Error Message Index

The first line is usually not the error#

Flow Config wraps unhandled failures in a generic dialog:

Something went wrong!
<the real message>
<the inner message>

"Something went wrong!" carries no information — it is the wrapper, not the diagnosis. The useful text is the second and third lines, and that is what to search for here. A ticket titled "Something went wrong" tells support nothing; the same ticket with the second line pasted in is usually answerable immediately.

Some areas wrap more specifically — detaching an event scheme set reports Something went wrong with the Event Scheme Set detachment, followed by the real cause. The rule is the same: read past the first line.

Infohub raises some of these; the platform raises the rest#

Many errors customers report are not Infohub's text at all. They are SQL Server, Windows, or .NET messages that Infohub passed through unchanged. That distinction decides where you look next, so each entry below says who raised it.

Errors Infohub raises#

The given key was not present in the dictionary. Infohub looked up an object by ID in a map that does not contain it — almost always during a template update, a model import, or a chart or form import, where IDs from a source model are matched to a destination model. Raised in several places, including template instantiation and the export repositories.

This is the single most durable error in Infohub's support history: it appears in tickets from 2022 through to this year. Individual causes have been fixed as they were found, so the version you are on matters. On 7.2.0, chart and form import were hardened against the specific case where source and destination ID spaces overlapped. If you hit it:

• Note exactly which operation produced it — template update, model import, chart import, form import. They are different defects with the same message.
• Full import reclassifies objects differently from a partial one; if the failure is on chart import, retrying as a partial import sometimes clears it.
• Send support the operation, the version, and the model export if you can. The message alone does not identify the cause.

An item with the same key has already been added. The mirror of the above: Infohub tried to add an ID that was already in the map. Reported against calculated event attributes. Same advice — the operation identifies the defect, the message does not.

Item type '<name>' is not valid (Parameter 'itemType') Raised by Flow Server when a chart or form contains a colour rule, filter, or repeating item bound to an item type that build cannot resolve — typically after a chart was imported from a model where that item existed, or after an upgrade changed what a chart component may bind to. Open the chart in Flow Config, find the component named in the stack, and re-pick its source item.

Multiple Matching Dependancies Detected Flow Config's dependency-mapping dialog, shown when a template update finds more than one candidate in the destination model for a dependency in the source. Note the spelling: the product's dialog says Dependancies, so searching for the correct spelling "Dependencies" will not find the product's own text.

It is a prompt, not a failure — Infohub is asking you to choose. The related message Could not locate a matching dependent. Check whether dependent has been released in the template or instantiated is the opposite case: no candidate at all, usually because the dependency was never released in the template or the instance was never created.

Could not find a limit type implementation for <type> in <package> <version> The limit plugin package registered against that limit type is not installed on the machine doing the work, or is a different version from the one the model expects. Deploy the package to that host. Expect this after moving an instance to a new server, or upgrading one component without the others.

Something went wrong using: Host: … Url: … Exception: … A Web Service or Netilion data source could not complete its HTTP call. The host and URL in the message are the ones Infohub actually used — check them against what you intended, because placeholder substitution happens before this point. The exception at the end is the real network or HTTP failure.

IDX10400 — Unable to decode… A licence key that lost or gained a character in transit cannot be decoded. Usually a key pasted from an email with a line break in it, or a key truncated by a form field. Ask for the key as an attachment rather than inline, and apply it unmodified. On 7.2.0 this no longer takes Flow Config's connect sequence down with it, so you can still open the dialog that replaces the key. See Licence Keys in Practice.

Something went wrong with the Event Scheme Set detachment. Execution Timeout. The CommandTimeout in the Flow.Config.Settings file is set to <n> seconds. See Timeout expired below — this is the one place Infohub names the setting for you.

Measure <id> '<name>' has <n> tag transform configurations (TagTransformIDs …), but only one is supported. Raised when a model is loaded and one measure carries more than one tag transform. A measure can hold a single transform; a second one normally arrives through a database-level copy or an older import. Before 7.2.0 the same condition failed every model load with An item with the same key has already been added, which named nothing — the new message names the measure and the transform rows so the extra one can be removed. See Tag Transforms.

Flow Bootstrap on <host> port <port> refused the registration: <status> <reason>. Flow Config asked the Flow Bootstrap service on that host to register the instance when deploying a platform, and the service declined — the text after the status code is Bootstrap's own reason. When it continues "The connection was sent with Windows authentication…", the instance uses a SQL Server login that has not been entered under System properties; enter the Username and Password and deploy again. The Config CLI's --register-with-bootstrap reports the same text and exits non-zero. See The Deployment View.

Errors SQL Server raises, that Infohub shows you#

Maximum recursion 100 has been exhausted before statement completion. A recursive query walked more than 100 levels. In Infohub this is a dependency walk — a measure's calculation chain or a model folder tree. Two causes, and they need different fixes:

• A genuine cycle. Measure A depends on B, which depends back on A. The walk never terminates and the recursion limit is what stops it. Find and break the cycle; raising a limit will not help.
• A chain legitimately deeper than the limit. Rare, but real in deep model hierarchies.

Reported repeatedly against tag dependency walks and against template metric updates. Fixes for the cycle case shipped in 7.2.0. If you are on an earlier version and the chain is genuinely acyclic, upgrading is the remedy.

Timeout expired. / Execution Timeout Expired. A statement ran longer than its command timeout. In Flow Config the big offenders are purging, deleting a model or measure, detaching an event scheme set, and opening a measure with a long history.

Flow Config's timeout comes from CommandTimeout in the Flow.Config.Settings file. The configured value is in minutes — Infohub multiplies it by 60 — and when it is not set the effective timeout is 60 minutes. Increase it before a large purge or delete rather than retrying the same operation. The Config CLI takes the equivalent as a command-line option; see CLI Option Reference.

A timeout is a symptom of size, not corruption. If routine operations time out, the database needs maintenance — see Database Maintenance and Engine Performance and Tuning.

Cannot insert duplicate key row in object '<table>' with unique index '<index>'. SQL Server rejected a row that violates a unique index. Reported against EventAttributeValue. The table named tells you which part of Infohub was writing. Capture the full message including the table and index name — without them the error is not actionable.

Transaction (Process ID n) was deadlocked on lock resources with another process… Two workers contended for the same rows. Infohub retries some of these automatically, so an occasional deadlock in the log is not necessarily a fault; a sustained rate is. Deadlocks during measure promotion and event-period writes were addressed in 7.2.0. If you see them steadily on an earlier version, that is the upgrade case.

A network-related or instance-specific error occurred… (provider: Named Pipes Provider, error: 40 — Could not open a connection to SQL Server) Often reported just as "error 40". Infohub cannot reach SQL Server at all. Check, in order: the instance name in the connection string, whether SQL Server is running, whether TCP/IP is enabled for the instance, and whether the port is open through the firewall. Nothing in Infohub fixes this.

Login failed for user '<account>'. The account Infohub runs as cannot authenticate to SQL Server. After a service-account change, this is the first thing to check — the Windows service account and the SQL login are two separate pieces of configuration and changing one does not change the other.

String or binary data would be truncated. A value was longer than its column. Usually a name or description pasted from elsewhere.

Errors Windows and .NET raise#

Error 1069: The service did not start due to a logon failure. The Flow Bootstrap service account's password is wrong or expired. Error 1068: The dependency service or group failed to start is a dependency, most often SQL Server not yet up. Both are Windows service-control errors; neither is an Infohub fault.

Object reference not set to an instance of an object. (NullReferenceException) A .NET null dereference. On its own it says nothing — what matters is the method name in the stack trace, which is why the log entry is worth more than the dialog. One known case, reported against ExecuteMeasureCollectionExpression, involved measure-collection expressions and was fixed in 7.2.0. Always capture the stack trace; see Reading Infohub Logs.

The added or subtracted value results in an un-representable DateTime. Date arithmetic went outside the range .NET can hold — in practice, an offset applied to a date that was already at or near the minimum, such as an uninitialised timestamp. Look at the time scheme, refresh offset, or event period involved, rather than at the data.

The pipe is broken. / Pipe is broken Inter-process communication between Infohub components failed because the process on the other end went away. The cause is the host process lifecycle, not the data source plugin the message appears under — a point worth making because the plugin is what gets blamed. Addressed in 7.2.0. On earlier versions, restarting the platform clears it until it recurs.

Could not load file or assembly '<name>' A dependency the plugin needs is not installed on that machine. For historian connectors this is normally the vendor's own client software — aahClientManaged.dll, for example, comes with the AVEVA/Wonderware client, not with Infohub. Install the vendor client on the host running the data source.

ERR_SSL_PROTOCOL_ERROR (in a browser, opening a dashboard) The browser could not negotiate TLS with Flow Server. The certificate, binding, or protocol version is wrong — see HTTPS and Certificates.

401 Unauthorized (from the Flow Server API) The token is missing, expired, or not valid for the instance being called. See Flow Server REST API.

When the message is not here#

Three things make an unfamiliar error answerable, and all three are usually missing from the first report:

1. The full text, including everything after "Something went wrong!" and the inner exception.
2. The operation — not "updating a template" but which template, onto what, and whether it was a full or partial update.
3. The version, exactly. Many of the messages above have had different causes in different releases, and the message alone does not distinguish them.

Reading Infohub Logs covers where each component writes, and which log holds the stack trace rather than just the summary line.

See also#

• Common Problems and Fixes
• Reading Infohub Logs
• Diagnosing Stuck or Faulted Measures
• Licence Keys in Practice