Attention and accumulation of Linux server availability skills

Attention and accumulation of Linux server availability skills-Linux Enterprise applications-Linux server application information. The following is a detailed description. Many Windows platform DBAs must be annoyed with the problem of "Restarting to take effect" When upgrading the operating system. This may be the reason why few may be willing to manage Windows platform databases. In fact, sometimes there are similar issues in Linux. to Patch the Kernel, you must restart the operating system unless you ignore it. The recent Slashdot news about Linux is worth noting. Ksplice: Rebootless Linux kernel security updates is a key technical improvement for DBAs who are very concerned about system availability.

To improve availability technology, careful planning at the early stage is an important part. For example, if the fsck of a large file system is faulty, the system automatically starts the fsck check if the number of mount times or exceeds a certain time by default. If a Linux Server running for a period of time crashes after reboot, the file system verification takes a long time, which is desperate. If you pre-process the problem initially, you can avoid unnecessary downtime.

In addition, it is essential to accumulate the "High Availability" technologies or skills during maintenance. For example, the Fdisk command cannot be used to re-read the partition table from the Kernel, but the partprobe command mentioned here is useful.

In the past, I have recorded how Linux can identify new LUN topics without restarting. This is useful even if the number of newly added Luns falls short.