn and try to never repeat it. And share with others so hopefully they learn.
Ps. Don't do knee-jerk late at night quick patches. For example don't stop a database that has run out of disk space, try to migrate the data in memory first... And also do proper backup monitoring, and restores. Having 30 days of 0 byte backups is not that helpful. :)
I run a replicated ClickHouse server setup, Clickhouse uses zookeeper to enable replication.
The zookeeper instance was not replicated.it was a single node.
The server on which zookeeper was running ran out of hard disk and Clickhouse went into read only mode.
Luckily,no data was lost while this happened because we use RabbitMq to store the messages before it gets written to the db.
Thanks to RabbitMq's ACK mechanism.
While several other users have posted takeaways for how to prevent this from happening, I'd be interested in if anybody has an idea of how this happened given the code that was posted?
Presumably, a managed DB service should essentially never be available on `localhost`. Additionally, it would be very weird for `config.DevelopmentConfig` to return the production database credentials.
> Thankfully our database is a managed database from DigitalOcean, which means that DigitalOcean automatically do backups once a day.
Do cloud providers provide a smaller window for backup. Are there better ways to reduce the backup window here for DBs? Would love to understand any techniques folks use to minimize the backup window?
I did something similar once - I had fiddled with my /etc/hosts and subsequently connected to the production database without realizing. I dropped a table but thankfully it wasn't much of a deal - the monitoring rang the bell and I recreated it a few seconds later. All that happened was that I had logged out several hundred users.
Someone might have copy&pasted it elsewhere and that propagated away. This is why writing code can also be dangerous in open dev. Whoever programmed anything also should be sensible enough to judge their own code whether it could be dangerous in the wild. Once out there (or worse: on stackoverflow) it could wreak havoc.
I don't understand how this happened if localhost is hard coded and the password is different. I don't think they fully understand why this happened. At least enough to prevent it from happening again.
That is why we can only access the prod db from a jump box in our shop.... And even then it's just certain people with less privileges than a sysadmin account. No way you can do this accidentally from your laptop then...
The script also unnecessary complicates things. If it just does the equivalent of rake db:drop, this incident wouldn't have happened, since postgres wouldn't allow a database with active connections to be dropped.
You'll find that most users of cloud databases are in this boat. For example, on GCP, deleting the database instance deletes the backups! You have to write your own software if you want to survive that button click.
The maturity of the article is laughable. I'm sure my age is the same as the people who wrote it, but this is unacceptable: dropping databases in prod is a serious issue, not a joke. I think the culture of the company is toxic and not professional at any level. #change-my-mind
Here is how we had our database deletion error, about 15 years ago. Our DBs had been on leased servers at a hosting company in New York City. They were getting out of the datacenter business so we had to move. We were moving to colocated servers at a Seattle datacenter.
This was the procedure:
1. Restore DB backups in Seattle.
2. Set up replication from NYC to Seattle.
3. Start changing things to read from Seattle, with writes still going to NYC.
4. After everything is reading from Seattle and has been doing so with no problems for a while, change the replication to be two-way between NYC and Seattle.
5. Start switching writes to Seattle.
6. After both reads and writes are all going to Seattle and it has been that way for a while with no problems, turn off replication.
7. Notify me that I can wipe the NYC servers, for which we had root access but not console access. I wasn't in the IT department and wasn't involved in the first 6 steps, but had the most Unix experience and was thought to be the best at doing a thorough server wipe.
My server wipe procedure was something like this.
8. "DELETE FROM table_name" for each DB table.
9. "DROP TABLE table_name" for each DB table.
10. Stop the DB server.
11. Overwrite all the DB data files with random data.
12. Delete all the DB data files.
13. Delete everything else of ours.
14. Uninstall all packages we installed after the base system install.
15. Delete every data file I could find that #14 left behind.
16. Write files of random data to fill up all the free space.
The problem was with step #6. They declared it done and turned it over to me for step #7 without actually having done the "turn off replication" part of step #6. Step #8 was replicated to Seattle.
It took them a while to figure out that data was being deleted and why that was happened.
We were split across three office buildings, and the one I was in had not yet had phones installed in all the offices, and mine was one of the ones with a phone. None of the people whose offices did have phones were in, so they lost a few more minutes before realizing that someone would have to run a couple blocks to my office to tell me to stop the wipe.
It took about 12 hours or so afterwards for them to restore Seattle from the latest backup, and then replay the logs from between the backup time and the start of the deletes.
After that they were overly cautious, taking a long time to let me resume the NYC wipe. They went right up to the point where I told them if we didn't start now we might not finish, and reminded them that those machines had sensitive customer personal information on them and were probably going to end up being auctioned off on eBay by the hosting company. They came to their senses and told me to go ahead.
Like this never happens in big corps: sure, mostly on departmental level but still enterprises, not startups. They try to escape the annoying and slow as molasses dba’s/devops and install/create/buy/saas systems to avoid the red tape. But then the same things as with startups go wrong. And rather often too; we often had these calls asking if we could restore.
Obviously there are plenty of large enterprise wide data breaches, which, I would say is actually worse than losing a day of data in a lot of cases. So also not so many satefy measures, again, worse than startups; at least they have an excuse of being understaffed and underfunded.
Step 0) automated backups
Step 1) full manual testing of said automated backup
Step 2) weekly test of the supposedly automated process
... Now you can continue to your hair-brained engineering "processes"