Important Storagenode Fix
v0.31.10 made it so that storage node will not accept more pieces if they are full. The used space calculation was still a bit wrong. v0.31.12 should fix the used space calculation. Again my apologies. Let us know if you have any issues.
7 Likes
In the next few minutes we will deploy the satellite with a new minimum storage node version. Please update your storage nodes because otherwise you will not get any uploads.
3 Likes
You guys are fast today!
3 Likes
Lets say we learned today a lot about which process we have to optimize and what is already working great.
3 Likes
Always good to learn something. Even though it might not always be pleasent.
own experience, is best teacher, take big price, but result is excellent.
Some new (to me) errors after the update to 0.31.12:
2020-01-30T20:30:38.630Z ERROR piecestore:cache error getting current space used calculation: {"error": "lstat config/storage/blobs/ukfu6bhbboxilvt7jrwlqk7y2tapb5d2r2tsmj2sjxvw5qaaaaaa/3v/}
2020-01-30T20:30:41.197Z INFO piecestore download started {"Piece ID": "NPAZAWHQ7DESYGVNACEQ5NVYCYLJ2FCGM4VCASMTMEYUGDFD3IFA", "Satellite ID": "118UWpMCHzs6CvSgWd9BfFVjw5K9pZbJjkfZJexM}
2020-01-30T20:30:46.139Z INFO piecestore downloaded {"Piece ID": "EWVJJRU3E5DLFAIL6GIESUJSMC3GRFGY5HP5IQVTL4Z7DNM4CQFA", "Satellite ID": "118UWpMCHzs6CvSgWd9BfFVjw5K9pZbJjkfZJexMtSkmKxvv}
2020-01-30T20:30:47.106Z ERROR blobscache newPiecesTotal < 0 {"newPiecesTotal": -329984}
2020-01-30T20:30:47.106Z ERROR blobscache newPiecesContentSize < 0 {"newPiecesContentSize": -329472}
2020-01-30T20:30:47.106Z ERROR blobscache newSatPiecesTotal < 0 {"newSatPiecesTotal": -329984}
2020-01-30T20:30:47.106Z ERROR blobscache newSatPiecesContentSize < 0 {"newSatPiecesContentSize": -329472}
Anything to be worried about? Seems to be working fine otherwise.
Edit: Web dashboard and SNOboard shows incorrect disk space. Showing used space starting at zero when starting the node. Increases as uploads come in.
There was used space calculation errors in the code. The current error displayed is to be expected. However, it would be concerning if it shows up again.
Storage Node with this version needs to recalculate the actual usage from scratch, hence itās starting from 0 for the time being. Once it finishes the walking all the files it should show the correct usage.
3 Likes
Okay, thanks for the quick reply. I will let it run and see what happens. Maybe this info can be added to the the first post @littleskunk ?
1 Like
have the uploads resumed, Iām v0.31.12 now for forty minutes now and my logs only get, get repair, put repair just curious
That one is an upload and it means everything is fine on your storage node.
ok great I was thinking plain put thanks
Iāve tried to do manual update with the instruction provided here but Iām still at 0.31.10
In order to manually update you need to remove the container and start a new container.
I think is what I did by following the provided instruction
I even try a few watchtower restart but it say a weird āWaiting for running update to be finishedā¦ā after the update to 0.31.10
Here the watchtower log
time="2020-01-30T19:21:55Z" level=info msg="Found new storjlabs/storagenode:arm image (sha256:d81cdf3ea58b88684029dfdd1b1eafb3154f2dcdf74ab918cbf0dc407dc93070)"
time="2020-01-30T19:21:57Z" level=info msg="Stopping /storagenode (c5a04eb40a2a7f7e1c8bb3712a59e293a0ebdde2b67614d235a16dd8955b344c) with SIGTERM"
time="2020-01-30T19:22:01Z" level=info msg="Creating /storagenode"
time="2020-01-30T23:02:22Z" level=info msg="Waiting for running update to be finished..."
time="2020-01-30T23:30:26Z" level=info msg="Waiting for running update to be finished..."Yeah it doesnt update by restarting watchtower
My Windows GUI updated fine and quickly to 31.12 from 31.10.
My previous update to 31.10 had paused and didnāt automatically restart in windows 10.
This update restarted fine. Itās good if the windows 10 system can update clean and quickly with these rapid updates coming out
everythings seems normal in the logs after the update. my downloads continue.
I get uploads but they all fail with ācontext canceledā error on my Raspberry Pi node. I have 86% used space.
1 Like