Last time I talked about reading System Logs on the SQL Server box, explaining why it is really important that DBA(s) should scan through the logs once a day on a critical production system. As I mentioned in my previous blog post , sometimes there are messages logged in as information, and at times it can be treated as an early warning before the system gets actual error messages – a sign of warning or an error. That is why it is important to read the information level messages. Let me tell you yet another case that I had where the disk sub system issue was reported as an information in system logs.
In this case, the system was suffering with the high disk I/O. The disk that we had replaced was used for writing backups. For a few days we observed that writing backups were longer than it was before. The number of databases were the same and the size of these databases were not drastically increased, though the time it was taking to write backups had increased significantly. Looking at the system logs I noticed some messages related to the disk. Searching for those messages lead me to some links pointing toward a disk issue, link among them. After working with others in storage admin they confirmed the issue too, and now they are procuring a new disk.
So, here is what I would say. When you start your day, spare few minutes to read the system logs. At Pythian, we have our home grown monitoring tool Avail which does this job for us reporting information, warnings and errors as a report.
Excerpts from the System Log:
Log Name: System
Source: Server Administrator
Date: 6/18/2015 10:55:55 PM
Event ID: 2271
Task Category: Storage Service
The Patrol Read corrected a media error.: Physical Disk 0:0:10 Controller 0, Connector 0
Learn more about our expertise in SQL Server.