Shulou(Shulou.com)06/01 Report--

This article focuses on "what are the typical problems of the server disk array". Interested friends may wish to take a look. The method introduced in this paper is simple, fast and practical. Let's let the editor take you to learn what are the typical problems of the server disk array.

1. Error in writing or reading data to the disk array on the server (such as I error, read and write delay failure, etc.), or unable to write data, or error in the writing process

1) check whether the SCSI line and SCSI card are loose.

2) determine the disk array type: if it is a disk array of SCSI interface type, the first idea is to check whether the host channel interface of the disk array is connected incorrectly. If the SCSI line is connected to the out port, this kind of fault phenomenon is very easy to occur. At this time, you must remind the user to turn off the server first, then close the disk array, wait a moment, and immediately change the SCSI line to the in port after the static electricity is released.

3) View the channel speed:

Under normal circumstances, the channel frequency of our disk array is 160MHZ (the corresponding transmission rate is 320MB/s). If it is found that the disk channel frequency is 80 or lower, it is recommended to check the connectivity of the components on the SCSI link one by one. If necessary, you can verify your judgment by changing the host channel and replacing the SCSI card.

two。 The server writes or reads data to the disk array at a slow rate.

1) check the physical connection, such as the SCSI line, whether the SCSI card is loose or not. It is recommended that the SCSI line not be too long, so as not to affect the rate of Imax 0.

2) ask the 320MB/s card model, if the disk array with the speed of 320MB/s uses the SCSI card to connect the disk array, the server will not be able to access the array or the read and write rate will be very slow.

3) check the channel rate. If you find that someone has changed the channel frequency or that it is displayed as 80 or lower, the rate will be very slow. Try to modify it to 160MHZ.

4) View parameter settings:

Such as reading and writing strategies

There are two write strategies in the infrotrend array: write back and write through, the default is write back. When using the write back write strategy, if the server writes data to the array, the controller will first write the data to the cache, and then write to the hard disk periodically. Because the cache Imaco rate is much higher than the hard disk Imaco rate, it will improve the performance. The use of write through will cause the controller to write data directly to the hard disk without caching, and the performance of the disk array can be improved by modifying the write policy.

There are read strategy and write strategy in promise array.

There are three reading strategies: read cache,read ahead,no cache. By default, read ahead,no cache does not go through the cache when reading data, and read cache does not read data from the hard disk when reading data. Instead, it first puts the data to be read into the cache and reads the data directly from the cache. Read ahead puts the data to be read this time and the data that may be read next time into the cache, so the reading performance is better.

The writing strategy is similar to infrotrend.

5) setting of stripe size

The controller writes or reads data to the disk array in stripes, and the stripe size ranges from

(8KB-1MB take promise as an example, different manufacturers may not be able to.) for different applications, different stripe settings can lead to great differences in performance.

There are two typical applications:

A) Multimedia or video-on-demand system: it is required that the number of data Imax O should not be too many, and the amount of data at one time should be large, otherwise there will be multiple pauses on the client side. Therefore, the stripe size is relatively large, and the stripe size should be set to 64K or larger, otherwise the performance will be affected.

B) Database file access: the number of visits is relatively frequent, but each time the amount of data is small, which is usually the modification of a table or several fields. In this case, the stripe is required to be relatively small, which is generally set to 8K or less.

In the online support of common disk array problems, another common problem encountered is that the server can not recognize the disk array. The general analysis ideas are as follows:

As a hardware device, the disk array is recognized by the server at two levels

The first is that when the server starts, the HBA card scans the hardware device, which has nothing to do with the operating system, because the operating system is not started yet.

The second is whether the disk array can be recognized in the system after the operating system starts.

Therefore, in the case that the server cannot recognize the disk array, it is necessary to locate the level of the problem.

1. If the disk array is not recognized in the HBA card when the server is started

1) check whether the disk array mapping is done.

2) check the physical connection, HBA card and optical fiber line, SFP or SCSI card installation, and whether there is any problem with the connection. If it is an optical fiber card, check whether the LED status of the optical fiber card is normal.

3) the physical connection is normal, but the disk array is still not recognized, which needs to be confirmed.

A) compatibility between HBA card and disk array

If the adaptesci card is not compatible with the SCSI array, please refer to the relevant compatibility list for details.

B) whether the SFP of SCSI line or optical fiber line is damaged, if it passes through the fiber optic switch, check whether the corresponding port light is normal.

4) the disk array is strictly tested before it is sold, and there is generally little possibility of failure. Only when the above points are confirmed to be normal will the disk array be considered:

If one channel of the disk array is blocked, it can be replaced to another channel to test whether it is normal. If the other channel is normal, it may be a controller channel failure. At this time, you can visually see whether there are signs of man-made damage such as bent pins at the entrance.

two。 If the disk array can be recognized in the HBA card when the server is started, but not after entering the system.

1) first check whether the driver is installed, and then reinstall the driver if necessary

2) if the disk array device is recognized, but the partition is not, the typical case is the windows system to see if the disk can be activated or rescanned for hardware changes

3) when connecting to the server, the optical disk array may sometimes fail to recognize a partition after the hard disk is replaced or rebooted. At this time, the server and the disk array can be restarted. If you pass the optical switch halfway, restart the switch.

At this point, I believe you have a deeper understanding of "what are the typical problems of the server disk array?" you might as well do it in practice. Here is the website, more related content can enter the relevant channels to inquire, follow us, continue to learn!

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.