Travel Mapping
User Discussions => Welcome & Notices => Topic started by: Jim on January 19, 2021, 11:47:37 am
-
The hardware has arrived for memory and disk upgrades to the server that houses TM and METAL. I hope to install them this week, possibly as soon as tomorrow (Wed 1/20). There will be multiple complete outages of TM and METAL, including this forum. Watch this thread, the motd on the main TM page, and the project status on GitHub (https://travelmapping.github.io/) for updates.
-
memory
Been doing some work on C++ subgraph generation this past week where memory bandwidth plays an interesting role; it's made me wonder what noreaster's setup is.
So what's it got?
DDR4?
What mHz?
Dual-channel? Tri-channel? Other?
-
Been doing some work on C++ subgraph generation this past week where memory bandwidth plays an interesting role; it's made me wonder what noreaster's setup is.
So what's it got?
DDR4?
What mHz?
Dual-channel? Tri-channel? Other?
Current:
32GB (8x4GB) 2400MHz DDR4 RDIMM ECC
Adding 8 more of the same to bump to 64GB total.
-
I am planning to do the upgrades starting around 9:15 AM EST on Wednesday, January 20. Expect TM to be unavailable at times throughout the day.
-
Looks like the DDR4 modules are rated for 2400MHz, but the processor only supports 2133...
On noreaster we have this:
Dual Intel Xeon Processor E5-2630 v4 (10C, 2.2GHz, 3.1GHz Turbo, 2133MHz, 25MB, 85W)
The 2133 here appears to refer to the memory clock:
https://ark.intel.com/content/www/us/en/ark/compare.html?productIds=92981
...which answers my other question: it's 4-channel.
Not bad. That's nominally 178% the bandwidth of lab2, 267% the bandwidth of lab3.
noreaster should be able to churn out subgraphs wicked quick.
-
The disk is in, but not yet partitioned or in use. Turns out there was a misunderstanding on the memory, and it has not yet arrived. We're up at the moment, but I will be doing a lot of reconfiguring so expect ongoing outages this morning.
-
The database is now on the new SSD. This post is in part to make sure the forum still works, but also to say TM service disruptions will continue today.
-
Things still look good after a reboot with the DB and some other things moved to the SSD. The site should be stable for the next hour or two, then I'll be making more changes.
-
The main TM site and the forum are now living on the new SSD. Hopefully this post will work!
-
I'm still doing some rearranging and configuring, but I don't expect any of it to cause more than very brief interruptions to the operations of the main TM site or the forum. Please let me know if you notice any new problems that might be a result of today's disk upgrade.
When the new memory chips come in, we'll have another outage, but that should be a lot less disruptive.
-
Noticed there was a 'theme' link issue since you moved the forum to a different directory. I've just fixed that issue.
-
wpt editor does not work anymore (FF, Edge, IE): https://travelmapping.net/wptedit/
Leaflet issue?
-
no, a general issue - on our end? When I load data, the table does not apear
Not sure whether there is a connection to the HW update. It worked yesterday but I had wpt editor opened during the upgrade.
Of course, no cache issue etc. (automatically deleted on closing browser)
-
[Exception... "Favicon at "https://travelmapping.net/favicon.ico" failed to load: Not Found." nsresult: "0x80004005 (NS_ERROR_FAILURE)" location: "JS frame :: resource:///modules/FaviconLoader.jsm :: onStopRequest :: line 253" data: no] 2 FaviconLoader.jsm:253:22
wpt editor + mapview + showroute do not work - because of favicon failure!
-
Map display not working in regions or individual route pages.
-
wpt editor does not work anymore (FF, Edge, IE): https://travelmapping.net/wptedit/
Leaflet issue?
It works on Chrome but the points on the map are not shown correctly.
-
HB doesn't seem to be working right. See for example https://travelmapping.net/hb/showroute.php?units=miles&u=tckma&r=de.de001
From chrome://version --
Google Chrome 87.0.4280.141 (Official Build) (64-bit) (cohort: Stable)
Revision 9f05d1d9ee7483a73e9fe91ddcb8274ebcec9d7f-refs/branch-heads/4280@{#2007}
OS Windows 10 OS Version 1909 (Build 18363.1256)
JavaScript V8 8.7.220.31
Flash 32.0.0.465 C:\Users\<removed my employee number>\AppData\Local\Google\Chrome\User Data\PepperFlash\32.0.0.465\pepflashplayer.dll
-
I can confirm that the waypoint browser isn't working on Chrome, Firefox, or (ew) Edge on my computer. "URL Placeholder" is shown in red text and the route shield is shown, but no table or map is generated.
-
I can confirm that the waypoint browser isn't working on Chrome, Firefox, or (ew) Edge on my computer. "URL Placeholder" is shown in red text and the route shield is shown, but no table or map is generated.
Ditto on Brave, Safari and Firefox on my iMac.
-
I'm not having any problems with the HB in my Chrome browser. I haven't cleared the cache, wishing to keep whatever functionality I have.
The HB doesn't work in Microsloth Edge on my computer, through.
-
Leaflet must have been cached in my browser for yesterday's testing. I've put leaflet where it needs to be, and so far so good on my end. Let me know if I've fixed it for all of you as well.
I'm sure there will be other problems, hopefully not as significant as this one!
-
HB working again here on my end Jim. 8)
-
Now works for me in Edge, too.
-
Yup, that fixed it. Thanks.
-
I suppose next time I move the files on the web server, I shouldn't spend the next day away from the computer. But the skiing was fantastic!
-
wpt editor does not work anymore (FF, Edge, IE): https://travelmapping.net/wptedit/
Leaflet issue?
It works on Chrome but the points on the map are not shown correctly.
The WP Editor on Chrome is back to normal.
-
Yep, working again.
-
Leaflet must have been cached in my browser for yesterday's testing.
Always test in a 'private' window, as that doesn't use the cache. ;)
-
I think everything is back to normal - wish we could say that for our lifes too :)
-
Turns out there was a misunderstanding on the memory, and it has not yet arrived.
When the new memory arrives and noreaster has to go offline, I don't suppose you could humor me (https://github.com/TravelMapping/DataProcessing/issues/375#issuecomment-764962926) and do a round of memtest86 with the existing DIMMs? :D
-
Turns out there was a misunderstanding on the memory, and it has not yet arrived.
When the new memory arrives and noreaster has to go offline, I don't suppose you could humor me (https://github.com/TravelMapping/DataProcessing/issues/375#issuecomment-764962926) and do a round of memtest86 with the existing DIMMs? :D
Can probably make that happen.
-
I'm finally planning to complete the hardware upgrades by doubling the server's RAM on the afternoon of Tuesday, March 16. Down time shouldn't last long, but TM, the forum and METAL will all be down during the upgrade.
-
Heading to the machine room shortly for the memory upgrade. Expect several minutes to maybe a half hour of down time between 1:00 and 2:00 PM EDT today.
-
What was supposed to be a few minutes of downtime to pop in the new memory chips turned into about an hour tracking down the fact that 2 of the 8 memory chips were a slightly different part and not compatible with the other memory in the system. The good news is that the system at least went from 32 GB of RAM to 56 GB, though not the 64 GB it was intended to be. Everything should now be back to normal. There will be a hopefully shorter outage whenever the correct chips finally arrive.
-
Oh fun. Sounds like you're not getting the full bandwidth benefit of n channels then?
-
Oh fun. Sounds like you're not getting the full bandwidth benefit of n channels then?
Probably not. Hopefully the two misfits will be replaced fairly quickly.
-
When the new memory arrives and noreaster has to go offline, I don't suppose you could humor me (https://github.com/TravelMapping/DataProcessing/issues/375#issuecomment-764962926) and do a round of memtest86 with the existing DIMMs? :D
Can probably make that happen.
Did this happen?
The mysterious C++ Pointer Value Loses 0x900000000 bug is back (https://github.com/TravelMapping/DataProcessing/issues/442).
It's weird enough that I can't write off bad memory as a potential cause.
-
When the new memory arrives and noreaster has to go offline, I don't suppose you could humor me (https://github.com/TravelMapping/DataProcessing/issues/375#issuecomment-764962926) and do a round of memtest86 with the existing DIMMs? :D
Can probably make that happen.
Did this happen?
The mysterious C++ Pointer Value Loses 0x900000000 bug is back (https://github.com/TravelMapping/DataProcessing/issues/442).
It's weird enough that I can't write off bad memory as a potential cause.
We didn't do any additional testing. I still haven't acquired the 2 additional ones to get up to the full 64 GB.
-
The Bad RAM Hypothesis is ruled out.
-
I have probably mentioned this before, but I spent a ridiculous amount of time in grad school tracing a bug that wound up being a bug in IBM's compiler, not my code...
-
Haha, I've run into one compiler bug (https://github.com/TravelMapping/DataProcessing/issues/327) already. At least that time I had something to Google; this time not so much.
Tracking down whatever this is ain't gonna be pretty.
I guess one relatively low-effort thing I can try is to install FreeBSD 13.0 and see if there's a new version of clang & things behave differently...
-
Still waiting on those two DIMMs?
-
Still waiting on those two DIMMs?
They were unavailable for a while and I never pursued it further. Maybe I still should.