Humor me...I talk to myself

Nick Name

Administrator
USA team member
I am still unable to sit at my computer for more than five minutes due to swelling. I had MAKO™ TOTAL KNEE REPLACEMENT SURGERY on the 14th of December. I'm still crunching WCG.
Somehow I missed this post when it was made. I hope you are healing well.
We are expecting a snow/sleet/ice event in our area beginning Saturday night, 15 January. I anticipate that we will lose electrical power. I will shut down my two computers in order to conserve the APS Back UPS 1500 battery for operating our LP gas tankless hot water heater. The hot water heater needs 10 milliamps for the igniter. Without power, we will depend on LP gas logs in the fireplace and an LP gas heater. Our predicted high temperature next Monday is 32 degrees F. I know, it is springtime in Alaska for some, but it is cold for this area. Our Siberian Husky is loving it.
It looks like I will be on the northern edge of this system. We'll either get a few inches of snow or nothing depending on the track. Right now it looks like mostly rain if anything. I'm sure this sounds terrible compared to what you're used to. Somewhere DrBob is laughing. :LOL:
 

Vester

Well-Known Member
USA team member
I live in Northwest Florida, you probably dont wont to hear me complain about my weather right now....
I saw 12 degrees in Pascagoula back in the late '70s and snow for 1/2 day. At least you have white beaches instead of Mississippi mud, but you still get the humidity in the summer. I lived in Ocean Springs until 1991 when I returned to my hometown, Anderson, SC.
 

Gandolph1

Well-Known Member
USA team member
I grew up in North Central Indiana (Muncie) so I'm no stranger to cold and snow.
Anderson, SC - You may not believe this but I'm pretty certain I've been to your home town. I think there was a paper mill there that I had the pleasure of doing some control systems work at. As I recall a fairly small place by paper mill standards, I think they made tissue paper. Nice area!
 

Gandolph1

Well-Known Member
USA team member
I cant recall if they were called First Quality at the time or not. Most paper mills went through a significant period of consolidation where a lot of the small plants went out of business and those that remained were bought and sold as companies came and went in this very capital intensive industry.
 

Vester

Well-Known Member
USA team member
Both computers, 48 tasks, are crunching WCG again. The heat pump needs all of the help it can get. It is 29 F and snowing.
 

Vester

Well-Known Member
USA team member
We had about 4 inches of snow and little sleet. It is 31 F going for 20 F (-4C) by morning.
 

Nick Name

Administrator
USA team member
No new snow but it was a brisk 6 degrees when I got off work this morning. I don't mind the cold to a point but I don't really care for those single digit or negative temps.
 

Vester

Well-Known Member
USA team member
I've been lurking at the WCG forums and didn't have to wait long for the first posts from members who wanted to know why there is no more work available. :D
 

Vester

Well-Known Member
USA team member
I'm still crunching hundreds of WCG tasks in queue. I had increased the cache to 14 days. Now my "problem" is to meet the deadlines.
 

Nick Name

Administrator
USA team member
I like to keep a low cache so that I don't get an excessive amount of tasks, and if I want to switch projects on short notice I don't have to abort a lot of work. I generally don't like going past a day. It can be a big problem when you join a new project and BOINC doesn't have good estimates on how long the work takes.
 

Vester

Well-Known Member
USA team member
I usually keep a short cache also, but I increased it to 14 days to get plenty before WCG quit sending new work. I will run out of work after 8 days of crunching, 22 February, unless I catch some of the jobs that others fail to complete on time.

I am setup to run NFS@home until WCG has new work.

Hopefully, the price of video cards will be reasonable this summer and I can do more than CPU work.
 

doneske

Well-Known Member
USA team member
I ran out of WCG work yesterday afternoon so I decided to get VBox working on the systems again after kernel updates caused it to fail during the kmod process. Got it working on the Dell server (that was probably the first mistake) and decided to try some of the VBox work on Rosetta. I haven't done Rosetta in quite a while but had heard about the new VBox work. Connected to the project and set cache very low trying to limit to only active processors (no queue). Since this server has 128 threads, down came about 88 work units before everything essentially locked up. I have 2 disks in a RAID 1 configuration and both disk lights were on solid. BOINC client process was running 100% and BOINC Manager was totally unresponsive. Not being one to just accept the problem and reboot I decided to try to find out what the issue was as some OS commands were still working (top, ls, cd, etc). Definitely looked like a I/O issue so went to the /var/lib/boinc directory and started there. To make a long story short, the problem was due to BOINC copying the 7.5GB vdi file from the projects directory into each active slot directory (88 at the time when I suspended everything in BOINC). All 88 hadn't been copied yet by the time I looked at the disk usage. My disk that started out with about 50 GB used was up to 497GB used. Used the du command to get a sense of the 'damage' and sure enough each slot directory was at 8GB. I don't understand the need to copy the .vdi file to each slot. Why not just refer to it in the projects directory like other projects do with their executables. I have a 900GB disk but it isn't going to be enough to run more than about 90 slots. If you run the VBox work at Rosetta plan on having about 8GB of disk space per slot minimum. EX: 8 slots = 64GB of needed disk space
 

Vester

Well-Known Member
USA team member
I've been in a whirlwind of Linux installations on my main cruncher. Don't try this at your home. I was running Ubuntu 20.4 LTE and the BOINC benchmarks were/are the best I've found. BOINC Manager was running flawlessly. Then I got stupid and decided to upgrade to Ubuntu 21.10, and could not get BOINC to run due to some root privileges that changed in the upgrade. I tried to find the necessary files and change the privilege with "sudo chmod 700 PathOfYourFile " and must have not found the necessary files/folders. Since that failed, I downloaded Ubuntu 20.10, installed it, and added BOINC Manager. BOINC Manager from Ubuntu Software does NOT work in Ubuntu 20.10. I reinstalled Ubuntu 20.4 and got BOINC running again
This is the time for trying new things since WCG is not running. :D
 

Nick Name

Administrator
USA team member
Rarely have I found any upgrade to have any benefits for crunching, in fact it's usually a negative. The system I was happiest with was on some version of Ubuntu 14. I made the mistake of trying to upgrade. Something broke and by that time Ubuntu 16.x was out and I couldn't go back because I couldn't find a copy of what I had. Usually I save the download images but apparently this time I hadn't, or I lost track of it. I eventually got that system running again but not as satisfactorily as it was. I only upgrade now for a compelling reason, like a newer application that requires a driver upgrade. Otherwise I stick with what's working.
 
Top