Linux disk space Full processing method

There may be two situations in which there is space under Linux

You can view disk space consumption by command df-h, and actually view the file (block) occupied by the disk block Df-i view the occupied (inodes) disk block and index node of the index node is full, resulting in the inability to create the file, prompting disk space is full. So please note that there is room for the view disk, but the creation of the file prompt is full, possibly the Inodes node is full via command: Find/-size +100m |xargs ls-lh lists files larger than 100M in the system du-h--max-depth= 1 View the size of the folder in the current directory can be tested in two ways, sometimes, occupy a lot of space is actually the system log or business accessories, such as small files, so according to the size of the file to find not necessarily can identify the disk space is occupied by the "directory" or file, the recommended use of the second log file large Web Access log, It is recommended to do a scheduled task log cutting, in days to do a compressed backup (can be implemented through scripting), to keep a certain time for viewing, there is a need to do log analysis. For example, in the case of a one-click installation package, the/alidata/log/directory is a log file that checks to see if the file size is full disk usage, and generally, the Access directory file is the log of the error. If the disk is slow, these logs are not important, then you can do the empty processing, Method 1, RM Delete, 2 echo "" >access.log a large number of small files due to business needs, resulting in a large number of small files, need to consider new additions to the disk or the use of OSS storage Services/HTTP Help.aliyun.com/all/11108271.html?spm=5176.383377.1996834802.6.7hpore can be based on the above steps to troubleshoot, if still unable to resolve, you can submit after-sales work orders, Server account password information, etc., for troubleshooting.



What if the server goes down?

Start the machine to see if it can be logged in, please check:

A, whether the application caused memory overflow or leakage caused B, whether the process is too high or constantly created, exhaustion of resources caused by C, whether the database program deadlock, the number of connections caused D, whether it is an application exception caused by E, whether the traffic load is too large to cause F, is a hacker attack caused G, If this is a mistake, you can query the system log to see if there is an exception record if you cannot view the failed scene.



Linux often crashes, snaps

Linux operating system server in the event of a crash or a Dayton situation, you can refer to the following items for overall troubleshooting:

A, check the server processes and services are not consuming too much memory or not free memory, when the server memory is exhausted, and then will quickly go down; b, see if there are cron (scheduled tasks, automatic tasks) in the/var/spool/cron within the corresponding time period; C, Check that the parameters of the Web server exceed the performance of the server, such as the maximum number of connections allowed, etc. d, the number of processes is high, the service is paralyzed, the machine suspended animation (for example, zombie process is excessive, has been occupying resources); E, cat/var/log/message see if the system log has an exception F, Check if the disk has bad blocks. (You can use the Badblocks command with the parameters) G, the kernel consumes too much, to see if there is an instantaneous resource consuming too large processes or services; h, ps-ef view process exception No, whether there is an attack

Linux disk space Full processing method