Netexpertise

I’ve noticed backups are getting slower and slower over time going from 500 to 2500MB/min throughput.
I run backups on 2 dedicated RAID0 disks hosted on a iSCSI MSA array first, and duplicate them on tapes. Since I want to use the same array for a new database project, I want to measure and tune access to SAN disks from Windows 2003 Server and Linux Redhat 5.
 

Environment

4 disks have been set up in a RAID 10 set.
All interfaces are forced in 1G speed on the SAN and the server side.
Servers are connected on to the SAN via 2 interfaces in a load-balanced modethat gives a max theorical throughput of 250MB/s (a bit less considering frame headers).
Jumboframes have been disabled, they’re not supported on the switches I’m using in this setup. They could give slightly better performance.
I also tuned the read ahead cache but I got no significant improvement.
 

Results

The tests were conducted with a simple dd on Linux and Diskbench on Windows.
 

Reads		MB/s	IO/s
Windows		70	1100
Linux		110	900

Writes		MB/s	IO/s
Windows		140	2100
Linux		140	400

Optimizations

There was no optimization whatsoever on the Linux platform. The partition was formatted with default ext3 filesystem on a LVM volume.
On the Windows side, there was no fragmentation since we started off with a brand new drive. Fragmentation degrades performance indeed. The disk is formatted in NTFS with default 4k clusters: Increasing the cluster size does not seem to be of much impact.
The partition has been aligned with the physical disk in diskpart to correct Windows 2003 Server caveat. This led a 5MB/s increase in the maximum disk throughput.
 

Then?

2 questions:
Why the HP MSA 2012i writes faster than reads!? Write cache?
 
Why Windows read operations maximum bandwidth gets stuck to 70M?
Throughput was roughly improved by 5M with above Windows optimizations but we seem to now reach a cap limit. I get the same results with whichever RAID0, RAID5 or RAID10!!

Tags: 2003 server, Array, HP, iSCSI, linux, MSA, performance, redhat, SAN, Storage, Windows

You can open any kind of document on Windows from an AS400 5250 session or a CL.
The type of file can either be an image or a picture (jpeg, tiff, bmp, png), a video (avi, mpeg), a PDF, an office document (word, excel, powerpoint), and even an Internet link (URL).
 
AS400 provides the strpccmd command in this purpose:

strpccmd pccmd('explorer C:\répertoire\fichier')

 
The document will open with the default application as if when double clicking the file in explorer; Eg: Acrobat reader for PDF files.
 
Some say you don’t absolutely need the ‘explorer’ command to the document to open but it fixes some issues:
- You do not to get a DOS window displayed
- URLs launch with the default browser, Internet explorer or Firefox (This wouldn’t work otherwise)
- Documents with a space in the path aren’t buggy anymore

Tags: AS400

Based on advice from a Microsoft techie, I decided to set up DFS (Distributed File System) on a few remote sites about a year ago. Sceptical at first, The rpoduct is so disappointing that I decided to warn people who haven’t made their mind up.
 
Here are 6 good reasons NOT to use DFS:
 

No Defragmentation

After defragmenting, a new replication occurs generating a huge traffic that could max out your lines bandwith for a few days if the folders size is big enough.
I haven’t tried but I wonder if the same thing wouldn’t happen when creating/removing the disk index?
 

File Locking

There’s no such thing as inter-site file locking since there 2 local copies of each files (or more).
 

Concurrent Access

Drawback from the previous point, data may be overwritten. If someone leaves a file open the whole day and saves it before he leaves the office, all changes made from another replication site will be lost.
Therefore, DFS is advised for write-protected files or files that can be modified from a single site.
 

Failover

If one of the DFS server becomes unavailable on a same site, some users are still redirected to that server and may lose access to their files. Wasn’t it one of DFS goals?
Of course, users are still able to manually select the DFS target but most of them probably don’t know the trick. The problem does not happen on all folders as a DFS server is randomly selected for each one of them. There’s also a cache that defaults to 1800 seconds, that’s 30 minutes.
Solution: Upgrade to Windows Server Enterprise which provides clustering and rising the price of your license.
 

Diagnostic Tools

No graphical tool (yet Microsoft speciality) is available to monitor what’s being replicated.
There are indeed those scripts connstat.cmd and iologsum.cmd written in Perl (Really!?) provided with Microsoft support tools. You first need to relocate the files in a path with no space or you’ll get the following error:

Can't open perl script "C:\Program": No such file or directory

Microsoft have posted a bug report on their website back in 2007 rather than fixing it!
Try to replace “@perl %~dpn0.cmd %*” by “@perl %0 %*” around the end of the script…
Script usage is not very intuitive and information not always relevant in my opinion. Usage with a practical example is available on Microsoft indeed.
 

Bandwith Throttling

Bandwith throttling isn’t possible as far as I know on Windows 2003 Server…
As well as that, to stop an ongoing replication, I stop the DFS service. Nothing happens, replication still goes on. I had to run ‘net stop ntfrs’ to stop the flow. Unfortunately, this also denies Active Directory from replicating if you’re on a domain controller.
 
In a nutshell, DFS has so many gaps and restrictions, it becomes difficult to use it for anything but content publishing across an enterprise.

Tags: DFS, replication, Windows