rpki-client broke today after creating 876’552 inodes and filling up the entire FS. A 16 GB FS is plenty data-wise, but the default inode limit is reached much faster than expected, requiring either a new VM with a higher percentage reserved, or more GBs.
It's all /var/cache/rpki-client .
For future reference, that's my current inode count:
https://paste.daknob.net/paste/XiHfWDb8AANbS5-0C2GYn9u9i8_PwHLaPfpDp8e27xU
@antonis hmm
(bytes)
tmpfs 10G 1.7G 8.4G 17% /usr/local/var/cache
(inodes)
tmpfs 32M 496K 31M 2% /usr/local/var/cache
Yeah this seems pretty rough by the looks of things, I put stuff ages ago into a tmpfs simply because the volume of churn is not worth putting on to the SSD (especially in bgp.tool's case where has to read every single file about every ten minutes)
athanasios
Antonis Chariton
benjojo