[SOLVED] 3.21.157.11 assigned/connects/no download/hangs

[Neil-B]

The alarm clock approach is working … Caught another one … This time a single close connection was enough to get me an assignment to another server

[Neil-B]

… and another one … most times now it seems that the first try by the AS sends me to 3.21.157.11 … Probably I guess as it sees this server has a very low assign rate and therefore should have capacity … luckily I got a different WS after only one connection closing cycle.

[Neil-B]

… and another one … looks like I might have a broken nights sleep trying to keep the slots running !! … This time was a close the connection, reassigned to this server, close the connection then get lucky and be assigned somewhere else pattern.

Changed the Title to "Workaround" as tbh what I am doing doesn't feel like a long term viable solution

With this server assignment rate down to 12.00 an hour I don't reckon it is just one or two folders having this issue … Might be time for some server tlc - a cricket bat might do the trick

Even as I typed that the assignment rate leapt up … to 36.00 an hour - I must have scared it

[Joe_H]

I posted a message for the work server admins about this, not sure if any projects are active on this WA, none showed for me when I just checked the Project Summary page.

[Neil-B]

The beta series 14801, 14802, 14702 are showing https://apps.foldingathome.org/psummary?visibility=BETA ... and there are a few more projects in that series which I guess are coming from that server ... https://apps.foldingathome.org/psummary?visibility=ALL shows there may be more and the beta forum announcement also.

I have posted here as it is a Server Issue (imho) not a Beta Project one - but it is I believe having probably a significant impact on those projects if they are still trying to run them.

[Neil-B]

… and another - this time the ASs assigned me to this server seven (yup that many) times before finally got a different WS.

Well, with over 24hrs nursemaiding FAH I think I really now need to get some sleep … I have one slot due to finish in an hour the other in just over 3 hours … at current rates I reckon the chances of either slot actually actively folding something when I get up in 5hrs is close to zero which is a bit of a shame.

I think I have generated enough evidence today to show this is a repeatable and consistent issue.

[Neil-B]

... and I was correct ... both slot "ready" and waiting for me this morning (one had been patiently waiting on the Established connection for 2 1/2 hrs the other just 1/2 hr so could have been worse) ... closed both Established Connections (but inactive) to 3.21.157.11 and both slots assigned to other WSs first time, getting WUs and immediately recommencing folding ... server stats still show this server having negligible connections per hour - from what I am witnessing all it is likely to be doing is stalling folders clients?

Hopefully someone will get a chance today to either sort the server out (if there is an issue as the behaviour I am witnessing is suggesting) or identify if it is some issue with my kit that is causing this (which although the evidence might stack up against this is I concede still a possibility).

I'll edit this post in a bit to confirm how long the slots were stalled for - all I did was kick the connections and get the slots working - but for now grabbing 30mins more shuteye until next WU is due to complete (and the next Connection fail is likely to need sorting).

[PantherX]

FYI, I have informed the researcher so let's see what happens

[HaloJones]

what's a Linux equivalent to TCPView?

[Neil-B]

Not sure how similar, but https://github.com/chipmunk-sm/tcpview may be a port?

I guess using netstat and killing the right process might be the same thing - Every time I ask colleagues for a "similar" tool on Linux to a gui one I use in windows they seem to go straight to command line type a couple of strings and do the same thing - they don't appreciate I need an idiot level gui based interface to process information

The important bit for me with the windows TCPView is that I can see this server has an Established Connection as a row, I can right click on the row and select close connection - and that is simple enough that I can't (easily) mess it up !!

[Neil-B]

… and just in case anyone was wondering … still happening - luckily the ASs seem to have more options at the moment so getting other WS assignments as well as this one so not always having to close the connection

So far over last 48 hrs or so since this issue started I have had around 30 assignments to this server every single one of which have had to be manually "closed" … I'll increase this tally at notable moments - I guess the next is likely to be 50 "failures" - may get there within the next 24 hours at this rate

[Kebast]

I just had to remove the beta flag from my CPU. I couldn't keep up with all the failures.

[HaloJones]

This is why I think the AS remembers people because I haven't had more than one or two of these and I have 7 GPU slots running.

[Bastiaan_NL]

I just had to remove the beta flag from my CPU. I couldn't keep up with all the failures.

I did the same thing an hour ago, and it started folding in a minute.

[Neil-B]

Probably more linked with the type/size of slots and possibly OS … I run large(ish) CPU slots on Windows … The projects I am aware of on this server are Beta CPU projects … Kebast, Bastiaan_NL, HaloJones, what types of slots do you run?

Removal of the beta flag overnight for me might have to be done - two nights with minimal sleep periods and I am a tad shattered tbh

Sorry just read previous posts a bit better … HaloJones - your GPU slots are unlikely to be assigned to this server at the moment as I believe the projects on there are CPU ones so you may not have seen this much (only off CPU slots if you run them) … Kebast - I guess it is you CPU slot that gets these given you have removed the beta flag off that slot …