Bug #10903
closed
Poor handling of disk full conditions
0%
Description
The disk on our SG-3100 recently filled up to 100%. The only symptom was that the DNS Resolver service failed, meaning it could no longer service DNS requests. Upon trying to re-start it, the service logged errors about failing to read the root.key and despite successful attempts to delete and re-create the file, it kept wiping the file upon the next start of the service and reporting the same errors again.
Eventually we realised and cleared some disk space and the service re-started OK.
Two suggestions come from this:
1) Can the DNS Resolver service not log something more useful if it fails to start due to a disk space issue?
2) Although the home page does indeed show the disk usage at 100%, could this event, which could affect many other services in unpredictable ways, not be highlighted to the user in a more obvious way; a pop-up alert maybe?
Updated by Jim Pingle over 4 years ago
- Category changed from Diagnostics to Notifications
- Status changed from New to Duplicate
- Target version deleted (
Future)
The DNS Resolver question is one you should ask upstream to Unbound developers -- We can't control what it decides to log when a fatal error occurs like that.
As for notifying about disk space, there are already feature requests for that (Like #10467).
Updated by Alan Ingram over 4 years ago
Jim Pingle wrote:
As for notifying about disk space, there are already feature requests for that (Like #10467).
This request suggests sending an eMail alert, which may be useful in some cases, but I think it would also be useful to alert the user as they logon to certain conditions (such as disk full) as I may have not been on the distribution of any eMail alert. Or maybe the eMail failed to send because the disk was full!
Updated by Jim Pingle over 4 years ago
Notifications like that happen using all available channels (GUI, mail, etc). There is no need for a separate ticket for each type of notification, it will all be handled on the same issue.