Migrating from Cumulus 1 to MX

Like other longer pages in this Wiki, the content is divided into sections, if you are interested in a particular issue, find it in the table of contents, rather than reading whole page!

Introduction

This page was inspired by this update from Cumulus 1 forum topic. That post was made in January 2020, and therefore relates to Version 3.3.0, which was the MX release that was in use at the time.

  This page was last partially updated for the MX release available in June 2021; that is no longer latest!

Appeal to contributors: Please work out any updates needed for this page, so it is kept correct for more recent releases!


You might want to read the related page at My_migration_from_C1_to_MX which describes how this worked in practice for release version 3.4.5 (build 3069).

Should I migrate to MX or not?

The legacy Cumulus will never be updated, and as more people migrate to MX there will be fewer people to help you if the legacy cumulus software gives you a problem. MX has a major advantage it runs on more devices, and for example a Raspberry Pi computer is far cheaper to leave on all the time to run MX (knowing it will never be rebooted by Windows updates). The current developer of MX offers support, and other people are also gaining expertise, so if you have a problem you will get help.


Of course MX started with very little of the functionality that the original Cumulus offered, but over time MX has gained more and more functionality. Features in the original Cumulus have been gradually added to MX, but also MX offers a lot that is not in the original Cumulus. See Compare C1and MX page for more on this. Unfortunately, there is no list anywhere of all features in Cumulus 1. About_Cumulus does not cover all features. Nor is there a list of features still to be introduced into MX. Therefore it is hard to help you understand whether functionality changes between the flavours will suit you.

There are Cumulus users who tried MX, then have migrated back (from MX) to the legacy product; either on a temporary basis (to use a missing feature before they return to MX) or in a very few cases decided to stay with Cumulus 1 (because they prefer it). It is worth noting here that the legacy Cumulus is a stable product with very few bugs, but MX is constantly being developed and thus likely to contain bugs (these obviously vary depending on release as some are fixed and others appear). To put this in context, MX now has a huge amount of functionality and it is difficult for the developer to test all the different features.

Cumulus 1 creates a special log containing just temperature and humidity values, this is NOT available in MX.

Although from version 1.7.5 onwards the original Cumulus counts air frosts and gale days, this feature is not in MX.

The This Period screen that is a popular feature of the original Cumulus software as it tabulates data simply, is not available in MX.

Migrating from original Cumulus to beta MX

The beta builds (version 3.0.0) of MX were designed to make it easy to migrate. This section is based on Steve Loft's wording taken from the support forum with minimal alteration for its new context.

You can move between [Cumulus flavours] fairly easily, but you should really read all the guidance.

  • In particular if you use decimal commas with Cumulus 1, then you MIGHT have issues when MX tries to read existing log files.
  • [For separator characters in dates and lists] whilst Cumulus 1 only takes settings from the control panel in Windows,
    • MX running in a Windows environment, takes settings from the locale you specify in a parameter when starting MX, or the Control Panel in Windows,
    • MX running in a non-windows environment takes settings from the locale you specify in a parameter when starting MX, or the default locale in your Mono installation
    • MX [may] take some [settings from] interaction between the two [the specified locale and the default locale], and MX may struggle to read file [lines] created by Cumulus 1 if the MX locale is not precisely same as settings when that file [line was] created.

While MX was in beta, there was limited documentation about what features were included and why, there was some fault reporting and a tracking list that showed when some of those issues were fixed, also initially the documentation on how MX had implemented features it did have was very sparse. The lack of a list of features in Cumulus 1, meant it remains difficult to track which Cumulus 1 features are or are not implemented in MX. Steve Loft said parts of MX were simply machine code level copies of parts of Cumulus 1 functionality, and parts were trying to offer better functionality, but he never said what was included in these two categories.

There was a list of enhancements requested by users for Cumulus 1, but that list was deleted before work on MX started. Despite that, it does seem that some features that were on the now lost list of enhancements for Cumulus 1 that never got implemented in Cumulus 1, have been implemented in MX, although again there is no definitive documentation.

How to install MX

The basic instructions for installing MX (see MX on Linux, MX on Windows OS, or Raspberry Pi Image, as appropriate for fuller instructions):

  1. Download the MX distribution (from Software for latest version or from https://github.com/cumulusmx/CumulusMX/releases if you want an earlier release version)
  2. Unzip into the location where you are going to install MX.

'Engine' and 'Admin Interface' for MX

When you run the original Cumulus software it displays a screen (see Cumulus_Screenshots#Main_Screen, and from there you navigate to other screens to view data or change settings.

MX is different, it consists of a stand-alone 'engine' which performs the reading and logging of data, uploading to a web site etc. This 'engine' is a command-line/terminal/console application which has no user interface. It does write diagnostic information to a diagnostic log, but many people run MX on a device that has no monitor and so the terminal output (if any) is not monitored.

When you successfully start MX, the engine is running, and it continues, until it is terminated by control C (or its equivalent in a Mac environment). (You can also run MX as a service, that has different ways to start and stop, discussed in the links given in previous section).

The separate admin interface(unfortunately this is called various names in the support forum, including "user interface", "dashboard interface", and "admin interface") is provided by virtue of the engine acting as a web server. You can view the admin interface by typing the URL of the built-in web server into your browser, either on the same machine, or on a separate machine sharing the same local network.

If you install MX on Microsoft Windows, then a few extra one-off steps are needed to allow this web server functionality:

  1. Windows Defender has to be told to allow all for Cumulus MX.
  2. Typically for other security packages, select "CumulusMX.exe" then right click and select "Change File Rating to Trusted"
  3. In any "Firewall" package, add port 8998 as one that was permitted.
  4. Using "command prompt as administrator" type netsh http add urlacl url=http://http://192.168.1.64:8998/ user=\users, the response "URL reservation successfully added" means it worked. This command is apparently to allow all users to bind to port 8998 (i.e. that used for the Cumulus interface). This also means you don't have to run the engine (CumulusMX.exe) in an administrator user, nor select "Run as administrator" from right click menu on the shortcut, nor set the properties for any shortcut to run in any special way.

The default URL if the browser is on the same machine as MX is http://localhost:8998/. Newer releases of MX will tell you an IPv4 address to use, such as by typing "http://192.168.1.64:8998/".

Equally, if "localhost" is already in use for another web server (that you already run on your device), unless you can differentiate purely by port number you may need to use the correct IPv4 address as above, even on the same device.

For security reasons, the admin interface should not be accessible via the public internet.

Migrating existing data files from legacy Cumulus to a recent build of MX

If you want to continue using the same weather station, and you are not moving to a different home, you will want to maintain the data you collected using Cumulus 1 in your MX installation. This is what is meant by migration, and is the subject of the rest of this article.

File names

Note that if you run MX on a UNIX based operating system (e.g. Linux or Raspberry Pi OS) all file names are case-sensitive, please read documentation to see where capital letters are required in those file names. Be aware that Wiki pages change first letter to a capital even when a file that must be all lower-case is being described.

Line terminators

If you are not running MX on a Microsoft Windows computer, ideally you should change the characters appearing at the end of every line in any file you move from Cumulus 1.

I say ideally, because although Microsoft is fussy, and determined to be different by insisting files match its choice, most other operating systems can tolerate different line terminators.

With Cumulus 1 (or MX) on Windows, each line in every file ends with both carriage return and line feed.

With MX on a Mac, each line should end with just a Carriage Return.

For all Unix-based operating systems (Linux, Raspberry Pi OS, and other variants), each line should end with just a Line Feed.

To change end of line characters, run each file through an editor designed for programmers. Various editors are available but "Notepad ++" is one that is popular on Windows, but can run on other operating systems. In its Edit menu, choose EOL conversion.

If you use PHP Hypertest Pre-processor anywhere on your installation, normally that is written for "Line-feed" as line terminator and any "carriage return" in your files may mess up the content unless you code in a "trim" to remove it. Hopefully, if you use PHP, you are technical enough to realise you may need to edit the code depending on which device it is being run on!

Configuration file

MX's Cumulus.ini has different content to the legacy cumulus.ini.

Cumulus 1 can recognise, in some circumstances, "cumulus1.ini", and other variants, not just "cumulus.ini".

MX only recognises "Cumulus.ini".

The old approach for migration, was to copy across your existing file, and let MX ignore all the parameters that do not apply to it. To add the parameters that MX does need, you would then go to the MX_Administrative_Interface#Changing_Settings pages and work steadily through ALL the options. For any parameters that were not set by the admin interface (and there were many "read-only" parameters in early releases of MX), one assumed these also existed in Cumulus 1 and so were already in your file.

At releases like 3.3.0 and 3.6.0, amongst others, there were changes to "read-only" parameters for MX. These could not be adjusted through the admin interface, and needed to be entered manually as listed on Cumulus.ini (MX 3.0.0 to 3.7.0) page. If you choose to migrate to an old MX release, to simplify some data file issues listed below, then refer to that page for the new parameters, and refer to Cumulus.ini (Cumulus 1) if you need to check if particular parameters were used by the legacy software. However, you should reuse your old file, and just use the admin interface to work through all settings pages and finalise the new settings.

Substantial changes to changes were made from release 3.8.0 onwards, hence Cumulus.ini is a new page applicable from that release. More changes were made in 3.10.0, and subsequent releases, as settings that used to have to be made directly in file were gradually moved to be advanced settings controlled in the admin interface. From release 3.12.0,many more new settings were added, more settings were removed, and some parameters in the file were renamed.

Consequently, if you are migrating from the legacy software to MX now, it is best to rename your old 'cumulus.ini file so it is is not seen by MX. Let MX create a new configuration file with just the parameters it needs using the various [MX_Administrative_Interface#Changing_Settings|settings pages in the admin interface]]. That will ensure you don't get muddled by parameters used for Cumulus 1 (but not for MX); and you do have all the parameters you do need, set correctly.

Settings

The settings pages in Cumulus 1 and MX work differently:

  • for Cumulus 1 you choose to save changes by clicking OK,
  • for MX changes are only saved when you click a Save button if one is provided.
  • If there is no Save button anywhere on the screen (as in Extra Web Files) then the setting is saved when you move to next field/line.

Be aware that you should restart MX after changing settings, as many settings are only read from file when MX starts.

Start date

When you first run any Cumulus software (whatever flavour) a parameter is added to the configuration file that documents the date Cumulus was first used.

For Cumulus 1, this parameter appeared in two places on the example web template for all-time records, but was otherwise ignored. Thus if you decided to import into your data logs readings from before you first ran Cumulus, the legacy software would be able to find those earlier records.

CumulusMX.exe uses this parameter to determine the first standard log file to start reading from, it will ignore any log files for earlier dates. Thus if you were to migrate your Cumulus 1 configuration file into MX, you should check the "StartDate=" line in the '[Station]' section is correct for your earliest data before you let MX read this configuration file for first time. If you need to make any edit, ensure you stick to exactly the same date format.

From release 3.10.1, MX allows you to edit this start date within the admin interface. It is "hidden" as an "advanced" setting, with a strict danger warning, but it is there!

Station connections

You can skip this sub-section if your weather station connects to mX either by uSB or by wireless connection.

If your weather station used a port to connect to Cumulus 1, that port was set on the settings screen as a number and stored in Cumulus.ini_(Cumulus_1) in the station section as a parameter in the format Port=n.

In Cumulus MX, as it runs on various operating systems, the port is specified using text (instead of a number), again you select it within settings, on Station settings page, but within Cumulus.ini in the station section the parameter is in the format Comport=tttttttt.

If your old parameter had a value of 3, and you are still using Windows, the new setting would have value of COM3.

A typical parameter value for other serial connecting devices might be "/dev/ttyUSB0".

RG11 Rain gauge

If you use a RG11 rain gauge:

  • Replace: RG11port=n and RG11port2 =n (Cumulus 1) where n is a number, by RG11portName=xxxx and RG11portName2=yyyy (Cumulus MX on Windows) where the value is a string with values as per previous paragraph depending on device on which Cumulus MX is running.

From release 3.10.1, there appear to be other new parameters, not yet documented.

Strings.ini

This is another configuration file, but it is an optional one. Please remember that the Microsoft Windows Operating System is case insensitive for file names, so "Strings.ini", "STRINGS.INI", and "strings.ini" are all treated as the same file by any Cumulus software.

If you install MX on another operating system, then the file system is case sensitive, in this case MX will only recognise "strings.ini".

If you have not created a strings.ini file in your (leagacy) Cumulus top level folder, then you have no file to move to your MX installation, and you should skip the rest of this sub-section.

The contents of the samplestring.ini file you get in your MX release distribution varies depending on the release you have downloaded.

Check your existing strings.ini file against the samplestring.ini file in the MX distribution you have. If the attribute names (left hand side of the equals sign) match for the parameters you selected to include in your strings.ini, then you can reuse your existing file. If your file includes attributes that are no longer in the MX samplesting.ini file, then you will need to edit your strings.ini file that is placed in the folder containing CumulusMX.exe.

NOAA style reports

The generation of reports is an optional feature, if you have never used it your (legacy) Cumulus Reports folder will be empty, then you have no files to move to your MX installation, and you should skip the rest of this sub-section.

Please see Reports_folder for full information. Cumulus software creates reports, it does not edit existing reports, so migration is simple. Just copy the contents of the Reports folder in your original Cumulus installation into the folder in the new installation.

For those of you who are more technical:

  • files created in Microsoft's Windows Operating System use two characters (carriage return and line feed) to end each line, while all other operating systems use a single character (line feed in most Unix derived systems). Apple Mac are again different in using just Carriage Return. This should not cause any problems.
  • files can be encoded (how individual characters are represented by binary codes) in different ways. There is more about encoding at Reports_folder#Encoding, the relevance here is that if your MX settings and Cumulus 1 settings use different encodings you may find some characters (e.g. degree symbol) do not appear correctly when viewing some of your reports.

"data" folder

You should copy all files in the data folder from your legacy Cumulus installation to your MX installation if you want to be able to see past data.

There are some complications:

Weather Diary

As explained on weather diary page, the Cumulus 1 weather diary uses a different file to MX, and indeed takes a very different model for how to store the information.

Therefore don't copy Log.xml file into the data folder of your MX installation. You will need to manually use the admin interface Edit menu to access the new diary and add your previous entries one by one.

Please note the editing interface has been changed from release 3.10.1, but the page linked to above may not have been updated. A major change in the upgrade relates to handling of time-zones.

Please also note that the Cumulus 1 weather diary permitted multiple entries to be stored for an individual day, but the MX implementation only permits a single entry per day. Also the configuration file defines a time when the entries stop applying that may not match the rollover time for other weather measurements.

Extreme Record files

The Correcting_Extremes page explains how extreme records are stored in the Category:Ini Files discussed below, with the changes to all-time records being tracked in Alltimelog.txt.

You will find that MX creates at least one file that was not in the older Cumulus (monthlyalltimelog.txt).

the .ini files

Your Cumulus 1 .ini files can be read in all release versions from 3.0.0 to 3.5.y. Steve Loft did make two changes in his beta MX, these mean your .ini files may look strange while they have some entries made by Cumulus 1 and some made by MX:

  1. all newly stored values will use decimal points in these files, (i.e. any decimal commas valid in the legacy software, are not used by MX).
  2. all newly stored time-stamps will use the ISO specification (ISO 8601 Data elements and interchange formats – Information interchange – Representation of dates and times) so new dates have year first, the parts of the date are separated by hyphens, and new times use colon (:) between hour and minutes (i.e. the regional settings that the legacy software adopted have been abandoned by MX)
  • MX releases from version 3.0.0 to 3.5.4 can read all the date/time entries in any file whether in the Cumulus 1 format or in the MX format.
  • MX will change any dates to the ISO format only when it needs to update that particular date/time.
  • Consequently, within a single file both formats will co-exist, you may see lines using Cumulus 1 format for the extremes that have existed for a while, and Cumulus MX format for any new extremes.

If you used decimal commas in your Cumulus 1 installation, you will make the migration to MX easier, by an edit of each ".ini" file to change those commas to full stops.

Equally, if you had any times not using a ":" as separator, you will make the migration to MX easier, by an edit of each ".ini" file to change those characters into colons.

MX releases at versions 3.6.x onwards

Migration from the legacy Cumulus to MX has become harder as MX has been developed for two reasons:

  1. It appears that releases at versions 3.6.x onwards have become more fussy about existing content in these files due to changes in the source code.
    • Effectively, the format of existing entries is expected to be same as any newly created entries, i.e. matching current locale settings
  2. The addition of Feels Like temperature and Canadian Humidity Index (Humidex) to the .ini files
    • Neither of these was stored by the legacy software or by earlier MX releases
    • The technique for adding this is described on Calculate Missing Values page

Despite this, most people migrating from Cumulus 1 to the newer MX releases, are doing so successfully, so it works for at least some old files.

Historical background to dates/times

When Steve Loft designed his original Cumulus (1), he had no experience to draw upon as to the best way to treat items like dates. He wrote the software originally just for his own benefit and did not need to worry about time zones. Subsequently as he enhanced his software to make it usable by others, he faced many issues on how to cope with different time zones, and different weather stations having different sensors. In Cumulus 1, he basically focussed on compatibility by keeping to his original design for the data log files (both those ending in .ini and those ending in .txt, and only adding extra fields to the end.

When Steve Loft took a new look at the data log files for Cumulus 2, he started with a new design, the principal change was that he decided to use UTC for all fields in them that reference dates and times. Steve struggled with Cumulus 2 largely because he was (at that time) not familiar with the C# language he was later to use for MX. It is fair to say that conversions between local time and UTC did contribute to his failure to get Cumulus 2 to provide the functionality he had offered in the original Cumulus.

When Steve Loft designed Cumulus MX, he was able to learn from his experiences with both Cumulus 1 and Cumulus 2, so he decided to use dates to an ISO specification (ISO 8601 Data elements and interchange formats – Information interchange – Representation of dates and times), but in the local time zone of the particular user, and therefore log files are not backwards compatible.

dayfile.txt

MX (except early releases) reads the whole of this file as it starts running, so all lines must use exactly the same format.

date format issue

For MX (from release 3.5.4 onwards), the date separator specified for the locale when you run MX must be used in all lines of this file. Please see Amending_dayfile#Date_Separator_in_MX for more detail of what MX expects.

Cumulus 1 allows the date that appears in the first field of each line to use any character (except space) to separate the day of the month, from the month, and from the last two digits of the year. Consequently "29/03/88", "29-03-88", and "29.03.88" are all acceptable in the legacy software (and indeed in some early MX releases).


  • See MX_on_Windows_OS#Parameter_for_changing_Locale for how to specify locale if you are running on Microsoft Windows (by default the locale settings are taken from the standard Windows settings application or Control Panel).
  • If you are running MX in a Linux operating system, the locale parameter can be used, but the default locale is determined by the mono-complete package, and that in turn depends on your device settings.

So when you migrate "dayfile.txt", run it through an editor designed for programmers. Many such editors (e.g. Geary, NoteTab Light, Brackets) are available.

For example "Notepad ++" is one that is popular. That editor has a "Search" menu, with a Replace... option within it. In that option, there is a "Find what" box, a "Replace with" box, and a "Regular expression" selection.

Suppose your latest lines use either "." or "-" as separator, and that is what MX expects, but you have some older lines using "/" as date separator. Since you know "/" does not appear anywhere but in the dates with an older format, putting "/" in find what box and either "-" or "." (as required) in replace with box will sort you out after you hit "Replace all".

It is all a bit technical, if MX expects "/", but you have "-" in some older lines. The complication comes because "-" may be used in value fields too, so you need to find a way of specifying a minus with two digits before it and two digits after it. For this correction you need to select Regular expression and then set Find what to "^([\d]{2})-([\d]{2})-" and Replace with to "$1/$2/". Please check this, as this was copied from a forum post by Mark Crossley and I have not verified it.

time-stamp format issue

The dayfile.txt contains time-stamps following any value that represents a highest or lowest in the day.

All such time-stamps must be expressed quoting hour and minutes, with a separator. The use of seconds is not accepted.

The legacy Cumulus was not fussy on the character separating the minutes from the hour.

MX will only accept colon ":" separator, all dayfile.txt time-stamps must be in "HH:mm" format. You will need to edit your old lines if any use a different separator.

value format issues=

If you moved your Cumulus 1 installation from one windows pc to another, it is just possible that you might have a mix of "decimal comma" and "decimal point" in your values, or you might have changed the field separator (normally ";" or ","). Again, these must be consistent in all dayfile.txt lines for MX, and must match what is defined in the locale used.


web server

Some people have been slow to migrate from the legacy Cumulus to MX because any web pages designed using a Cumulus template file that works with the legacy software will not work with MX. It is not appropriate for this page to explain how to edit your Customised templates, but it can briefly cover the different approaches for default web pages.

From release 3.10.1, the new default web pages have a totally new look (designed by Neil Thomas), and include responsive code allowing them to look better on wide screens and on narrow screens (such as those on smart mobile phones). MX creates a number of Category:JSON Files that are used to convey the data from your local MX installation to the web pages installed on your web server. Mark Crossley's Steel Series is used for the MX gauges page. A single "use default web pages" setting sets up the settings required, see New Default Web Site Information page.

Steve Loft's Cumulus 1 came with a set of example template files (designed by Beth Loft) that could generate web pages to upload to your web server. It also produced a set of images representing standard graphs that could be uploaded to your web server and shown on the default "trends.htm" page, a set of images representing wind speed and direction that could be shown on the default "gauges.htm" page these were based onWeb Dashboard Components for FreeWX and FreeWX-Wi (no longer available) and mixed images generated by Cumulus with plots drawn using JavaScript routines (before "Canvas" functionality)., and a moon image that was shown on one web page.

Library software

Any non-technical person can skip this sub-section!

MX uses a number of library software utilities like Highstock, jQuery, boot strap, and others, see MX_Basic_info#Library_software_for_your_web_server. Be aware that MX determines the versions of these it seeks, and they may not match the versions needed for anything on your web server that is not supplied in the MX release distribution. A browser tries to reuse components that are already loaded, so there is a possibility of the wrong version being loaded.