repeated missing data for one or several servers

[Rambo]

Hi,
One or more servers for historical data ($TOPIX-TSE) are not functioning properly. There have been repeated occurrences of missing data in recent weeks. I have reported the problems through the bad tick report webpage. But for some reason, the problem has not received any attention. As a result, I have to spend time each day, after logon, to check whether there is a data problem. About half of the times, there is a data problem; I have to logoff and logon several times, checking data each time, until I get connected to a server without missing data.

The following is a list of the problems I have reported. Could you please take a look at the issue and help to have the problem resolved?
The software I use is Esignal 10.2.

reporting date, date and time of missing data
21-Nov, 11-Nov data aft 13:00 missing
23-Nov, 10-Nov data aft 13:00pm missing
23-Nov, 12-Nov one 10m bar outside trading session at 12:21.
23-Nov, 17-Nov one 10m bar outside trading session at 12:21.
23-Nov, 17-Nov 15s data missing, 9:15 to 10:45
23-Nov, 20-Nov 13:01, one 10m bar missing
25-Nov, 21-Nov 15s data after 13:00pm missing
29-Nov, 25-Nov interval data missing from 13:00pm to 15:00pm
1-Dec, 25-Nov Interval data missing from 13:00pm to 15:00pm

[Rambo]

Missing data - today's (Dec 1) data sheet

today's (Dec 1) data sheet: Afternoon data for Nov 25 are missing.

[Rambo]

missing data - 15s, Nov 28 after 13:00

missing data - 15s, Nov 28 after 13:00

[ScottJ]

While we look into it, are you sure some of these gaps are NOT related to exchange holiday or half-holiday hours? Not familiar off-hand with that market but the cut-off times appear so clean it looks like holiday hours to me.

Thanks.

[Christine B]

Rambo,

thanks for reporting this issue. Please note that we have received your reports thru the bad tick reporter and we are currently looking into this issue.

As a work around the data is available in the Time and Sales window.

Please feel free to contact us directly if you see any problems.

Regards,

[Rambo]

Thanks for looking into this

Hello, Christine and Scott:

Thank you for taking a look at the issue.

There were no holidays or half trading days among the dates with data gaps listed here. In fact, I was able to eventually connect to a server without data gap on each occasion.

Thanks for the suggestion of the time and sales window. Unfortunately, it is not very helpful for my purpose, as no indicators can be applied to time and sales window.

I clicked on the “contact us” link. It launched an email page, but with a website address – Esignal’s contact us webpage. I suppose you are referring to LiveRep provided on the page.

With a couple of exceptions, my experience with the LiveRep on technical issues has been good. On the other hand, with a couple of exceptions, my experience with the LiveRep on data issues has been bad. The answers I got most of times were something like “well, we get our data from the exchange”, meaning, I guess, there is nothing that can be done about the problem. For this reason, I stopped using the LiveRep for data issues some time ago.

The attached is what happened today – historical data before Nov 26 were missing. Again, this did not happen to all the servers. I was able to eventually connect to a server without the problem.

Many thanks for looking into this.

Regards,
Rambo

[Christine B]

Rambo,

I am sorry to hear about the bad experience whilst using LiveRep; I was actually suggesting you call us when you see this problem. And as I don't know where you are based, I provided the link with our office phone numbers and opening hours.

We seem to have different issues here; The first one is that some servers seem to be missing data while others don't. If you contact us in real time eg phone or via live rep, our support should be able to locate which server you are connected to and pass on this information to our Operators for further investigations.
Ideally any current data issues should be troubleshooted in real time so it would really help if you can contact us as soon as possible if you experience any problems.

The 2nd issue here is that although we do have the data on our servers (T&S as well as on tick charts) we seem to be missing tick bars (1t, 15s, 20p etc) on certain time periods.
We were able to replicate this on different versions and have forwarded it to Product Management for further investigations. Unfortunately I am unable to provide you an ETA as of when it will be fixed.

Regards,

[Rambo]

a possible cause?

Christine:

Thanks for the clarification and for looking into the other data issue I raised in an earlier posting about 15s charts.

I understand it takes time to resolve these issues. I am glad that you are making an effort to address them.

For what it’s worth, here is a thought about one possible cause of the data gaps in minute charts. Earlier today, when I accidentally turned off the data manager, the charts stopped updating. When I re-launched the data manager some time later, the charts started to update again, but with a gap left in the charts for the period when the data manager was turned off. Is it possible that something similar could be happening to some of your servers? Something similar to the data manager for your servers may be experiencing regular breakdowns, resulting in data gaps.

For my problem, just re-launching the whole page would get ride of the data gap by getting all the historical data from the server instead of what are stored in my computer. If the data gaps on the servers are generated in a similar way, I hope there is also such an easy solution, assuming that your colleagues are alerted to the interruption.

Thanks again for looking into the issue.

Regards,
Rambo

[Rambo]

Please help, the problem is getting worse

The problem is getting worse. Today, $TOPIX-TSE data for Dec 5 afternoon are missing
for all the servers. I reported this to your Australia supporting desk by phone. They
confirmed the problem.

I understand it takes time to correct the problem from the source. But in the meantime, it
is fairly easy for you to write a short program that checks on the possible gaps regularly
so that you are alerted to the gaps and correct them promptly when they happen. Given
the high frequency of the problem, I would appreciate it very much if you could be
considerate enough for the users to make such a small effort.

[Rambo]

Data sheets

Attached is data sheet with the gap on Dec 5, from 13:01 to 15:00.

[Rambo]

No justification for allowing the problem to continue

The missed data were not restored until the afternoon session. The problem could have
been easily avoided by a few lines of program codes that check the data integrity
regularly, given that the frequency of the problem has been reported for some time
already.

Frankly, you do not have to care too much about your customers or the data quality, just
for the sake of making good use of whatever resources made available for your data
maintenance, such an easily preventable, yet disabling problem should not be allowed to
continue.

[Christine B]

Rambo,

thank you for the continuous information.

We are currently looking into this issue and it seems that there are issues with those servers which carry the data during certain market hours.

This is currently still investigation, hence I don't want to confirm it until we have done more testing.

Could you please send me a privat message including your username and your timezone when you see this issue.
I will look into it and hope we can find the cause of the issue so we are able to fix it soon.

Thank you for your patience.

Kind Regards,

[Rambo]

Thank you for your attention

Christine:

I started to report this problem by bad tick reporter about a month ago. It does not seem
to be clear how much longer it is going to take to have the problem corrected at the
source. While it is fully understandable that the testing and correction will need time,
would it be possible that before pinpointing the source of the problem, you take steps to
actively monitor the data and correct the problem promptly when it happens rather than
wait for users to report the problem?

It happens so frequently. It takes such a long time for the report from the user to reach
your data department – half a day was gone before today’s problem was corrected. And
it is so easy to write a small program to monitor the data. You probably need to run the
program only once on each server when it comes online every day. The monitoring will
also help you to map out the problem servers and the timing of the occurrences quickly.

I will send my user information to you by PM. But today’s problem happened to all
servers and it was duplicated by your Australia office. The other problem, added data
outside market hours, was duplicated by your London office a few days ago. And
obviously, I was not connected to all the problem servers at any given time. So the help
my account information can offer in resolving the issue is limited.

Thank you for your attention.

Best regards,
Rambo

[Christine B]

Rambo,

we are currently unable to always replicate the issue, hence it your account information/settings will hopefully help to identify the problem.

I can understand your frustration with this issue and I will speak to my colleagues so they can keep an eye on this issue more closely in the future.

Hopefully we will have more detailed information soon.

Regards,

[Rambo]

Data gap on Dec 8 appeared today

Christine:

A data gap between 13:00 to 13:41, exchange time, on Dec 8 appeared today for
$topix-tse.

Please be reminded that most times, the problem occurs only to some of the servers, not
all of them. So if you randomly connect to a server, rather than systematically test all the
servers, you may not see the gap.

In fact, when I reported the problem today, your supporting staff at Australia office did
not see the gap initially because he was connected to a different server. But he was able
to reproduce the gap after connecting to the same server as I did.

I am setting aside 30 minutes each morning to check the problem and report it.
Hopefully, I will be able to help you map out a set of the problem servers eventually.
Could you please kindly keep a record of the problem servers reported to you? Let me
know if you need the name of the supporting staff I talked to today.

Obviously, Esignal staff could easily do the checking each day as I am doing
and have the problem corrected when it appears. It puzzles
me why the burden ought to be borne by the user to check and report the problem, and to make
sure that the data are without the gaps each day, not mentioning the risk that many people
may be making trading and investment decisions unwarily with the impaired data.

Thank you for your attention.


Regards,
Rambo