@testbug
I am not an IT guy but I think when you clone the machine and use it as a test environment to install an update, you can't simulate the users activities so you maybe can't be sure how it goes when it's in live..
What I am worried about is the maintenance page. The page is so ugly and unprofessional, why as a supposed serious company don't you install at least a more friendly maintenance page....
And then come to my mind: why do you want a more friendly page when everyone left the office and are not planning on coming back
You are right, you can't simulate real user activity. BUT: I think they are not where they are testing user activity. I bet the MS SQL, Mariana, Redis whatever DB got updates or even got upgraded. I personally know that even updateing PHP to a newer version can bring big big problems.
What if they are just buying time for their exit? Shouldn't we do something?
What
can you do? The only thing goes in mind - that you should monitor exchange wallets for unexpected activity.
But even if you see something - what really can be done?
I've hear roumors about people forming groups, organizing wallet balances and stuff. I know, that like on every other exchange, there are people with several million dollars on the exchange. I personally would
START TO COMMUNICATE MOREDoes it make any sense that the maintenance takes so long? Their "system upgrade" should be as simple as getting more connectivity from their host, right? This shouldn't take more than a few minutes. Unless I am missing something.
My experience shows me, and i bet other guys from IT branche can confirm, that sometiems you want to do something "quick", like: "Hey, this shouldnt take too long, let's roll out that update". Someone has made a mistake and out the "let's roll out that update" is a "fuck this, how can we ever again have a working system?
A company like this most likely has:
- internal backup
- backups of database
- backups of wallets
- complete machine backups
- INSTANT restore if something is realy fucked up
- backup to tape
- backup to other servers
- backup to cloud
I have no idea what company is doing the IT work for kraken, but to be honest you should think about changeing your IT company. This is completely unacceptable.
Here is one szenario which came into my mind:
Maybe the kraken servers have been real physical machines and no one has ever thought about to migrate these machines into virtual machines. There are converter out there, being able to even virtualize an very very very very very very old linux system. If you start the virtualization there are
two szenarios: You fire up the virtualization and everything is fine. It will need a big amount of time but it will finish without problems. After that you change some settings on the VM and you can start the server and its good.
On the other hand (szenario two) you start the virtualization and it will stop/hang/freeze after a few minuts. If this happens, my experience says that i can take HOURS to get it fixed, as it can be required to do several fixes to get it running. Sadly after EVERY SINGLE FIX you start the virtualization again and receive a new error... This has to be repeated as often as required, and only god knows how often it has to be.
@kraken:Please communicate more. Even if it's like: we have still 100 more errors to fix, but we have already fixed 500 errors. Something like that should be ok.