[Chris Gordon]

[Douglas Zhande] “1. I would like to know if dual port hba fc is faster than a single port hba , both with same same speed.”

Yes, you have twice the bandwidth to the array.

You have several ways of doing this:

1. Have an array that supports multiple paths to the same LUN simultaneously and multipathing software on the host that handles that.
2. Break your data across multiple LUNs on the array and zone different LUNs to different HBA ports. This means you need to balance what goes on which LUN yourself to get balanced/max throughput. This will look like many different disks on your host.
3. Similar to #2, but you use an LVM or something similar to stripe a filesystem across all of the LUNs present to the host.

Now think about why you want or need this. The premise would be that you need more than you can get out of a single HBA — the most you find these days are at 8Gbps. Do you really need that? Do you have an array that can handle that? How much money do you have? If you really think you need this, you are probably in a situation where you have (or should have) a Storage Administrator on staff.

[Douglas Zhande] “2. Is Qlogic adapter PCI – e 1.1 faster than PCI X 2 qlogic adapter”

Wikipedia has great descriptions of the different PCI versions and the bandwidth they provide. Combine that with the speed of the particular HBA and the more limiting component is your bottle neck.

[Chris Gordon]

In Final Cut Express, you’d use Apple Intermediate Codec for the HD material.

[Chris Gordon]

It’s been some time since I’ve had to use it, but I’m pretty sure you can just have it change the gamma but not re-encode. I meant to write up the details of what I did, but never got around to it. Next time I need to do the correction, I’ll be sure to document it for future use.

[Chris Gordon]

I’ve used JES Deinterlacer (https://www.xs4all.nl/~jeschot/home.html) in the past to adjust the gamma value of an h.264 file after render from FC. Pretty much leave all of the other settings alone and just adjust the gamma value.

[Chris Gordon]

You mentioned an online conversion tool.. Remember that the conversion between B (bytes) and b (bits) is simply a factor of 8. There are 8 bits in 1 byte. You can easily multiple or divide as appropriate and not need to hunt down any other tool that may give you doubt.

[Chris Gordon]

I’ll second that. I loath web mail — never found a web mail solution that comes close to a native client. Webmail is OK for quick checks from home or elsewhere, but I couldn’t use it as my main client. I’m also against the idea of “everything in a web browser”.

I’m very much a fan of providing open protocols and standards and letting people use the tool that they prefer or works best for them.

[Chris Gordon]

On my Linux and FreeBSD servers I do — that’s generally the way you configure samba. You’ll probably find relatively little specific to tuning on OS X, but should be able to follow, in general, the Linux notes for which there are many.

One of the most common things to adjust are some of the socket options, specifically the send and receive buffers. In the [global] section, you would have:
socket options = TCP_NODELAY SO_RCVBUF=8192 SO_SNDBUF=8192

Other things to look at is tuning your tcp stack. Again, there are a number of articles on the net about this. The best values are something you have to figure out for your specific use. I’m running with 9K jumbo frames (more because I can than I need to) and have these settings in /etc/sysctl.conf on my Mac workstation and using ttcp, I can get close to 900 Mbps (far more than I expected).
kern.ipc.maxsockbuf=500000
net.inet.tcp.sendspace=250000
net.inet.tcp.recvspace=250000
net.inet.tcp.mssdflt=8940

Now as you get into this, a couple of notes:
– SMB is not known to be the fastest protocol in the world.
– You’ll want to see if you’re getting limited somewhere else such as CPU or disk. Assuming you tested from the same exact machines with AFP and are now trying SMB, this would give you the same disk IO and network performance leaving CPU as the big thing to watch.

[Chris Gordon]

This isn’t so simple for an enterprise with a lot of security requirements not to mention legal and other similar challenges. For many companies there is a lot of proprietary information contained in emails (think intellectual property, merger/acquisition discussions, info that could contribute to insider trading, etc). Handing this all off to some outside company just won’t work.

My point is that a solution that works for one shop and is a great thing may be impossible or a “resume updating event” in other settings.

[Chris Gordon]

OS X uses Samba for its SMB server. There are many docs on the net and at https://www.samba.org on tuning SMB performance. This will require you to edit /etc/smb.conf.

[Chris Gordon]

Those Xeon CPU and all of that ECC RAM are power hungry. If you’re not giving the machine good power, you can have all sorts of issues including crashes, disk corruption, etc. Make sure your UPS is rated for the power draw of your machine — you may need something beefier than what your G5 needed.