And that is the problem.

---

Pluto will always be a planet to me.

Seti Ambassador
Not to late to order an Anni Shirt

Unless they set their "Community Preferences" to receive notification of PMs only on their Account Page...

Which is the default setting, and requires no action at all.

That's what I thought, but couldn't remember......

So we are back to the earlier suggestion - a volunteer sending a direct email to the problem cruncher.

---

Donald
Infernal Optimist / Submariner, retired

Unless they set their "Community Preferences" to receive notification of PMs only on their Account Page...

Which is the default setting, and requires no action at all.

That's what I thought, but couldn't remember......

So we are back to the earlier suggestion - a volunteer sending a direct email to the problem cruncher.

Or fixing quota so they don't get WUs to trash.

How about in the next version of Bionic or seti the defaults are changed to receive notices and a email as the default in stead of what it now only getting them on there accounts page .

Unless they change it , so if it's a set and forget they will be notified one way or the other

---

I was not talking about those who, for whatever reason, do not get a PM. But those who on getting a PM that one of their crunchers is sick reply with a message along the lines "So what, it's still working"....

There are of course those who have changed their email account so the notice that a PM has been sent will not be received, and those who don't actually know they are running SETI for whatever reason....

---

Bob Smith
Member of Seti PIPPS (Pluto is a Planet Protest Society)
Somewhere in the (un)known Universe?

I was not talking about those who, for whatever reason, do not get a PM. But those who on getting a PM that one of their crunchers is sick reply with a message along the lines "So what, it's still working"....

have you had that happen Rob ?

Gees some people then , maybe they just need to be banned outright if that's there attitude or am i being to harsh .

---

I don't know why Seti doesn't just deal with the problem. When the Admin is notified of a "rogue" user that is trashing workunits, then simply block their account. If they are genuine users they will soon realise that they aren't getting any more work and ask why. If they are "set and forget" people, or someone that has inherited one of those rigs without knowing, they won't care anyway.

well they are short on labor so possibily finding the time to do it mite be a problem .

Not shore what you mean about wringing of hands !

---

I don't know why Seti doesn't just deal with the problem. When the Admin is notified of a "rogue" user that is trashing workunits, then simply block their account. If they are genuine users they will soon realise that they aren't getting any more work and ask why. If they are "set and forget" people, or someone that has inherited one of those rigs without knowing, they won't care anyway.

Why all this wringing of hands?

There are currently probably several hundred possibly more accounts that would need to be blocked, then unblocked when they say they have got it working, then blocked again when it transpires they haven't

The best way is to repair the problem with the servers and just stop them getting new work, one fix instead of hundreds. Either the person will give up or they just might come here and ask what the problem is, someone can then explain and hope they fix it, of course if they don't then they still won't get any work.

Seems to me the simplest way.

---

I don't know why Seti doesn't just deal with the problem. When the Admin is notified of a "rogue" user that is trashing workunits, then simply block their account. If they are genuine users they will soon realise that they aren't getting any more work and ask why. If they are "set and forget" people, or someone that has inherited one of those rigs without knowing, they won't care anyway.

Why all this wringing of hands?

There are currently probably several hundred possibly more accounts that would need to be blocked, then unblocked when they say they have got it working, then blocked again when it transpires they haven't

The best way is to repair the problem with the servers and just stop them getting new work, one fix instead of hundreds. Either the person will give up or they just might come here and ask what the problem is, someone can then explain and hope they fix it, of course if they don't then they still won't get any work.

Seems to me the simplest way.

+1 and also the cheapest way of sorting the problem out .

The best way is to repair the problem with the servers and just stop them getting new work, one fix instead of hundreds.

In theory yes I agree. But who is going to repair the problem with the servers? Do they have any impetus to do so. And even then, what about people fiddling with the code at 11pm at night :-) Is it a case of just a dozen accounts causing 90% of the problem? If so a fairly simple quick task to block them.

And don't take their word for it they have fixed it, demand proof. There are plenty of technically savvy users here that could be used to deal with this, apart from admins, Devs, and Mods. Seti can't be held to ransom like this, time the project put it's foot down.

So you are quite happy with various people fiddling blocking various accounts but don't feel it is a good idea to stop the problem at source.

Doing it your way means someone has to keep an eye out all the time for the problem machines.

Rater than a "fix and forget" solution.

---

At projects where there is a higher (much higher) ratio of staff to users - like at CPDN and Einstein - the message board moderators take the technical role of identifying problem hosts and reporting them to the administrators.

That helps both sides - there is a consistent and trusted communication channel (the moderators are also the most experienced users), and the staff can just act on the reports without needing to spend time cross-checking. CPDN, in particular, have a regular policy of 'minusing' bad hosts and sending a direct pro-forma email from staff to any user minused. Some respond, fix the problem, and go back to crunching: some don't.

im kinda new to this, i checked all the users to see if i was on the list haha luckily i am not

At projects where there is a higher (much higher) ratio of staff to users - like at CPDN and Einstein - the message board moderators take the technical role of identifying problem hosts and reporting them to the administrators.

That helps both sides - there is a consistent and trusted communication channel (the moderators are also the most experienced users), and the staff can just act on the reports without needing to spend time cross-checking. CPDN, in particular, have a regular policy of 'minusing' bad hosts and sending a direct pro-forma email from staff to any user minused. Some respond, fix the problem, and go back to crunching: some don't.

Not quite right approach IMHO.
We need to keep in mind that system (stock one) should be designed as fully unattended one. This blacklisting nevertheless breaks this conception cause requires action from user (besides his initial act of giving computational resources at project's disposal).
What would be better IMHO: define blacklisting as long enough quota of 1 tasks per day. W/o all those inflation modifiers that cause so much grief for GPU hosts. And lift blacklisting automatically when N days (project/tunable) w/o issues passed.
Then system will keep ability to repair itself w/o direct user involvement.
What I mean: user can replace broken GPU (cause his favorite shooter freezes too often), he can reinstall OS/drivers (again, cause his favorite shooter lags and wife's solitaire crashes OS ;D )... all this will return host to usable for SETI/BOINC. But all this don't imply that user will contact project boards/admins/mods and ask for unblock.

I know that at CPDN computers giving invalid results are blacklisted. Their owners can send a mail explaining what they have done to mend and ask for being readmitted. Note that there are no GPU tasks at CPDN. Some tasks are sent only to Windows users, some to Windows and Mac and some only to Linux. I am running both Windows 8.1 and SuSE Linux 13.1 at CPDN. No invalid results so far.
Tullio

---

Agreed, it's not an ideal or a technically elegant approach, but it is an approach, and it works - having evolved over the years - in the CPDN context, which is very different to the SETI context. In particular, CPDN tasks are designed and expected to take days, or even weeks to complete, and involve downloading many megabytes of data: a task is a big investment for the project, and a potential cost (in download fees) to the user. 'Serial task killers', as they're known internally, are - in relative terms - a bigger problem for CPDN than they are here, and this is an appropriate level of response.

I just finished a CPDN task in 32 days on this not very fast Linux box, vintage 2008, alongside a vLHC task. CPU is an Opteron 1210 at 1.8 GHz, RAM 8 GB.

---

To bad "rouge users" can't be publicly shamed via twitter, Facebook, etc...

No way to identify anonymous accounts.


Keifer

---

Heck, why stop at public humiliation. I say we cane them. 20 or 30 rounds of that would put them back in line. Seti could even charge admission to the event and that would help with the funding issues. a WIN WIN.

Until then though, we probably should do a less severe approach like limiting the amount of WU they would receive.


Bob

(edited - fixed a missing word)

---

Sometimes I wonder, what happened to all the people I gave directions to?
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
SETI@home classic workunits 4,321
SETI@home classic CPU time 22,169 hours

To bad "rouge users" can't be publicly shamed via twitter, Facebook, etc...

No way to identify anonymous accounts.


Keifer

With such attitude maybe it was very clever idea to hide own hosts :P
Even if some really malice connections were made to project it only means that BOINC system not enough secured from such attacks - ToDo for devs.
Quota system definitely requires editing. It useless for GPU issues. And those issues are from innocent participants, not malice ones. Users, that just can't or won't spend additional effort for this project besides initial act of donation of computational resources.
I'm quite sure that % of really malice hosts (that is, those that trash tasks on purpose being fully aware about that trashing and actively refused to fix it) is neglectible small. All those "user punishment" ideas are absolutely wrong way to go.
Own BOINC quota management - that's what really matter.

I apologize if my thoughts were misunderstood. I was thinking of perhaps another way to get the attention of those few who tend to "set and forget" as well as those who feel it is "beneath them" to respond to sincere inquiries regarding the possible failure of their systems.

I meant no disrespect to those who, despite the responsible tending of their systems, compile invalid or inconclusive wu's due to circumstances beyond their control.

Personally I will always welcome any advice on how to make my system run more efficient. Feel free to examine my work and ask questions about my rig as it is not hidden.


Keifer

---