[wip] add first version of setup for creating disk with failures #173
Conversation
fbalak
left a comment
fbalak
left a comment
There was a problem hiding this comment.
Looks good. Just a few questions...
Also please add PR description.
| # Get number of failed sectors till next failure | ||
| repeated_set=$(echo "${repeat_failure_every}*1024*1024/${sector_size}" | bc) | ||
|
|
||
| #block100=10 |
There was a problem hiding this comment.
Why is there this commented code?
There was a problem hiding this comment.
This code is here so people can try it with human easily countable data sizes. Should I remove it?
|
@mkudlej are the created devices "permanent"? (do they survive reboot?) |
|
@dahorak I honestly don't know if they are permanent. I think that these changes are not permanent. It is similar as any other device mapper changes made by lvm. You should store them in configuration files, see conf files for lvm disks. |
|
@mkudlej would it be possible to create it permanent (store the configuration in configuration files) directly in the |
|
@mkudlej I wasn't able to create any volume on the devices :-/ (I've tried it multiple times with similar results). |
mbukatov
changed the title
|
I guess we are trying too hard here, mkfs fails on such device, so there is nothing left to try. |
|
Actually it would be better to try to:
Errors simulated by the flakey device should not prevent mounting of the brick and gluster daemon started. |
5 participants
No description provided.