log full of "failed to write" std::bad_alloc

[Michael Wohlwend]

Hi,


I have an 8 node cluster and want to add another 8 nodes. The nodes nodes are
in a new rack.

Space used on each existing node is around 3tb on hdd, replication is 3

I added a new node with the new rack configuration.

After the weekend the node has disk usage of 8TB, is marked DN and the log is
flooded with "failed to write, std::bad_alloc.

No idea whats the reason of this, hdd space is 16 tb (2 disk a 7tb as raid0)


Any help would be highly welcome


Michael