Dell EqualLogic PS4000: Hands-on Review Part 3

Share Button

By TechHead guest contributor, James Pearce: EqualLogic’s PS4000 iSCSI storage arrays are targeted at the SME sector, particularly for VMware virtualisation, as well as “branch office” use for larger companies.  The XV model is reviewed, with 15k SAS drives, dual controllers and dual power supplies.

This is quite a complicated product, so I’ve split this into four sections:

Dell EqualLogicPart 1 – The PS4000

Part 2 – EqualLogic Networking with Force10

Part 3 – System Management and Monitoring

Part 4 – Performance

Part 2 was a bit deep – fortunately this part is basically a run-through of a few screenshots…

 

Part 3 – System Management and Monitoring

Management

Management is via three core applications,

  • The built-in web interface (Group Manager), providing LUN management, most system settings, (one way) RAID level migration, remote replication etc
  • SanHQ, a SQL-Server based application that continually gathers stats from the arrays via SNMP and provides details system stats such as IOPS rates, NIC Mbps, command latencies and so on, and
  • Auto Snapshot Manager / vmware Edition (ASM/VE)

There is also a Telnet and FTP server to retrieve diagnostic logs, and a serial port console used for the initial “out-the-box” setup and some diagnostic tasks (such as resetting the admin account (‘grpadmin’) password via the special username ‘recoverpassword’).

ASM/VE is a slightly clunky app that connects in to vCentre Server to provide storage layer snapshot control with various scheduling options.  Although it works well enough, one can’t help but wonder if this would have been better as a vSphere Client plug-in – separate applications just spell confusion and potentially disastrous mishaps in my book.  In this article I’ll cover the other two core apps – SanHQ and Group Manager.

 

Initial Configuration

The initial configuration is completed through the console port using Hyper-terminal (look out Windows 7 users!).  It’s very straight-forward since a (text-mode) wizard asks a few basic questions like the management IP address – and once that’s on, setup can be completed using Group Manager from a browser.

The system immediately commences a background initialisation task, so the full capacity is available immediately, all be it with reduced performance.  But even so, the model reviewed here completed this task in just a few hours.

 

Group Manager – A Whistle Stop Tour

Group manager is the core PS4000 management application suite and for many users, it’s all that’s needed.  The software will configure all units in the group and enables the user to drill down to an individual unit if required.

Dell EqualLogic - Group Manager

The ‘storage virtualisation’ approach is evident throughout – Storage volumes (LUNs) appear quite separately to the PS units (members).

Dell EqualLogic - Group Manager

Functions like creating a LUN, and even RAID migration on a member, are about as easy as it gets, and almost everything can be changed on-the-fly.  Firmware upgrades too can be done online, the system ‘failing over’ to the secondary controller and then back again whilst each controller is patched (it’s worth noting that EqualLogic recommend scheduling downtime for this however).

Dell EqualLogic    Dell EqualLogic

One notable limitation is that RAID level migration is a one-way process, regardless of how much storage is actually in use.  Hence RAID-10 can only be selected in the initial build, and requires a complete system reset, destroying all data, to get back to it.

Group Manager also enables the email alerting capabilities to be configured, the PS arrays then providing notification of critical events such as capacity warnings (when using thin provisioning and snapshots) and of course hardware warnings.  For the PS series there is no ‘dial-home’ alerting, so the administrator will still need to log hardware failures with Dell Support to get replacements.

 

SanHQ

SanHQ is a great application, collecting and recording performance metrics and alerts and keeping this available for historic charting.  Combined with the vCentre Server’s performance monitoring capabilities at the ESX host and VM levels, the administrator can gain a view of their infrastructures demands and limitations as a whole that simply wasn’t possible with physical servers.

This monitoring and historic charting capability can provide even the SME with sound basis to their IT budgets and (when actually used!) avoid the all too familiar pitfalls of performance bottlenecks appearing in systems that have to budget for upgrade available.

The client application seems to have a memory leak and when left open for weeks at a time, it will eventually stall whatever it’s running on completely.  This might need some consideration if the server (or VM) it’s running on is also managing other key tasks, such as UPS controlled infrastructure shutdown for example.  Of course the application need not be run on the server at all and might be better installed directly on an administrative desktop.

Charting

The charting capabilities of the application provide all the key metrics right from a 5-minute real-time view to historic views of the past 30 days and even specific SanHQtime windows beyond that.  As with vSphere, the available resolution is reduced the further back is viewed, with data from three-months back being pretty well smoothed (but still potentially extremely useful).

One of the most useful views is the “IO” screen (right), providing IO sizes, latencies, and throughput in both IOPS and MB/s.

The depth of information available is impressive, right down to the queue lengths against individual disks (below left), whilst the network view provides yet more great detail, helpfully including packet retransmit rates that can have a horrid impact on iSCSI performance (below right).

SanHQ  SanHQ

The application also provides data about remote replication, if configured, and provides a further mechanism for email altering.  Unfortunately there is no ability to gather or generate statistics on key metrics for remote replication without having a partner configured however – it would be helpful for the SME assessing remote replication options to gain an understanding of the performance hit, the space requirements and the bandwidth requirements ahead of signing on the line for more EqualLogic hardware.

 

Coming Next

In the final section, I’ll look at the raw performance with RAID 50, 5 and 6 and see how this stacks up when compared to several other storage solutions – the software-based StarWind iSCSI solution, local disk, and an old Dell/EMC AX-150.

_________________________________________________

About the Author

James PearceJames is regular guest contributor to TechHead and is a Kent based qualified accountant, currently working in information security and technical architecture with  most of his  time “being spent on virtualisation and business continuity at the moment”. Check out his new virtualisation and storage blog here for more interesting and informative posts.

 

Technorati Tags: ,,
Share Button
  • http://davtechonline.com Dave

    hey mate,

    interesting read, but you may want to link this from part 2 – i had to manually enter the url to get here ^_^

    so, whens part 4 out? i would like to see some performance benchmarks as we are currently rolling one of these out and having massive read time issues..

    • James Pearce

      Hi Dave, many thanks for the feedback. I’ll sort the links now. Part 4 is pretty much done and should be up very soon :)