Rong,
From the Archiver Appliance’s website, it has statistic data about the PV connections, drops, and etc.
Regarding the timestamp, beyond your timestamp, I guess that all PVs inside your IOC got processed?
From our deployment at APS accelerator side, we did occasionally see some data point drops.
But over, it’s pretty healthy.
Thanks,
Guobao
From:
Tech-talk <tech-talk-bounces at aps.anl.gov> on behalf of Rong Huang via Tech-talk <tech-talk at aps.anl.gov>
Date: Monday, September 23, 2024 at 6:23 PM
To: Michael Davidsaver <mdavidsaver at gmail.com>
Cc: tech-talk at aps.anl.gov <tech-talk at aps.anl.gov>
Subject: Re: archive appliance some PVs no data
Hi Michael,
Thank you very much for your reply.
I searched for keyword "dropped" in all tomcat instance directories where each log directory locates, I find nothing:
archiver@strawberry:~/epics/archive$
grep -r "dropped" engine/logs/
archiver@strawberry:~/epics/archive$
grep -r "dropped" etl/logs/
archiver@strawberry:~/epics/archive$
grep -r "dropped" mgmt/logs/
archiver@strawberry:~/epics/archive$
grep -r "dropped" retrieval/logs/
You are right that on the detail page, the problemed PV's "how many events so far" is 0.
Paul Sichta from PPPL emailed me and mentioned the possibility of a wrong PV timestamp. I find out our troubled PVs have the timestamp of year 1990, and I will discuss with our IT staff to find out what changed
to our IOC booting host PC.
Our IOC cmd file uses `putenv ("EPICS_TS_NTP_INET=10.1.0.3")` to sync time with the PC.
On 9/23/24 12:20, Rong Huang via Tech-talk wrote:
> Hello,
>
> I installed Archiver Appliance last year while APS was down. To learn how to use it, I started archiving some PVs, and paused them. I "resumed" the archiving now while APS is back, but some PVs can be archived, some PVs cannot.
Some general AA troubleshooting advice...
I would suggest looking at the "error" reports pages,
paying special attention to those with names beginning
with "PVs by dropped ...".
Also, look on the details page for one of the PVs in question.
eg. look for the "How many events so far?" line. Refresh a
few times to see that it increments when that PV should update.
Also look for "How many events lost totally so far?" and see
that it is zero, or at least not incrementing.
> I use
http://strawberry.ls-cat.org:17665/mgmt/ui/index.html <https://urldefense.us/v3/__http://strawberry.ls-cat.org:17665/mgmt/ui/index.html__;!!G_uCfscf7eWS!dGt06d4Y35r21tzRJQ1P4MfKQL9FjIyz9ZfpC9EqT_x-zgfhGmPeWFfshx886REUI6bsf-LGb559zQ0dOJG_bOkZ$>
to Achive/Pause/Resume archiving.
>
> I use
http://strawberry.ls-cat.org:17668/retrieval/ui/viewer/archViewer.html <https://urldefense.us/v3/__http://strawberry.ls-cat.org:17668/retrieval/ui/viewer/archViewer.html__;!!G_uCfscf7eWS!dGt06d4Y35r21tzRJQ1P4MfKQL9FjIyz9ZfpC9EqT_x-zgfhGmPeWFfshx886REUI6bsf-LGb559zQ0dODFD8Xz_$>
to view/retrive data.
>
> The problem, using "21:C1:MO:E:ActPos" as an example, the archViewer website showing one value straight line, "113" Kev, for any display time ranges. "113" comes from the mono E reading when I started archiver appliance last year but not the current E reading.
>
> I deleted the PV "21:C1:MO:E:ActPos" and its data from archiver appliance, re-entered the PV and started its archiving, the mgmt page showed it "Being archived", "appliance0", "Connected? true", "Sampling period: 10.0".
>
> archViewer web page shows a flat line of "0" for any time range.
>
> There is no data file of this PV saved in /arch/sts/ArchiverStore/21. (Even there is no subdirectory of "C1" in /arch/sts/ArchiverStore/21 after I manually removed it and deleted all 21:C1:* PVs)
>
> Restarting archiving appliance (sampleStart.sh stop & sampleStart.sh start) does not help.
>
> But some PVs can be archived after "resume" (some of them might have noise, therefore the values were varying all the time in the past).
>
> I am using apache-tomcat-9.0.86. mysql 8.0.37-0ubuntu0.22.04.3, I installed Archinver Appliance using single_mechine_install.sh.
>
> Thanks a lot!
> Rong
>
|
