Hi All,
What is the best tool or way to show the IT guy(s) that the network speed is the cause of the problems for TopSpeed files?
Any suggestions please?
Hi All,
What is the best tool or way to show the IT guy(s) that the network speed is the cause of the problems for TopSpeed files?
Any suggestions please?
Hi @Marius_Nel
Iām not sure, with the information you have given, that there is enough information for anyone here to be able to advise you or the āIT guyā on whatever issues are being seen.
So it might be best if you could provide answers to the following:
Hopefully with extra information someone may be able to help more.
Regards
Mark
I deal with this all the time and it can be a struggle since 1) what is Clarion/Topspeed 2) oh itās not using SQL 3) ISAM WTF??? I get this all the time. I usually use Process Monitor to show them all the file open\close and I/O and eventually they see the lag times. Good Luck. Be sure to let us all know how you did with this ābattleā
Markā¦
This one client complains about speed and the client is at the end of the discussions with the IT guy. SO I need to proof to him/IT Company that it is something with his network setup and not the system as he likes to points out to the client the whole time. (One of hundreds of other clients that works 100%)
The client complains that if they are more than one user using the system it is to slow to use, but if they log in on the one back office and use the system it works 100%. (The same system and or network)
Server is a windows 11 pro PC with SSD drive
Will check the Op Locks but I asked if that was off and they replied yes.
A lot of data corruptions lately.
Wired network running at 1GB
I need a tool or way to show them it is their settings and not the system.
Thank you Kevin, will give it a go.
we are running Windows 2019 servers and everyone logs in via RDP and access TPS files on different servers and we run reports and import\exports all day long. We have up to 100+ users active running different programs accessing files w/o slowness.
Thatās usually a symptom of oplocks, which allow client cacheing of data to speed things up. with 2+ stations oplocks are nullified. So thatās why 1 is fast (client cache) and 2 run at normal speed (no client cache)
Problem with oplocks is you can only turn it off if using SMB1, which in latter versions of windows not active, so itās a bit of a problem. It can be done, itās just painful.
You might try giving more time to background processes on the server system, thats a system control panel setting.
Other than that Iām not sure what can be done, other than go to ipserver or sql
i can only say I never managed to make app residing on the windows server share and using tps as db run well with multiple users running the same app at the same time.
Very simple ones - yes, but with just a bit more records, unfortunately not.
That was one of the reasons I moved to SQL. And.I still have some that need the conversion to SQL.
My advice - definitely use the IP Driver. It will improve speed on even the crappiest network - including wifi. And you wonāt see any more TPS corruptions either. But Oplocks must be turned off. It also helps explaining to MS-indoctrinated IT support guys when you mention the magic words Client/Server, data packets etcā¦
Mike
This fight started for me in 1986 when ArcNet and Token Ring were the kings. I have found choking is often effective. I have found that once you have proven your point multiple times over, it is your only option. And, it never ceases to amaze me that each generation (training, not age) of Infrastructure guys revert to the same arguments⦠Oh, well, what are you going to do???
Fast, short-term solution is to get an RDP server in there.
I have done both IP Driver and RDP. For short-term gains, Iād go with RDP.
Going the IP Driver route could potentially involve a lot more work and a whole lot of testing under multi-user conditions. I had to deal with file error #27 with just one user under IP driver and I had to workaround program lockups with multiple users.
In my previous job, I was the network guy and the app used TPS files. I never was successful in preventing TPS corruption. I tired a lot of different things. Speed was never a complaint. The app was always run in a RDP session and run locally on that server. I switched to SQL Server and never looked back. Never had speed or corruptions with SQL Server. All my work now used SQL Server. Also allows for multiple app platforms accessing the same databases. I have PHP (both backend and web) and Clarion apps accessing the same database. Works great.
BTW: I never had TPS corruption when the apps ran on a Novell Network.
Same reasons & results in long ago moving to PostgeSQL. The ease of creating aggregate functions/subqueries for summary data is reason enough, but many other advantages as well.
Fast, short-term solution is to get an RDP server in there.
I ran it on a RDP server and still had corruptions. When users left the app open and session closed or then they suspended a session and not close. Lot of UE errors can happen. I left the same RDP server in place and switched to SQL and corruptions/database errors went away.
RSchofield, do you happen to know if STREAM( ) command was used throughout the code in the app or if the .TPS files were being stored on a different server than the RDP server or on a NAS? I can imagine that either case could trigger problems. Otherwise, I just find it hard to believe that there were corruptions when running under RDP with the .TPS files on a drive within the RDP server. Although, I can think of a case where live backup software can cause a problem with a file system like .TPS where the files are constantly opened and closed and the backup software tries to squeeze in a back.
To measure the network speed, I use the free version of āLAN Speed Testā because the supposedly 1Gbps LAN usually runs at 100Mbps because of bad cable connections, sloppy LAN card setup, etc. It helps with the performance of Microsoft Access data files, so Iām sure you will get some useful data for TPS files too.
āYou canāt solve a problem until you can measure it.ā
Also, check out the oplocks settings and other advice offered above.
Word of warning: getting the IT guys to actually run LAN Speed Test may be a problem, because they donāt like having to fix their own setup problems.
No STREAM and app and TPS in same folder. No live backups either. I tried many things to try and fix the problem, but never did. Windows isnāt designed for ISAM files. Netware was great. The same app running on NetWare never corrupted the TPS file, Windows always did.
Hi All,
Thank you all for the feedback. I will use these to reply to the client and IT. Letās see what they say and if they can get it sorted.
I really appreciate all your feedback.
Iāve had an issue with slow multi-user tps file access on a network for a long time. Read all the comments about oplocks, SBM etc. and resolved to just live with the situation. Yesterday a client threatened to switch to another product due to the unworkable situation so I started to look for alternatives.
I noticed that the āEnclose RI code in transaction frameā option was checked. I unchecked this and set all the āIndividual file overridesā - āUse RI transaction frameā settings to āNo.ā
This has made an incredible difference to the multi-user experience on the network.
One screen with summary totals from a related file now runs in 4 seconds instead of 3 minutes!
Iām thrilled, but also concerned as Iām not sure if Iāve compromised on database integrity.
Looked in the Help - I donāt make use of the āSave Button Transaction Frameā or āProcess Transaction Frameā extension templates anywhere in the app, so I think I have just switched off superfluous processing.
Any thoughts?
Unmanaged switches, ie switches with no interface where you can set things like vlans, routing tables etc can bog down with too much data.
Too much data can just slow things down.
On Servers I used to look after I used to have dual and quad nic cards inside the server as this could be a bottleneck.
Cabling can be an issue, where possible also use shielded Cat6, if cable is trunked running alongside power cables, the magnetic flux from the power cables can interfere with the network cable itself. 3 phase can mess things up alot if the cable is not good. Check the cable termination is sound as well. You can get a bit of kit where you plug the cable into the device and it tests the quality of the cable in a variety of ways.
Managed switches also sometimes have basic cable testing facilities where they can even calculate how long the cable is to within a metre.
Cableās should be no longer than 200 metres as the time it takes for the signal to reach the other end falls out of sync with the window of time set aside for the signal to reach the other end. This then causes packet retries.
Packet retries can bog down a network very very quickly so having a managed switch which reports the data throughput and send out emails when high volumes are reached can be helpful.
Thereās just loads of ways a network can bog down and grind to a halt.
Thereās some thoughtsā¦