• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

New SETI for BOINC-servers probably online in 2 weeks time...

Rattledagger

Elite Member
Allow me to have some fun stoking the fire with some pre-announcement info:

1. I started up all three splitters just to make sure they were working nicely in tandem. One splitter was plenty to handle our current demand. We cannot "build up" workunits as we currently only store about 500,000 on disk at any given time (when we get about this amount, the splitters go to sleep). But now I have at least proven that when demand increases, we'll be able to keep up.

2. New equipment has been acquired (and actually arrived earlier than expected). Much setup/testing needs to happen before we make any wild claims. The good news is that this stuff is in our lab *now* (when we expected it to arrive in early January). I don't celebrate Christmas, nor travel during this time of year, so I (and other team members) will be working on and off over the next two weeks trying to bring all the new servers on line.

Feel free to use this information to adjust your predictions about when we are going to shut down classic SETI@home.

- Matt (Lebofsky)
 
Ye gads, it looks like we can really start to take this seriously!! I'm not gonna reach my goal 🙁

Now we can begin guessing if BBR has a sandbag that would allow them to challenge us


Edited for spelling...
 
I refuse to accept the end of SETI Classic until I can no longer get WUs from anywhere and the last WU crunched leaves my Q......

Does sound like the end is near. But I'm not holding my breath. It could be 2 weeks, it could be 2 months. 🙂
 
To answer some questions:

The cross platform GUI work is happening simultaneously with any new hardware installs. Not sure of the status of that (not my department), but it'll probably be working by the time any new systems are up.

The Dell server is old news - the new equipment I'm referring to was Sun V40z, etc. which is listed in some other post below. It'll be a vast improvement over our current database server (a Sun 220) - it will have faster disks (hardware RAID), more disk space, much faster CPUs (and the ability to expand - we're starting with 2), and much more memory (8GB with the ability to grow vs. 2GB max).

As well, what I'm working on specifically right now is some donated equipment - a whole batch of boards given to us by a private donor to be named later in some technical news item. I was able to use these parts to assemble a whole new E3500 system that is slightly more powerful that our current E3500 system. We will use this new E3500 to help migrate off the slow, bulky, non-RAIDed disks on the old E3500 and onto much faster storage. This will be our new back-end science database server, and with more CPU/RAM power and faster disks we will be able to clear out the backlog of SETI@home classic data to be reduced within a few months (and then free up a 4-cpu E450 for other use). As well, the validator and assimilator will be able to insert data much faster as well. I'm busy waiting for a solaris 9 install on this machine, hence the influx of free time to jot down this little missive to y'all.

- Matt (Lebofsky)

The new BOINC database-server is:

Sun V40z + Sun StorEdge 3510:
Dual Opteron 844, 1.8 GHz (max 4 cpu)
8 GB memory (max 8 GB/cpu)
RAID-array, max 876 GB unformatted if 15k-disks, 1.72 TB if 10k-disks. (not specified what Seti is using)
 
Originally posted by: Rattledagger
To answer some questions:

The cross platform GUI work is happening simultaneously with any new hardware installs. Not sure of the status of that (not my department), but it'll probably be working by the time any new systems are up.

The Dell server is old news - the new equipment I'm referring to was Sun V40z, etc. which is listed in some other post below. It'll be a vast improvement over our current database server (a Sun 220) - it will have faster disks (hardware RAID), more disk space, much faster CPUs (and the ability to expand - we're starting with 2), and much more memory (8GB with the ability to grow vs. 2GB max).

As well, what I'm working on specifically right now is some donated equipment - a whole batch of boards given to us by a private donor to be named later in some technical news item. I was able to use these parts to assemble a whole new E3500 system that is slightly more powerful that our current E3500 system. We will use this new E3500 to help migrate off the slow, bulky, non-RAIDed disks on the old E3500 and onto much faster storage. This will be our new back-end science database server, and with more CPU/RAM power and faster disks we will be able to clear out the backlog of SETI@home classic data to be reduced within a few months (and then free up a 4-cpu E450 for other use). As well, the validator and assimilator will be able to insert data much faster as well. I'm busy waiting for a solaris 9 install on this machine, hence the influx of free time to jot down this little missive to y'all.

- Matt (Lebofsky)

The new BOINC database-server is:

Sun V40z + Sun StorEdge 3510:
Dual Opteron 844, 1.8 GHz (max 4 cpu)
8 GB memory (max 8 GB/cpu)
RAID-array, max 876 GB unformatted if 15k-disks, 1.72 TB if 10k-disks. (not specified what Seti is using)

Gads ... 😕 ... 1.72TB of SCSI drives

*drooooooll* 😀
 
I know some people that might help ya 😉...........though I can't remember if you've already been helped!😱

Originally posted by: Rattledagger

- Matt (Lebofsky)

The new BOINC database-server is:

Sun V40z + Sun StorEdge 3510:
Dual Opteron 844, 1.8 GHz (max 4 cpu)
8 GB memory (max 8 GB/cpu)
RAID-array, max 876 GB unformatted if 15k-disks, 1.72 TB if 10k-disks. (not specified what Seti is using)[/quote]

Man! sounds like their getting 1 whopping upgrade!:Q😀😎

 
Nothing "official" have been posted on either disk-speed on capasity of the RAID-array, the quoted is just the max for S3510 in single 2U-array. Looking more closely, it seems it's possible to use upto 8 expansion-packs in 1 RAID-array, this means max capasity is 7.8 TB or 15.7 TB depending on 15k or 10k-scsi-disks.

Since "done" wu is being purged from the BOINC-database, any science-info have already been copied to the science-database, it's doubtful they need so much capasity in the BOINC-database as 15.7 TB. 😉
So the most likely configuration is hopefully 15k-disks, meaning formatted capasity of 400+ GB since RAID 0+1 or whatever you're calling this...


Originally posted by: Wolfsraider
clear out the backlog of SETI@home classic data to be reduced within a few months

hmmm...

"Thus far we've run the results from 999 tapes through the result verifier. This represents 71.4% of the SETI@home database."
This indicates the "classic"-validator is over 1 year backlogged, so most likely "a few months" isn't how long till "classic" is killed off, but how long before the backlog is removed...
 
Originally posted by: Rattledagger
Nothing "official" have been posted on either disk-speed on capasity of the RAID-array, the quoted is just the max for S3510 in single 2U-array. Looking more closely, it seems it's possible to use upto 8 expansion-packs in 1 RAID-array, this means max capasity is 7.8 TB or 15.7 TB depending on 15k or 10k-scsi-disks.

Since "done" wu is being purged from the BOINC-database, any science-info have already been copied to the science-database, it's doubtful they need so much capasity in the BOINC-database as 15.7 TB. 😉
So the most likely configuration is hopefully 15k-disks, meaning formatted capasity of 400+ GB since RAID 0+1 or whatever you're calling this...


Originally posted by: Wolfsraider
clear out the backlog of SETI@home classic data to be reduced within a few months

hmmm...

"Thus far we've run the results from 999 tapes through the result verifier. This represents 71.4% of the SETI@home database."
This indicates the "classic"-validator is over 1 year backlogged, so most likely "a few months" isn't how long till "classic" is killed off, but how long before the backlog is removed...


Ah ok Sorry 🙁
 
Seti@home have split and distributed wu from 1317 tapes, and got many results in return.

But for various reasons it's not certain a result is "correct", example due to transfer-error, client-bug, hardware-problems (often due to too much overclocking) and cheaters. Before the results is used for anything except increasing the wu-score, they're therefore passed through a validation-process.

This "validation-process" have currently looked at 999 tapes, this is 71.4% of all distributed wu.
In other words, the seti-"classic"-validator is currently 28.6% backlogged.


From Technical news, November 17, 2004 - 20:00 UTC:
As part of SETI@home classic ramping down, we are busy cleaning up the old on-line database (an intemediary database that doesn't exist in BOINC) as well as preparing a new master database that will contain classic and BOINC SETI@home data on a system with higher capacity/throughput.
 
Back
Top