Affected system: PI Data Archive 2010 (3.4.385.77) / 6500 tags
On a production server a PI-to-PI interface caused several overflow event queues. When I've started up interface with wrong "history recover" values, it made this huge "history recover" and caused > 30 overflow event queues and currently there are > 500,000,000 total overflow events. Now PIARCHSS.EXE is busy with processing the data and only snapshot values are shown at client software. Snapshot / Archive statistic tells me: on average 5000 events per update cycle of statistics are processed. PI-to-PI interface is stopped, but other interface has to continue to feed the PI server.
PIARCHSS.EXE is now "backfilling" data to existing archive files. I assume it might be faster while PI server is offline. At TechSupport article 3081OSI8 - Reprocessing corrupted event queues - PI Server 2010 SP1 and before there is the following command listed:
piarchss.exe -if C:\PI\EVQ\pimq0001.dat -of e:\pi\piarch\piarch.102 [...]
My questions here: Does it make sense to process all overflow event queues offline? If yes, how I can find the right archive file for given event queue? Parameter -of is required.
Another idea is to setup a second PI server, let it process all event queues and merge those archive files to existing PI server. But it sounds too risky for me, because I hadn’t such case in the past. Does someone have experiences with this or a similar solution?
Thank you in advance for any feedback.