I m preparing a solution of health monitoring. how we can monitor the archive corruption?
Jinmo is correct that since PI Data Archive 2016, the PI Data Archive will check for the errors indicating archive corruption on writes to the archives, and flag them as corrupt if these errors are encountered. In addition to looking in PI SMT, there is a performance counter that tracks the number of archives marked as corrupt. Perhaps you could use this to help monitor archive corruption.
\\<Hostname>\PI Archive Subsystem\Corrupted Archives Count
The above screenshot is taken from PI SMT -> IT Points -> Performance Counters and can be gathered using the PI Interface for Performance Monitor.
I don't believe we have an official way of monitoring for archive corruption because the corruption occurs at the record level and not at the file level. A possibility would be to have a PowerShell running periodically that uses pigetmsg.exe to output the SDK logs of the PI Data Archive to a .txt file, searches the .txt file for the errors linked to archive corruption and then deletes the .txt file.
The errors that indicate archive corruption can be found in this KB:
Edit (After Adam Fink answered)
I don't believe we have an official way of monitoring for archive corruption before PI Data Archive 2016. The corruption occurs at the record level and not at the file level. A possibility would be to have a PowerShell running periodically that uses pigetmsg.exe to output the SDK logs of the PI Data Archive to a .txt file, searches the .txt file for the errors linked to archive corruption and then deletes the .txt file.
With the PI Data Archive version 2016 or newer, you can use the performance counter Adam suggested in the below post.
Since PI Data Archive 2016, we do offer an visual indicator within PI-SMT Archives Plug-in so that you can view which archive is corrupt and can reprocess them at once.
Have you take a look at it before?
Hi Jinmo ,
ya i have seen the indication. but i m going through some monitoring process via points.. Is there any health points available for indicate Archive Corruption??
Thanks Adam Fink
In addition to what others have mentioned, you may also want to monitor PI Archive Subsystem\Failed Events/sec.
Although failed events do not necessarily indicate a corrupted archive by counting failed archive event additions, we can get a sense of when the archive is rejecting data due to corruption. There are many "expected" reasons as of why there could be a failed event but one proactive approach is to look at the message logs when and if the number spikes.
Retrieving data ...