You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
It is possible that monitored nodes have their hard drives put into "Read-Only" mode, and then Netdata Agent+Cloud will have every individual chart in an error state for the node, but the node itself will not be in an "error state" in Netdata Cloud. This happens at least in Ubuntu as this is the default behavior for failing hard drives, it may occur in other cases.
Picture attached for reference,
Description
When a node has a disk drive enter read-only mode, that information should be relayed in Netdata Cloud, probably flagged as a "red error"?
Arguably, any drive, not just the root drive as in my case here. This may not be the universally correct though, as presumably someone could have purposefully put a drive in read-only mode. I think the root drive being "ro" can clearly be considered an issue.
Importance
nice to have
Value proposition
The value proposition for this feature would be to allow users to quickly understand when their nodes have disk drives enter read-only mode. Especially in my specific case, with the root drive entering RO mode because of a failing disk drive, the entire node is non-functional, and ideally I would be notified of this so that I can get it replaced or repaired ASAP.
Proposed implementation
Even if the root drive is in this state, the read-only information can still be gathered from:
/proc/mounts
It can be parsed out of the drive info there, flagged as "ro", here is an example from a ubuntu server using LVM.
Problem
As a netdata user,
It is possible that monitored nodes have their hard drives put into "Read-Only" mode, and then Netdata Agent+Cloud will have every individual chart in an error state for the node, but the node itself will not be in an "error state" in Netdata Cloud. This happens at least in Ubuntu as this is the default behavior for failing hard drives, it may occur in other cases.
Picture attached for reference,
Description
When a node has a disk drive enter read-only mode, that information should be relayed in Netdata Cloud, probably flagged as a "red error"?
Arguably, any drive, not just the root drive as in my case here. This may not be the universally correct though, as presumably someone could have purposefully put a drive in read-only mode. I think the root drive being "ro" can clearly be considered an issue.
Importance
nice to have
Value proposition
The value proposition for this feature would be to allow users to quickly understand when their nodes have disk drives enter read-only mode. Especially in my specific case, with the root drive entering RO mode because of a failing disk drive, the entire node is non-functional, and ideally I would be notified of this so that I can get it replaced or repaired ASAP.
Proposed implementation
Even if the root drive is in this state, the read-only information can still be gathered from:
/proc/mounts
It can be parsed out of the drive info there, flagged as "ro", here is an example from a ubuntu server using LVM.
/dev/mapper/ubuntu--vg-ubuntu--lv / ext4 ro,relatime,data=ordered 0 0
The text was updated successfully, but these errors were encountered: