What are the errors, if any, that would be observed on /var/log/messages if an SSD is failing or has failed? I would like to know how such errors would be worded so that it could be picked up by an automated program to warn the user to replace the degraded disk before the second one fail.
Asked
Active
Viewed 137 times
-2
Question Overflow
- 2,103
- 7
- 30
- 45
-
How you set up the RAID? Software RAID via mdadm? Via a hardware RAID card? – Hennes Jan 23 '13 at 15:15
-
@Hennes, Via Hardware RAID – Question Overflow Jan 23 '13 at 15:16
-
@QuestionOverflow What type of hardware RAID controller is being used? – ewwhite Jan 23 '13 at 15:24
-
@ewwhite, wow, that is very specific. I am just asking because I intend to set up such a system, not that I have it now. :p – Question Overflow Jan 23 '13 at 15:26
-
With hardware RAID, it depends heavily on which controllers you use. They all report errors differently. Adaptec, HP, LSI, etc. – ewwhite Jan 23 '13 at 15:27
-
Please read our [faq] in particular http://serverfault.com/faq#dontask – user9517 Jan 23 '13 at 15:27
-
@ewwhite, I see.. But would such errors be reported on `/var/log/messages`, and how does one looks like? – Question Overflow Jan 23 '13 at 15:29
-
No, you would not see errors in most cases, but it depends on which controllers you use. – ewwhite Jan 27 '13 at 13:54
1 Answers
1
This depends on:
Your server hardware, storage backplane, OS type, whether you're using a hardware RAID or software RAID, the type of SSD, etc. Are you also looking at S.M.A.R.T. errors?
If hardware RAID, the failure would be reported the same way any disk would. What type of hardware controller is being used?
If software RAID, mdadm.
ewwhite
- 197,159
- 92
- 443
- 809
-
Yes, I am also looking at SMART via smartmontools, but will anything be shown on `/var/log/messages` if there is a failure? The OS is Fedora 17. – Question Overflow Jan 23 '13 at 15:22