[O & M personnel? Why? 14. clean the room and treat the guests again.

[O & M personnel? Why? 14. clean the room and treat the guests again? Why? 13. The LDAP server cannot log in to discuss the analysis ideas. http://www.2cto.com/os/201304/205318.html Sometimes there are updates in the wood? Communications Administration? Song? South stealing heart? Yunxiao ?.. [O & M personnel? Why? 14. clean the room and treat the guests again? Why? 13. The LDAP server cannot log in to discuss the analysis ideas. http://www.2cto.com/os/201304/205318.html Sometimes there are updates in the wood? Communications Administration? Song? South stealing heart? Yunxiao? /Div> However, I think this kind of square is not good, maybe the subsequent articles are all like this. are people in O & M? Tai Tsui? Are you happy to receive porn? Why? Why? /Div> However, if you look down on the logic, trying to find the answer will waste more time. This is very dangerous. I remember Sherlock Holmes mentioned: making guesses without facts as a reference is a terrible mistake. People with incorrect feelings always use facts to set their own guesses (arguments) instead of deriving conclusions based on the obtained facts in the correct way, it is important to see whether the obtained facts match the facts (none of them). you must first know the fault site !! The following are some practices that make it easier for you to understand the so-called "facts" in Sherlock Holmes. when you look at the official website, please smile. (1) the past and present of the problem will not be immediately merged into the server. you need to clearly understand this the number of known server conditions must be clarified, as shown in the following figure: (2) who is there? The truth is, a mountain cannot be used together, unless, And haha. who is online now? Which users have accessed it? Do you want to know? Oh, so let's see what happened to the two of them. ● w ● what happened before last 3? In the late autumn, you have to turn over the old accounts. if you are thick or not, you can use the following command: ① command: history ② variable: HISTTIMEFORMAT: The two men are poor brother and brother of Yiba shuishui. if they leave, they will be crazy. (4) What is the process running now? This is relatively easy. there are major routes to Rome, such: ● pstree-a or, ● ps aux, and listen to the network service, not much to talk about, bright three kill guys ● netstat-lntp ● netstat-lnup ● netstat-lnxp: What is the meaning of netstat? I don't know what the room means. why is the CPU and memory here? I feel like I have a bit of experience solving the equation. I have used several commands/tools to answer a few questions. this seems to have returned to the distant junior high school. at that time, I was still young. for example, the equation was: ● free-m ● uptime ● top ● htop for example, the problem is: ① is there any spare memory? Is swap used? ② Are there any remaining CPUs? How many cores? Are there too many core loads? ③ Where does the maximum server load come from? What is the average load? Middleware hardware ● lspci ● ethtool ● dmidlecode if there are many servers or bare metal status, you can check: ① find the RAID card (do you want to bring the BBU backup battery ?), CPU, spare memory slots ② is the NIC set? Is it in half duplex status? Is the speed 10 MBps? Is there a TX/RX error? ⑻ IO performance ● iostat-kx 2 ● vmstat 2 10 ● mpstat 2 10 ● dstat -- top-io -- top-bio these commands have many benefits for backend optimization N: ① check the disk usage and whether the disk is full? ② Is swap switching mode enabled (si/so) ③ who occupies the CPU? ④ Dstat can see who is performing IO: whether the mount point is MySQLD or PHP or other fuse mount points and file systems ● mount ● cat/etc/fstab ● vgs ● pvs ● lvs ● df-h answer the following questions: ① total number of mounted file systems ② whether there is a dedicated service file system ③ file system mounting options? ④ whether there are large files deleted but the space is not released ⑤ whether there is space to expand a partition kernel kernel, interrupt, and network ● sysctl-a | grep... ● cat/proc/interrupts ● cat/proc/net/ip_conntrack ● netstat ● ss-s answers the following questions: ① display all existing connections, the netstat may be slow, and the ss can first understand the overall ② in different states (TIME_WAIT ....) how to set TCP connection time ③ whether conntrack_max is set to be large enough to cope with traffic on your server ④ What are SWAP switching settings? For workstation, it is good to set swappiness to 60. 5. is the interrupted request evenly allocated to the CPU for processing? Logging system logs and kernel messages ● dmesg ● less/var/log/messages ● less/var/log/secure ● less/var/log/auth answer the following questions: ① check the error and alarm information, such as whether there are overloaded connections ② check whether there are hardware errors or file system errors ③ analyze these error events and the previous suspect points for time comparison and timing task ① is there a scheduled task that runs too frequently? ② Have some users submitted hidden scheduled tasks? ③ When a fault occurs, is there a backup task in progress? ⒀ The system logs of other applications here are huge, broken, crazy, (N adjectives are omitted here), and projects such: oracle alert and trc file Nagios logs .... after so much trouble, I think, I have some understanding about the fault scene. at this moment, we can start troubleshoting, and the article is coming to an end. since it is the end, the problem is coming again. everyone has heard of the old and difficult problem that "fruit and vegetables are increasing prices in the last mile". This is the last kilometer problem on the Internet, come on, don't worry, let's not talk about the last thing. when the server is troubleshooting the problem, let's look at the present. come, line up, 1, 2, 3. let's shout together, we the target is "the first 10 minutes "!!!