|
|
|
|
|
by alexandrerond
1684 days ago
|
|
I run some machines on Hetzner with 10x disks sw-RAIDed in a very hands off fashion. From time to time I look how they're doing. Usually a few disks have blown up, or are erroring or have smart errors. I then need to open 1 ticket per disk to get it fixed. They get replaced with really old disks (25k+ hours) which suffer same failures soon after. At one point support complained that I was making too many requests to change their broken disks as if it was my fault that 20% were screwed. Then they treat you like an idiot as much as they can. They ask you to provide disk slots numbering for disks which are impossible to get/know for a customer because they cannot read the serial number labels on the chassis and generally assume customer's fault to not know where they plugged something. Sometimes machines do not restart after changing the disks and they don't realize so have to follow up again. Network isolation doesn't exist so any weird network behavior from the machines towards then LAN gets you blocked. If something is wrong you have to beg for some console access that they need to enable on demand only to find they plugged something wrong. But they're cheap so... |
|