The master has a trick to deal with repeated disk loss.

Recently, I encountered a very depressing thing. The virtual disk allocated to a server on the CX3-10C is always dropped. The reason is unknown at first, but it is normal now. It’s hard to figure out?

The two virtual disks on the video server are gone. The storage array management terminal prompts that the fiber is not connected. Then I guess that the Netcom person has touched the line or unplugged it. Because they have recently touched the line and prepared to relocate the machine room, it is very likely to be knocked out. The phone told Chen Qian, for possible reasons, I went to the computer room to check the server fiber connection and deal with this failure. After I went there, I found that the fiber was loose and the indicator light was off. After plugging in and restarting, everything is fine.

But the next morning, that is, yesterday morning, the problem was that the two virtual disks on the video server were gone, but the storage array management terminal showed that everything is working properly, and the server-side powerpath is also Normal, there is no error, just a few prompt errors in the system log.

The cause of the problem is unknown. The server also prompts normal, but it does not have a disk, which is very surprising. Then I restarted the server, because many problems with the windows system are normal after restarting. After restarting, it is normal, but I don’t know the reason. At least I think there is no problem on the storage array. The fiber switch is definitely no problem, even if There is a problem with a certain port. I suspect that the biggest one is that Netcom touched the line again. I called the person in charge of Netcom to ask if anyone went to the computer room. They said that no one went to the computer room. I was very suspicious, but no. The solution, because our computer room has no access control system for the time being, and there is no way to judge it. I will temporarily put this question here, and the log does not see what the reason is.

Depressed is two hours later, the disk is gone, the prompts are normal, the error log is the same as last time, I think it will be a fiber problem, went to the computer room to put the fiber and The fiber switch ports have been changed, and the restart of the machine is normal, but it is still useless. The original problem occurred at 3 o'clock in the afternoon, and the restart is normal again. To further observe the problem, the error is the same every time, that is, the quantity. A little different.

Because there is still one thing that is more important than this, I will go to work in the afternoon and go to do another thing. After 8:00, I will call my colleagues on the road and say that there is a problem. I said that you should restart it first. Return to normal, research immediately after returning, monitoring at the same time, and backing up the above data to prevent the disk from being unloaded several times, but the backup is broken. Because it is very late, the server access is not very high, at night. There is no overtime processing. To tell the truth, even if I work overtime, I don’t know what to do. I went home and thought about it. I decided that the rest can only be a problem with the HBA card. The port has also been changed. The fiber has also been changed, or it is repetitive. This problem occurs, intermittent, really annoying … …

After the morning restart, contact the dell service engineer to discuss the problem solution, consult a similar problem, he looked at the storage array The output of the support, command of the spa, spb, system log, and fiber switch (I don't understand that, he let me grab it for him). Then I also concluded that the HBA is broken. I was already in the computer room. Since there are two HBAs on this server, I connected the fiber to another HBA card, and then re-registered the HBA card to form a storage array. Effective access to links. Needless to say, the switching process is the general registration process, very simple, and the problem is solved. Everything is back to normal. The whole is a magical oolong event!