It's back again, all my CNODE work units are taking up to 2 hours or more to run, not 30 minutes.

OONI Probe does not seem to be affected, however an occasional PERF does run long as well but not all.

Been happening last few days.

Conan

The CNODE work units are still taking up to 3 hours to run for me, averaging well over the hour mark.

What happened to 30 minutes?

Conan

---

It's back again, all my CNODE work units are taking up to 2 hours or more to run, not 30 minutes.

OONI Probe does not seem to be affected, however an occasional PERF does run long as well but not all.

Been happening last few days.

Conan

The CNODE work units are still taking up to 3 hours to run for me, averaging well over the hour mark.

What happened to 30 minutes?

Conan

Well I found the basis for my long running work units.

After a reorganization of my internet connection (I moved the modem/router from my office to the front lounge room, re-ran the telephone lead-in cable to a new 8P8C (RJ45) wall connection and plugged a RJ11 phone jack into that for the modem, fibre only gets to the node then copper phone lines to the house, speeds aren't brilliant but OK), my internet started playing up.
I bought a 50M (metre) pre-made Cat6 SSTP cable, this shielded cable was to be my backbone cable from the modem through the ceiling back to my office where my computers are.
I ran the cable up through the wall across the ceiling down into my office, plugged it in and my internet started dropping out.
My gigabit HP ProCurve switch had been working fine with a Cat5e cable from the modem to the switch before the move. Constantly unplugging and replugging the cable in to get internet back was a hassle so I called the company and they sent a replacement cable.
This did the same thing, not as bad but just as annoying. So I spent more money and bought a crimping tool and connectors and then cut off and re-terminated the first cable, but nothing changed.
So I bought a new switch from TP-Link (the HP is at least 8 years old) only to find it did not work at all with the cable I had bought, no internet at all.
Put the old switch back in.
Went out today to a local electrical wholesaler and bought a normal 50 Metre Cat6 cable and all started to work.
A download that had taken 3 days to get to 60% done downloaded in seconds, my internet speeds jumped with speed tests going from 0.1Mbps to 35-50 Mbps.
With that CNODE work units dropped back to 30 minutes.

With CNODE needing an internet connection all the time the ultra slow internet (slower than the old dial up speeds for those of us that can remember that) was causing the work units to drag out trying to get it's data through.

So my $65.00 bargain cable, with the extras ($35.00 crimping tool+connectors, $150.00 switch, $75.00 new Cat6 cable), ended up costing $ 325.00, some bargain eh.

So not an issue with the project but with my home network, which I still have to re-wire through the ceiling yet to make it permanent (it is draped through the house during my testing).

At least now I wont have to check at 3 in the morning to see if I am still connect to the net.

Conan

---

10+ years ago, for about $50, I bought a cat5 cable kit, probably from Newegg.

Best networking investment I ever made.

500 feet of cat 5 cable, 100 connectors, and a crimping tool. I forget if it came with the cable tester, or if I bought that separately.

After all these years, I still have plenty of cable and connectors left.

Besides making normal short run cables, it's invaluable for running cables through walls or ceilings because I can drill a hole the size of the cable and not have to worry about making holes larger enough for the connectors. Run the cable first, then put the connectors on as the last step.

Best part is I always know the cables *work*, because I made them myself.

And when you need that super long 150" cable, it's no harder to make than a 5" cable.

I only run 1GB ethernet, so CAT 5 is sufficient. More than that is wasting money.

---

Want to find one of the largest known primes? Try PrimeGrid. Or help cure disease at WCG.

Much longer Runtime, same credit!?
Name PERF_TESTS_0D_T1665234001_705938_2_30
Anwendung iThena PERF
erstellt 8 Oct 2022, 13:00:53 UTC
autorisiertes Ergebnis 311097093
gewährte Punkte 2.00000000
Mindestanzahl 1
Anfängliche Kopien 1
max # von Fehler/Gesamt/Erfolg Aufgaben 1, 1, 1

Name PERF_TESTS_0D_T1670582702_568909_2_30
Anwendung iThena PERF
erstellt 9 Dec 2022, 10:45:12 UTC
Mindestanzahl 1
Anfängliche Kopien 1
max # von Fehler/Gesamt/Erfolg Aufgaben 1, 1, 1
Fehler Zu viele Ergebnisse insgesamt
I abort it after 37 min runtime. It stucks at 99.999%.

The tasks run much longer! I stop all.

PERF run times have gone back up to over 2,400 seconds (40 minutes) instead of 179 seconds (3 minutes) on Windows.

Conan

---

I had 3 sets appear to be stuck at 100% after more than an hour on Linux.
No new tasks set and all aborted.

Paul.

Watched a PERF in BOINCTUI Raw View.
CPU 0.000000 after 1100 seconds, completed PERF show .02 to .04 on results page.
Fraction done rising in ever smaller increments, now 0.995339.
Little activity on LAN light, probably just BOINCTUI connection.
Will let it run but appears broken to me.

Paul.

FIRST, FIX WHATEVER IS WRONG WITH THE APPLICATIONS THAT ARE RUNNING FOR HOURS ON END AND CLOGGING MY INTERNET CONNECTION.

THEN, FIX THE UNRESPONSIVENESS OF YOUR WEBSITE THAT'S PROBABLY RELATED TO THE APPS PROBLEM.

---

PERFs now running OK in usual 4 minutes.

Paul.

PERF runs from about 10:40 UTC on 8-4-23 ran about 11,000 seconds (3 hours) and next set appear same.
From the unusually high amount ready to send probaly same for all.

Paul.

PERF runs from about 11:00 UTC 8-5-23 appear bad, no CPU used and run over usual 4 minutes.

Paul.

Back to normal.

Paul.

My Linux PERF and CNODE work units are not ending.
Since maintenance the work units are running for ever.
I have a PERF that has reached 100% after 1 hour but still running (usual time is 3 to 4 minutes)
I have CNODE work units that have passed 2 hours and not past 79% and still running (usual run time is about 30 minutes)

This is happening on all 3 of my Linux computers.

I will abort them and see if the next ones run OK or not.

Conan

EDIT:: new ones are doing the same, running forever.

OONI work units don't seem to be affected.

---

Confirmed and still bad as of 14:00 UTC.

Cnode & perf bad after about 10:00 UTC Tuesday 8th Aug.

OONI appear OK.

---

Paul.

PERF & Cnode appear normal since about 15:30 UTC (only had a few so far).

---

Paul.