What would cause this type of error? I am requesting HD,sym,10; multiple times. I request a symbol, wait for it to complete, then request the next symbol. Sometime between 500 and 1200 symbol requests, I'll get the error.

When I was running this, I'm on a network behind a Cisco IOS firewall. I don't know if that has anything to do with it, but I thought it might make a difference. I usually run my ap on my DSL, but has been down for the last week, so am using it on a different computer on a different network.

I am using the trial 4.0.0.3 IQFeed version, which appeared to work well in the other environment.

Sounds like your missing a reference object...

Make sure all implicit or explicit external references are present or consistent with your path declarations and pointers...

I don't think this is the problem as I get data. After a few hundred requests, I start to get the errors. I think it has something to do with the firewall configuration as I was able to connect the laptop to a dsl line, and things ran much better. Not perfect, but better.

I'm hoping a developer might be able to tell me what that 'socket' error message might indicate so I can go back to the firewall people and see what they've done.

I believe there some sort of control process running that attempts to identify users that are "abusing" the history system. When too many requests are received within a certain timeframe, the connection gets broken and needs to be restarted.

Note that I have never seen any official confirmation of this -- this is just my experience after using the feed every day for more than 2 years.

I'm having a similar problem. The historical database is so SLOW. It used to be much faster. What happened in v4.0?

This may be the same issue discussed in other threads about making too many request via the socket interface for history data. We have corrected this in a test build and it has been confirmed by a few users. Contact developer support via email if you would like to test this build as well.

dn,
As for the slow requests, we haven't heard this from others. You didn't specify what requests are slower in 4.0. As documented in the last developer notice, if you are requesting daily data and request ALL data by specifying "0" for number of days, you are now getting 3 times the data because we increased the amount of historical data available. This isn't a 4.0 issue, but it is something we changed a couple weeks ago. It will take longer to get the full historical data on a symbol since there is more data. You can change your request to pull 1000 datapoints (which was the max we had available previously) and you should see it just as fast as it used to be.

---

Jay Froscheiser
DTN - Trading Markets

Jay,

The test release to which you are referring, is that the 4.0.0.3? If so, I was getting the !ERROR! messages with that release.

Do you have a subsequent release I can test?

4.0.0.3 is the one I was thinking. Your issue must be different, as the .3 version has corrected everyones socket issues with multiple intraday requests. We may need to dig into the daily requests and see if there is something different in there. I don't see how the issue you are seeing could be related to a firewall. If it was, all request should have the issue. When will you have your DSL back up so you can see if it was somehow related to the network or configuration?

---

Jay Froscheiser
DTN - Trading Markets

If needed, I can send you another version with more logging in it. This would help us to see what the socket error is. Thanks!

Natalie

---

Natalie Hannan DTN Market Access, LLC.

Yes please.

I just sent a test build to your email address. Thanks!

Natalie

---

Natalie Hannan DTN Market Access, LLC.