Well that is one a day now with transient http errors

Over the years never figured out to what WCG techs will / will not act / react. Presume this is below the interest threshold... not standing out in the [daily?] problem reports... no automaton monitors that alert techs by text or otherwise of bigger active issues.

Seen only one -119 [MD5 fail] and one -120 [RSA Key] since May 28. What I'm not sure about if this befalls task parts or app parts or both [think the latter], neither am I sure what comes out the CDN or directly from WCG servers. Setting the extra log flags [see Cc_config.xml wiki manual], could highlight the paths and paint a picture.

Is this random, a race condition, configuration [proxies/firewalls/Av are suspects]? Anyway, 'only' 32 posts in this thread [The OP title still not pointing readers to the specific problem, 'Errors' is just not a useful search term, so maybe you can edit the title of this post into your OP, so more wake up to 'ah, saw this in my logs' ;?)

I have several errors from FAH2 over the last couple of days. Just started crunching this project after completing 10 yrs on UGM. Example Event Log and error text from the Results Status below. If I read everything correctly above this is confirming that something is getting touched on the server end but not getting completely updated before download is attempted.

6/26/2016 9:30:13 PM | World Community Grid | Started download of fahb.FAH2_000083_avx38743_000019-in1.dms
6/26/2016 9:30:13 PM | World Community Grid | Started download of fahb.FAH2_000083_avx38743_000019-in2.dms
6/26/2016 9:30:14 PM | | Project communication failed: attempting access to reference site
6/26/2016 9:30:14 PM | World Community Grid | Finished download of fahb.FAH2_000083_avx38743_000019-in1.dms
6/26/2016 9:30:14 PM | World Community Grid | Temporarily failed download of fahb.FAH2_000083_avx38743_000019-in2.dms: transient HTTP error
6/26/2016 9:30:14 PM | World Community Grid | Started download of 67a33e5185cf85aea18580006239336a.rst
6/26/2016 9:30:14 PM | World Community Grid | Started download of 35fd59a27f466175f548a94679f2e33c.inp
6/26/2016 9:30:15 PM | | Internet access OK - project servers may be temporarily down.
6/26/2016 9:30:15 PM | World Community Grid | Finished download of 35fd59a27f466175f548a94679f2e33c.inp
6/26/2016 9:30:16 PM | World Community Grid | Started download of fahb.FAH2_000083_avx38743_000019-in2.dms
6/26/2016 9:30:16 PM | World Community Grid | Finished download of 67a33e5185cf85aea18580006239336a.rst
6/26/2016 9:30:17 PM | World Community Grid | Finished download of fahb.FAH2_000083_avx38743_000019-in2.dms
6/26/2016 9:30:17 PM | World Community Grid | [error] MD5 check failed for fahb.FAH2_000083_avx38743_000019-in2.dms
6/26/2016 9:30:17 PM | World Community Grid | [error] expected e9694d4072d74208c06ca15a3b5075f6, got dd2e3999773b1cfde20fda6aa10e17e5
6/26/2016 9:30:17 PM | World Community Grid | [error] Checksum or signature error for fahb.FAH2_000083_avx38743_000019-in2.dms

Result Name: FAH2_ 000083_ avx38743_ 000019_ 0048_ 025_ 0--
<core_client_version>7.6.22</core_client_version>
<![CDATA[
<message>
WU download error: couldn't get input files:
<file_xfer_error>
<file_name>fahb.FAH2_000083_avx38743_000019-in2.dms</file_name>
<error_code>-119 (md5 checksum failed for file)</error_code>
<error_message>MD5 check failed</error_message>
</file_xfer_error>

</message>
]]>

The question I have for Techs is, why are there still tasks in circulation from the earlier project months? E.g. the just reported task with error.

FAH2_ 000084_ avx38747_ 000061_ 0046_ 026_ 0--

Highest I've seen mentioned is batch 284 like over a month ago, the above is only generation 26 in the 30 sequences series of a batch 84. At 4 days deadline each task, the sequence would be done in at most 120 days [measured was ~78 days needed at start of project for a whole non-ls series]. Are there reruns being done for sequences that did not do 'minimum-per-run' i.e. 7-8 trickles? Credit is one thing [for computing effort], getting useful continuation is another i.e. if results do not meet the minimum needs, why not generate a task from the last good one and recompute immediately instead of kicking of partial as of for instance 0020000?

All supposition on my part based on available info, like to emphasize that I have zero problem with such reruns... just want to understand all down to the pound, penny, cent, escudo and deutschmark ;O)