I'll set it to the default of 2.

If that does resolve the errors, it would still indicate there's an underlying problem- for whatever reason you're unable to download multiple files at the same time.

---

Grant
Darwin NT

I'll set it to the default of 2.

If that does resolve the errors, it would still indicate there's an underlying problem- for whatever reason you're unable to download multiple files at the same time.

But that is not really the case ..... I can and do download multiple files at the same time. I do it regularly. Only recently is this happening with BOINC projects. After the outrage today, when I turn the computers back on to contact the project after it gets back up, I regularly get blocks of 40-45 files requested for the 1st, 2nd and 3rd contacts to get back to my 300 task quota per machine. Einstein and MilkyWay are asking for work at the same time that Seti asks for work. I have never had a download error during those massive requests for work for all projects. It's always happens during the middle of a normal processing day, usually just a single task request that generates the error. So, one error per 1000 download requests?? This has got me stumped. What has changed in the environment? Based on timelines ... it all started when I began using Kaspersky Internet Security. My suspicion is that the AV program is not obeying my exclusions or is slowing the systems down right at the exact time a downloaded task needs to be written to disk.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

I'll set it to the default of 2.

If that does resolve the errors, it would still indicate there's an underlying problem- for whatever reason you're unable to download multiple files at the same time.

But that is not really the case ..... I can and do download multiple files at the same time. I do it regularly. Only recently is this happening with BOINC projects. After the outrage today, when I turn the computers back on to contact the project after it gets back up, I regularly get blocks of 40-45 files requested for the 1st, 2nd and 3rd contacts to get back to my 300 task quota per machine. Einstein and MilkyWay are asking for work at the same time that Seti asks for work. I have never had a download error during those massive requests for work for all projects. It's always happens during the middle of a normal processing day, usually just a single task request that generates the error.

In which case changing the number of simultaneous downloads shouldn't make any difference.

Based on timelines ... it all started when I began using Kaspersky Internet Security. My suspicion is that the AV program is not obeying my exclusions or is slowing the systems down right at the exact time a downloaded task needs to be written to disk.

So it would be a good idea to uninstall it & use a different programme to see if it is a the heart of the issues.

When fault finding it's best to go for
1 the easiest things to check, no matter how obvious, or unlikely.
2 the most likely cause, even if it is a pain in the arse.

As you eliminate possible causes, then what's left- no matter how strange or unlikely it may seem- becomes the most likely.
Unfortunately intermittent faults by their very nature can make it extremely difficult to eliminate possible causes without any room for doubt that you have actually eliminated that particular possibility.

---

Grant
Darwin NT

it all started when I began using Kaspersky Internet Security. My suspicion is that the AV program is not obeying my exclusions or is slowing the systems down right at the exact time a downloaded task needs to be written to disk.

What exactly did you exclude? No point in excluding folders on disk - it's fairly clear nothing is making it as far as the disk. Did I see one error log where only 40 bytes - out of 367 KB - made it as far as the BOINC client which would do the writing to disk of anything which a disk scanner would subsequently investigate?

You need to see if Kaspersky has a separate set of controls to manage active internet traffic - maybe a whitelist of trusted source domains?

What has changed in the environment? Based on timelines ... it all started when I began using Kaspersky Internet Security. My suspicion is that the AV program is not obeying my exclusions or is slowing the systems down right at the exact time a downloaded task needs to be written to disk.

You might want to see if you can correlate the timing of the download failures with the times that Kaspersky is downloading its definition file updates on any of your machines. I know that McAfee, which I have here on my daily driver, seems to hog the whole system when it's doing that update. And that probably affects the throughput for my whole network for the duration of the download.

Also try Googling Testing Internet Reliability

There was some very good links last time I looked. Your not interested in speed, but reliability.

What exactly did you exclude? No point in excluding folders on disk - it's fairly clear nothing is making it as far as the disk. Did I see one error log where only 40 bytes - out of 367 KB - made it as far as the BOINC client which would do the writing to disk of anything which a disk scanner would subsequently investigate?

You need to see if Kaspersky has a separate set of controls to manage active internet traffic - maybe a whitelist of trusted source domains?

I took quite a bit of Googling to come up with an answer to your question. Very easy to "blacklist" a URL or IP address in KSS. No information at all on how to "whitelist" though. I found an solution I'd like to try. Does anyone know the IP addresses of the SETI download servers, the ones that actually feed me work when I ask for it?

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

What has changed in the environment? Based on timelines ... it all started when I began using Kaspersky Internet Security. My suspicion is that the AV program is not obeying my exclusions or is slowing the systems down right at the exact time a downloaded task needs to be written to disk.

You might want to see if you can correlate the timing of the download failures with the times that Kaspersky is downloading its definition file updates on any of your machines. I know that McAfee, which I have here on my daily driver, seems to hog the whole system when it's doing that update. And that probably affects the throughput for my whole network for the duration of the download.

Seems that Kaspersky is downloading its definition file updates at the same time every day. I haven't correlated the download error times with Kasperksy definition file download times yet. I set the [Skip data file verification] setting on this computer the other day and no errors yet. I didn't on the other main cruncher and I got another download error on it yesterday. Just set the toggle on that machine. Will continue to monitor.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

Does anyone know the IP addresses of the SETI download servers, the ones that actually feed me work when I ask for it?

Current download server is 208.68.240.127 - but it can vary. It's probably safe to pass through any traffic from 208.68.240.* - SETI@Home owns 208.68.240.1 and 208.68.240.254, so I think we can assume the whole subnet. Other addresses in the range are used by other SETI servers.

Current download server is 208.68.240.127 - but it can vary. It's probably safe to pass through any traffic from 208.68.240.* - SETI@Home owns 208.68.240.1 and 208.68.240.254, so I think we can assume the whole subnet. Other addresses in the range are used by other SETI servers.

!UhOh! My solution probably won't work then if I have to use the global mask 208.68.240.* The solution I found in a computer tutorial said to whitelist an IP by bracketing the IP address with *
A double ** at the end will likely null the solution.

I will try *208.68.240.1* and see what happens.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

If you can only enter a single IP address at a time, the address you asked for was 208.68.240.127 - use that, with whatever delimiters are required specifically by Kaspersky.

I suggested replacing the final octet with an asterisk as a generic wildcard - again, use whatever wildcard is specific to Kaspersky.

Another, more general, 'whole subnet' designator might be 208.68.240.0 - I don't think that is a valid single address - but 208.68.240.1 certainly is, so Kaspersky might interpret that as whitelisting one server only - the wrong one.

The IP address you stated was the download server resolves via reverse DNS to "Vader" That is the download server marked as disabled on the SSP. Is that just a mixup in the SSP listings? I have put 208.68.240.127 which resolves to http://vader.ssl.berkeley.edu in the "whitelist". You can only add URL's to the "whitelist", not IP addresses.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

The IP address you stated was the download server resolves via reverse DNS to "Vader" That is the download server marked as disabled on the SSP. Is that just a mixup in the SSP listings? I have put 208.68.240.127 which resolves to http://vader.ssl.berkeley.edu in the "whitelist". You can only add URL's to the "whitelist", not IP addresses.

I got the current IP address by pasting the download server url from client_state - 'boinc2.ssl.berkeley.edu/ into a ping cmd window to resolve it.

Doing the same today resolves to 208.68.240.119 - so maybe they've restored round-robin DNS. That resolves as

Pinging georgem.ssl.berkeley.edu [208.68.240.119]

The full answer is

C:\Windows\system32>ipconfig /displaydns

Windows IP Configuration

    boinc2.ssl.berkeley.edu
    ----------------------------------------
    Record Name . . . . . : boinc2.ssl.berkeley.edu
    Record Type . . . . . : 1
    Time To Live  . . . . : 212
    Data Length . . . . . : 4
    Section . . . . . . . : Answer
    A (Host) Record . . . : 208.68.240.119


    Record Name . . . . . : boinc2.ssl.berkeley.edu
    Record Type . . . . . : 1
    Time To Live  . . . . : 212
    Data Length . . . . . : 4
    Section . . . . . . . : Answer
    A (Host) Record . . . : 208.68.240.127

Round-robin it is.

I've changed to http://georgem.ssl.berkeley.edu and also added http://boinc2.ssl.berkeley.edu. That gives the program what it wants. I guess the round-robin will sort it out.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

I see that I picked up another download error after I changed my global preferences for all machines to skip data verification. So that didn't work apparently. Was running with the wrong download server whitelisted at the time so that is also in consideration. Will continue to monitor with no verification and georgem whitelisted.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

I see that I picked up another download error after I changed my global preferences for all machines to skip data verification. So that didn't work apparently. Was running with the wrong download server whitelisted at the time so that is also in consideration. Will continue to monitor with no verification and georgem whitelisted.

'Skip data verification for image files' applies to visual image files only - JPEGs and other eye-candy for the screensaver or Manager's 'Simple View'. You would expect scientific data files to be checked always and unconditionally.

Are you still getting [small number of bytes] received, expected [larger number, typically 367 KB] in your event log?

If so, it might be interesting to locate the file on your disk (be quick, it'll be deleted after the download failure is reported back to the server, so five minutes later), and peek inside it with an editor to see what those bytes actually say. I once came a range of Internet routers (Two-Wire, distributed in the UK as 'BT Business Hubs') which sent back messages which were designed to be viewed in a browser when things went wrong. Of course, if the http request wasn't initiated by a browser, that wasn't very helpful...

Yes, that is probably what I will have to do, catch it in the act and move it before it gets deleted so I can look into it. Since I don't sit on top of the computers all day long, I will have to have some luck to see it happen in real-time and then investigate. Yes, still getting the same truncated message for the failed files. Did see for the first time an entry for checksum or signature error. That wasn't in the logs in the previous errors.

30-Jun-2016 02:30:55 [SETI@home] Scheduler request completed: got 6 new tasks
30-Jun-2016 02:30:57 [SETI@home] Started download of blc3_2bit_guppi_57451_66326_HIP117463_OFF_0014.30807.0.17.26.253.vlar
30-Jun-2016 02:30:57 [SETI@home] Started download of blc3_2bit_guppi_57451_66326_HIP117463_OFF_0014.30807.0.17.26.250.vlar
30-Jun-2016 02:30:59 [SETI@home] Incomplete read of 43.000000 < 5KB for blc3_2bit_guppi_57451_66326_HIP117463_OFF_0014.30807.0.17.26.253.vlar - truncating
30-Jun-2016 02:30:59 [SETI@home] Finished download of blc3_2bit_guppi_57451_66326_HIP117463_OFF_0014.30807.0.17.26.253.vlar
30-Jun-2016 02:30:59 [SETI@home] Finished download of blc3_2bit_guppi_57451_66326_HIP117463_OFF_0014.30807.0.17.26.250.vlar
30-Jun-2016 02:30:59 [SETI@home] Started download of blc3_2bit_guppi_57451_66326_HIP117463_OFF_0014.30807.0.17.26.247.vlar
30-Jun-2016 02:30:59 [SETI@home] Started download of 13my10af.11525.501.4.31.132
30-Jun-2016 02:30:59 [SETI@home] File blc3_2bit_guppi_57451_66326_HIP117463_OFF_0014.30807.0.17.26.253.vlar has wrong size: expected 365373, got 0
30-Jun-2016 02:30:59 [SETI@home] Checksum or signature error for blc3_2bit_guppi_57451_66326_HIP117463_OFF_0014.30807.0.17.26.253.vlar
30-Jun-2016 02:31:01 [SETI@home] Finished download of blc3_2bit_guppi_57451_66326_HIP117463_OFF_0014.30807.0.17.26.247.vlar
30-Jun-2016 02:31:01 [SETI@home] Finished download of 13my10af.11525.501.4.31.132
30-Jun-2016 02:31:01 [SETI@home] Started download of blc3_2bit_guppi_57451_66326_HIP117463_OFF_0014.30807.0.17.26.9.vlar
30-Jun-2016 02:31:01 [SETI@home] Started download of 13my10af.11525.501.4.31.126
30-Jun-2016 02:31:02 [SETI@home] Finished download of blc3_2bit_guppi_57451_66326_HIP117463_OFF_0014.30807.0.17.26.9.vlar
30-Jun-2016 02:31:02 [SETI@home] Finished download of 13my10af.11525.501.4.31.126

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

Did see for the first time an entry for checksum or signature error. That wasn't in the logs in the previous errors.

30-Jun-2016 02:30:59 [SETI@home] Checksum or signature error for blc3_2bit_guppi_57451_66326_HIP117463_OFF_0014.30807.0.17.26.253.vlar

See Your post (27 Jun 2016):
http://setiathome.berkeley.edu/forum_thread.php?id=79812&postid=1799102#1799102

---

 


- ALF - "Find out what you don't do well ..... then don't do it!" :)
 

Did see for the first time an entry for checksum or signature error. That wasn't in the logs in the previous errors.

30-Jun-2016 02:30:59 [SETI@home] Checksum or signature error for blc3_2bit_guppi_57451_66326_HIP117463_OFF_0014.30807.0.17.26.253.vlar

See Your post (27 Jun 2016):
http://setiathome.berkeley.edu/forum_thread.php?id=79812&postid=1799102#1799102

Guess I missed that or didn't remember.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

I believe my download errors have been resolved and the matter closed. I haven't seen any more errors since white-listing georgem and vader, the download servers in Kaspersky Internet Security 2016.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)