IBM Storwize V7000 Unified - description, connection and configuration of storage systems. Software and basic setup

RAM

IBM servers are sensitive to memory module configuration. Often, after an independent upgrade - adding memory or replacing it - the server does not boot, or sees less memory than actually installed. Fortunately, in such situations, you don’t have to guess for a long time about the causes of the failure: on the diagnostic panel (if there is one), two indicators, Config and Memory, light up.

Therefore, before upgrading memory, be sure to study the specification of what type and size of memory is supported by your server. The number of processors in the server is also of great importance - the order of placement of modules in slots depends on this. This also needs to be clarified in the specification.

In general, the situation with memory is exactly the same as described in the article about, for example. Briefly:

Observe memory channels.
Install ECC REG 1(2)Rx4 memory in dual-processor systems and UDIMM in single-processor systems.
Place the same amount of memory on each processor.

What should you do if you inserted the memory according to the instructions, but the server still does not work and the Memory indicator is treacherously lit? In this case, you will have to check different options:

This memory type is not supported by the server. Please check the specifications carefully.
The memory turned out to be “broken”. Replace the ruler with the exact same one and check if the server starts up.
The slot on the motherboard is clogged with dust. This is a fairly popular reason if the server has been in operation for several years, and even more so if you are not its first owner. Blow out the slots with compressed air.
Bent contact in socket. This happens very rarely, but it still happens: the memory refuses to work due to a bent contact in the processor socket. If the previous options did not help find the cause of the failure, remove the processor and carefully inspect the socket. If you are one of the lucky few, you can try to carefully straighten the bent contact, but this is entirely at your own peril and risk.

Many system administrators are faced with the fact that when checking RAM using MemTest86 they receive error messages even in modules that are known to work, or on the same tracks. This is especially common in M4 generation servers. This is not the fault of the machines or the memory: MemTest86 is not recommended for testing server memory. If the memory starts to fail, the server will report this through the diagnostic panel. It is better to check memory on IBM servers using standard self-diagnostic tools.

Drives

We have already repeatedly mentioned that it is not at all necessary to install “native” drives in servers. Neither IBM nor other vendors produce them; they only purchase them from well-known manufacturers, re-flare them and glue their logos on them. Therefore, you can easily save on upgrading or restoring disk arrays by choosing analogues instead of “native” drives. The two to three times difference in price justifies this, especially when it comes to refurbished servers. You can easily find model correspondence tables online, for example:

However, situations of incompatibility of “non-native” drives with the server are still possible. In this case, the server does not boot normally or does not see the drive. This is usually solved by installing fresh RAID controller firmware. By the way, it is recommended to update the firmware of the backplane/expander; the IBM Bootable Media Creator (BoMC) application will help you with this.

When you turn on the server and pass the POST check, an error may appear:

A discovery error has occurred, please powercycle the system and all the enclosures attached to this system.

This indicates a problem with one of the drives. It's easy to spot: the lights on its sled are constantly blinking, even when all other media have passed the test and have stopped blinking.

There are also more exotic problems with the disk subsystem. For example, when using RAID-1 in the MegaRAID Storage Manager proprietary application, errors like:

ID = 63
SEQUENCE NUMBER = 48442
TIME = 24-01-2016 17:03:59
LOCALIZED MESSAGE = Controller ID: 0 Consistency Check found inconsistent parity on VD strip: (VD = 0, strip = 637679)

Most often, this does not indicate a dying disk, but a parity error - a mismatch of data on the primary and secondary disks. Possible reasons:

Often such errors appear immediately after configuring a new array or after replacing one of the disks.
During a pancake surface diagnostic session, the disk is initialized and I/O operations are performed. On RAID-1, this can result in a temporary volume mismatch, which is automatically corrected the next time it is checked for compliance. This does not happen during any diagnostic session, but when the stars align:
- o A RAID controller without caching is used, or the Write Through mode is activated.
- o Lack of RAM, in which the disk is actively paging.
- o Just very heavy disk usage.

To solve this problem, it is recommended to reduce paging activity from the disk: use a RAID controller with caching and increase the amount of RAM.

Firmware and software update

A curious problem can occur when installing Windows 2012 or Windows 2012 R2 from scratch - the newly installed operating system does not see a single drive. Moreover, this happens not only with IBM servers. The fact is that all drives in the server are connected via RAID, and the mentioned OS versions do not have built-in drivers for working with RAID. And so they simply ignore them. What should I do? The most reliable way: use the IBM ServerGuide utility. When installing the OS, it forcibly inserts all the necessary drivers for a given model and version of the operating system. Please note that the OS image must be installed from a disk, not from a flash drive: ServerGuide will not work with an image on the same USB drive from which it is launched.

When purchasing servers, there are situations when you need to first update all the firmware, and then roll out the system. This can be done using the aforementioned IBM Bootable Media Creator:

Boot from a bootable USB flash drive or disk.
Run BoMC as Administrator.
Select what you want to do: update and/or run diagnostics.
The program will ask where it can get the drivers: download it itself or extract it from the archive you specified.
Select the medium to record the boot image: flash drive or disk. The recording can take several hours, don't worry, the program hasn't frozen.
After recording is complete, boot from this media and follow the instructions.

This procedure also helps in a number of problematic situations. For example, if you do not wait for the Integrated Management Module update to complete and click the “cancel” button, then on subsequent boots the server may not be able to load the IMM and uses the default settings. You can first try to recover using the “UEFI & IMM recovery jumper” on the motherboard, thanks to which the firmware IMM image is loaded.

But if it doesn’t help, then use the update procedure via BoMC.

There are also more unpleasant situations when, according to the law of meanness, during the installation of a more recent BIOS version, a power failure occurs.

After this, the server can no longer load the main firmware and uses the backup one. If the standard BIOS recovery procedure does not help, then do... a downgrade: install an older firmware than the one that was before the power failure. Usually this helps. After this, you can try to install the latest BIOS version again. As they say, one step back is two steps forward.

Benefits of IBM servers

IBM servers are deservedly very popular:

These are simple and very reliable machines.
Excellent expandability even on initial models and a rich scope of delivery.
IBM servers are usually cheaper than competitors and are not inferior in performance. For example, the M3 and M4 generations are cheaper than their counterparts from HP (Gen7 and Gen8) and Dell (11G and 12G).
The most inexpensive consumables. Easy to find in Russia.
Convenient diagnostic panel on many models.

The main thing in which IBM servers are inferior to competitors is that they have a very long “cold” start.

Tags: Add tags

It can be said without exaggeration that servers from IBM belong to the class of devices that, without exaggeration, can be called an “all-in-one” product. The user can buy a ready-made solution for their needs, configured and ready to use. In recent years, the company has concentrated its efforts on creating information equipment and implementing modern concepts based on 64-bit operating systems. However, the question is constantly asked about how to configure an IBM server, since, in addition to configuration by the manufacturer, you can install flexible software solutions for your own needs.

Company offers

IBM servers are not named by model or series like products from other manufacturers. There is a single concept, today it is System X, which is an integrated set of hardware and software solutions. Various modifications are available within the system depending on the hardware capabilities.

Tower servers can easily meet any needs of a small company or make it so that the deployment of a mobile office can be done in a matter of hours;
The rack design of the Rack allows for easy modification of capabilities by changing the number of processors, memory, network cards and types of RAID controllers. As part of the software concept, servers can be integrated into a single structure with scalability capabilities;
IBM blade servers are an ideal solution that allows you to host production facilities using separate units that combine high performance and the ability to create your own data warehouses.

In order to select a server, it is convenient to use the so-called configurators. These are online services that, based on basic requirements, can offer a specific performance. The user needs to determine the hardware requirements based on his tasks, decide how to configure the IBM server (which base software to use), select the type of case and specify the parameters. Let's try to list a couple of options in table form.

For more complex queries (when selecting a database server), you should clearly indicate the number of hard drives, the type of disk controller, memory for each processor, and so on. Configurators provide the opportunity to fill in as complex data as you like, decide how to configure an IBM server and select a specific hardware model.

Software and basic setup

As a minimum server setup, you will need to configure the raid controller and organize a convenient setup and maintenance system.

Data storage organization

All IBM servers make it easy to create any complex system of disk arrays right at the BIOS level. To do this, when loading the server, you need to go into the configuration program. This is usually done using the "F2" button.

The built-in management utility allows you to select individual drives, combine them into arrays and create RAID levels. The basic set includes common options, however, you can also build RAID 10 if you know the structure and define the disks yourself. The utility is quite convenient and intuitive.

Detailed settings are available for each created virtual disk. It includes the cluster size, access mode (the disk can be made read-only, for example), lazy or direct write mode, type of exchange (direct or using a specific cache size), and so on. After initialization, the disk array is completely ready for use and there is no need to select drivers and configure the controller after installing the operating system.

Integrated work environment

Probably the most popular solution for web application or service servers is IBM Web Sphere. It is an application server designed primarily for centralized management of the entire production environment. The toolkit includes many analysis tools - you can test a website or service for performance and monitor resource usage. It also presents backup tools that provide both complete data protection and the ability to easily transfer entire applications or services to another server.

Brief summary

Naturally, it is impossible to describe all the rich functionality and options for setting up servers in this article. Buying an IBM server means purchasing a hardware and software complex. Within a single integrated environment, IBM offers a variety of specific tools for virtualization, automated monitoring and load balancing. A huge advantage is several tools for creating data access sources. Almost all standards are supported, with the possible exception of Oracle and MS SQL, which can be integrated into the application server, however, they represent their own environments.

Even if you don’t plan to purchase proprietary customized software, any IBM solution leaves no room for error. IBM Web Spere, Lotus Domino and others make it easy to organize a well-thought-out data protection system. IBM designs its products in such a way that the client can in any case ensure round-the-clock availability of data, operability of services and functioning of the server in a “24 hours a day, 7 days a week, 365 days a year” mode.

In this article, we will look at the issue of installation and configuration on CentOS 7. In this manual, the installation of the trial version will be demonstrated. WebSphere, but it's no different from the full version, so it doesn't matter.

So, let's go!

1) Preparing and configuring the OS

In our work we will use the new CentOS 7. Surprisingly, out of the box it needs a lot of finishing to work, so be prepared for this. So, install the minimal version without graphics and let's go. Through the interface - immediately set up a network so that there is Internet... this will make your life much easier :)

Let's install the basic software... which for some reason is not included in the package:

Yum install net-tools nano wget

Now let's check our hostname and we'll fix it hosts(edit as you like):

Nano /etc/hostname nano /etc/hosts

Ifconfig -a

To fix this, you first need to fix it a little grub:

Nano /etc/default/grub

At the end of the line “ GRUB_CMDLINE_LINUX" need to add " net.ifnames=0 biosdevname=0“. You'll get something like this (not necessarily 1 in 1):

GRUB_CMDLINE_LINUX="rd.lvm.lv=rootvg/usrlv rd.lvm.lv=rootvg/swaplv crashkernel=auto vconsole.keymap=usrd.lvm.lv=rootvg/rootlv vconsole.font=latarcyrheb-sun16 rhgb quiet net.ifnames=0 biosdevname=0"

We rename our network interface to normal, classic “ eth0” and let’s rebuild:

Mv /etc/sysconfig/network-scripts/ifcfg-ens32 /etc/sysconfig/network-scripts/ifcfg-eth0 reboot

Setting up the network:

Nano /etc/sysconfig/network-scripts/ifcfg-eth0 DEVICE="eth0" ONBOOT=yes BOOTPROTO=static IPADDR=1.1.4.185 NETMASK=255.255.248.0 GATEWAY=1.1.1.9 DNS1=1.1.1.10 DNS2=1.1.1.90

Disable the extra one Network manager and let's rebuild:

Systemctl stop NetworkManager systemctl disable NetworkManager reboot

We check whether the system is designated as a thread IPv6:

lsmod | grep -i ipv6

If the messages contain references to IPv6, but it will be, then we proceed to disabling it:

Nano /etc/default/grub

At the beginning of the line “ GRUB_CMDLINE_LINUX" need to add " ipv6.disable=1“. You'll get something like this:

GRUB_CMDLINE_LINUX="ipv6.disable=1 rd.lvm.lv=rootvg/usrlv...

Create a new config and save the result:

Grub2-mkconfig -o /boot/grub2/grub.cfg

Reboot:

Let's check again and make sure everything is fine:

lsmod | grep -i ipv6

Adding to the system EPEL(all sorts of packages “burdened” with licenses) repository for CentOS 7:

Wget http://dl.fedoraproject.org/pub/epel/7/x86_64/e/epel-release-7-2.noarch.rpm rpm -ivh epel-release-7-2.noarch.rpm yum repolist

The new OS uses a “master” daemon that controls the other daemons. This systemd, which was introduced instead of outdated initialization scripts init.d. A new firewall is also used, firewalld instead of iptables. Let's check its operation and open the ports we need (9080 and 9443):

Systemctl status firewalld firewall-cmd --permanent --zone=public --add-port=9080/tcp firewall-cmd --permanent --zone=public --add-port=9443/tcp systemctl restart firewalld

As a matter of fact, this is where the OS configuration ends and we proceed directly to the installation IBM WebSphere Application Server Liberty Profile 8.5.5

2) Install WebSphere

We will need an account IBM. After regular registration, you can download any software (for development purposes, it is also called trial version).

They do not allow you to download the software directly. We download universal Installation Manager, and then through it we can download the software we need. Archive content BASETRIAL.agent.installer.linux.gtk.x86_64.zip unpack it into the was folder and then upload it to the server in /root

We grant permissions and start the installation:

Chmod -R 775 /root/was cd was ./installc -c

First thing, Installation Manager will ask us to enter our login and password for the IBM account. Press p and enter your credentials:

We select only the following items for installation (installation manager, websphere liberty and java sdk for it):

But we won’t install fixes. They are not required for installation, besides they are buggy and install with an error:

Final message. What is installed and where:

After that, we wait. How much to wait? Depends on your Internet speed and server load IBM. You will need to download about 500 MB, or even more. Be patient... What's going on? The installer connects its repositories and downloads the ordered software from it. Everything is beautiful.

The successful installation message looks like this:

Theoretically, it is also possible to install all this through response files, without dialogs. But this option also requires already installed Installation Manager, so in our case this is not relevant..

So, that's it! we installed IBM WebSphere Application Server Liberty Profile 8.5.5 and necessary for its operation Java! Congratulations! We will now look at what we can do next.

3) WebSphere setup

a) Starting WebSphere

Let's create our test server:

/opt/IBM/WebSphere/Liberty/bin/server create PROJECT

Created. The folder appears: /opt/IBM/WebSphere/Liberty/usr/servers/ PROJECT All settings and future modules will be located in it. To launch this joint venture, you need to add the line host=’1.1.4.185′ (with our IP), above httpPort=’9080′ (this is here: /opt/IBM/WebSphere/Liberty/usr/servers/PROJECT/ server.xml ). An example of such a config:

Let's launch:

/opt/IBM/WebSphere/Liberty/bin/server start PROJECT

Going to the address http://1.1.4.185:9080, we will see the following:

This means that everything is fine and the websphere has started.

b) Installation of the administration module

This item is optional. But with the administration module it is more convenient to work with the web sphere. Through it, you can stop and start modules individually, without having to stop the entire server.

So, install this module:

/opt/IBM/WebSphere/Liberty/bin/featureManager install adminCenter-1.0 --when-file-exists=ignore

To log in to the admin area as an admin, use the account: admin/password. And under the user: nonadmin/nonadminpwd.

Its login address is: http://1.1.4.185:9080/adminCenter/ The admin panel looks like this:

All! The administration module is installed.

c) Installing an extension module

Also, you need to install on the Websphere extended packages (an extended set of libraries and binaries), this is done extremely simply:

/opt/IBM/WebSphere/Liberty/bin/featureManager install extendedPackage-1.0

d) Installation of modules

We come to the most interesting part. Installing modules in Liberty. How to do this? There are 2 ways, through the /opt/IBM/WebSphere/Liberty/usr/servers/PROJECT/ folder dropins and /opt/IBM/WebSphere/Liberty/usr/servers/PROJECT/ apps
From the catalog dropins modules are picked up and installed automatically. From the catalog apps– they must be manually registered in the server.xml config. An example of a config to which the module is connected via apps:

To run the SP not in the background and with logs, run the command:

/opt/IBM/WebSphere/Liberty/bin/server run PROJECT

e) Pros

Testing has verified that it is enough to copy the /opt/IBM folder to another server and everything will work out of the box. Very comfortably. Those. we can set up the joint venture we need in advance and supply the entire software package at once. And “Liberty Websphere” is very lightweight and starts/stops very quickly :)

Published

Today's post will focus on IBM Storwize V7000 Unified.

Let's look at connection and initialization issues, and also conduct a small performance test.

First, some background information on:

IBM Storwize V7000 Unified is a unified data storage system with the ability to simultaneously provide block and file access (SAN and NAS). File access via NFS/CIFS/FTP/HTTPS/SCP file protocols. As well as local and remote file replication. Well, plus all the usefulness and goodies inherent in the original Storwize V7000, and these are: Thin Provisioning (virtual allocation of disk space), FlashCopy (creating snapshots and clones of volumes), Easy Tier (multi-level storage), Data Migration (data migration), Real-time Performance, Metro and/or Global Mirror (remote replication), External Virtualization (virtualization of external storage systems), Real-time Compression (data compression).

The system consists of the V7000 itself and two file modules (a kind of system x server with specialized software installed on it) united in a cluster under the control of a single graphical interface, as they say at IBM - one system, one control, one unified solution.

Installation and initialization of the system is quite simple, the main thing is to make sure that the switching is correct and have a clear understanding of the procedure, and it also does not hurt to visit the IBM Storwize V7000 Unified Information Center (http://pic.dhe.ibm.com/infocenter/storwize/unified_ic /index.jsp?topic=%2Fcom.ibm.storwize.v7000.unified.132.doc%2Fmanpages%2Fdetachnw.html)

Example of IBM Storwize V7000 system switching

To initialize, perform the following procedure:

Click “Launch GUI” and a browser will open according to the ip specified in the Management IP item, where we will see the system initialization process. Upon completion, having specified all the necessary parameters, a GUI that is already familiar, but filled with new items, awaits us.

If something went wrong and a problem arose during initialization, you should pay attention to the file “satask_result.html” located on the flash drive with the utility; as a rule, it contains the number of the error due to which the failure occurred. Re-initialization is unlikely to succeed if at least one of the system elements has already been configured, so all settings must be reset. The reset is performed as follows: on the storage system itself, you need to go to the service graphical interface of the controllers (the IP address can be changed using the same InitTool utility, the default address is 192.168.70.121/122), switch node1 and node2 to service mode (“Enter” Service State”), then on the “Manage System” tab clear the system information of the selected node, then go to the “Configure Enclosure” tab and reset the system ID (check the “Reset the system ID” box and click “Modify”), this sequence of actions must be done for both controllers (selecting node1 and node2 on the “Home” tab in turn), after which you must reboot the storage system. To delete the configuration on file modules, you need to reinstall the system from the included disk, after executing the commands on the loaded modules, username/password (root/Passw0rd), then ( $ rm -rf /persist/*), and check that the file has been deleted ( $ ls -ahl /persist/*), insert the disk and reboot ( $reboot), installation will begin automatically after confirmation (press “Enter”).

There are several graphs of system performance with block access.

Host, OS Windows Server 2012, tested two local disks presented via FC, one 100Gb with RAID10 of 4 SSD 200Gb and the second 100Gb with a pool consisting of 3 RAID5, containing 19 SAS disks (300Gb 15k), two The RAID groups included seven disks, and the third group had five. Testing was carried out with the IOmeter program, two specifications “100%Random-8k-70%Read” were used - test in 8kb blocks, 100% random access, 70% read operations, 30% write. And “Max Throughput-50%Read” - test in 32kb blocks, 100% sequential access, 50% read and write operations. The queue depth had a value of 64.

Hello everyone, today I want to tell you how to configure the IBM Storwize v3700 storage system. The most basic diagram:

IBM Storwize v3700 storage system as standard with the ability to connect servers via iSCSI and SAS. Installed 4 disks of 600Gb
two IBM 3650 m4 servers, without local disks, with two single-port SAS HBA cards
cross-to-cross connection, fault-tolerant - each server HBA adapter is connected to its own storage controller

The task is as follows:

Connect to storage system for management
Update the firmware to support SAS connections
Create array from disks, RAID level 10
Since we have servers without hard drives, we create a separate LUN for each server to install the Windows server 2012 operating system
We create one common LUN that will be accessible to both servers. It will be used to create a MS SQL 2012 cluster, more precisely for storing databases

The storage system comes with a special flash drive; it is used for initial configuration, namely setting the administrator password and service IP address for connecting to the web interface. From the flash drive on the computer, run the InitTool.bat utility

Since we just took the storage system out of the box, select the Create a new system option

We set the IP address by which we will connect to the storage system

System initialization process:

We safely remove the device from the computer and take out the flash drive.
We look at one of the storage system controllers. We will need to insert the USB flash drive into one of the connectors with network management interfaces. But before that, you need to make sure that on the upper right side of the controller three indicator lights are sending the correct semaphore signals, the left one is on, the middle one is blinking, the right one is off.
After the flash drive is placed in a USB port (any). The right icon (exclamation) starts flashing. You need to wait until it stops going out, after which you can remove the flash drive and return it back to the computer to complete the wizard’s steps.

Through a browser (IE8 or Firefox 23+ is recommended) we go to the web interface. The default login password for IBM Storwize v3700 superuser is passw0rd (separated by zero)
Now is the time to update the firmware; to do this, go to the menu Settings -> General -> Upgrade Machine Code. The firmware was downloaded in advance from the official website ibm.com. In our case, this is Version 7.1.0.3 (build 80.3.1308121000). It includes an Upgrade test utility, first we load it onto the storage system, and then the firmware itself.

The storage system automatically detected 4 installed disks. The system assigned three of them to POOL, and assigned one to hot spare. If there were more disks, it might make sense to leave such an automatic setting. In our case, it is better to repartition the disks differently.

Delete the automatically created Pool, Delete.

We get 4 free disks from which we will create RAID 10

Click Configure Storage, then select which RAID we want to create and how many disks will be used for it.

Set the name for the newly created Pool. So as not to get confused in terms. We create a RAID or array from free disks, the resulting free space is Pool. Then we will cut the pool space itself into pieces, the so-called LUNs or Volume, and now they can be presented to servers (hosts).

Pool has been created

Create a new LUN (here it is called Volume) in our pool, Select Generic-Create

set the volume size

Thus, using the volume creation wizard, we make 3 moons. As planned, two 100Gb each for server operating systems. And one common size of 500Gb for creating an MS SQL 2012 cluster

Now you need to tell the storage system which servers (host) are connected to it. In the basic configuration there are only two connection options - iSCSI and SAS. We have two servers that are connected to Storwize v3700 via SAS

At this step, we indicate to the storage system that our first server is connected to it with two SAS cables, which are plugged into two SAS HBA cards with identifiers (16-digit) in the server. Thus, we add both servers, each with two identifiers.

We present Volumes to servers. In other words, we assign access rights. In the screenshot, HOST_LUN_TOP is intended only for the first server, because its operating system will be installed on it. And the second server cannot see this LUN. Unlike SQL_LUN, which must be accessible to both servers, because the MS SQL cluster databases will be located on it.

At any time you can view and change access rights to the LUN (Host Mapping)

At this point, the storage setup can be considered complete. You can start installing Windows server 2012 R2 on the server and when you select a hard drive, two LUNs from the storage system will be available, one 100GB in size, the other 500GB. This will mean that the setup was successful.