Feature Request: support RAID for TrueNAS widget pool metrics #3493
Replies: 18 comments 36 replies
-
|
Thanks for posting a discussion about homepage here on GitHub. Homepage now uses a fancy new AI tool thanks to our friends at Glime, so the answer below is the AI bot's attempt to answer your question. If you feel it's sufficient you can mark the answer and close this post, otherwise feel free to leave it open of course! Thanks for using Homepage! ❤️ Hello, Based on the information provided, it seems that you have already configured the TrueNAS widget correctly with the However, there are a few things we can check to troubleshoot the issue:
If after checking these points the issue persists, it might be a bug on Homepage's end or an incompatibility with the latest TrueNAS SCALE version. In that case, I would recommend reporting the issue on the Homepage GitHub repository, providing all the details and screenshots you've shared here. If you need further assistance, please let me know! This comment was generated by Glime. |
Beta Was this translation helpful? Give feedback.
-
|
TrueNAS API keys do not have permission settings. The API point is clearly correct, otherwise the Homepage widget wouldn't be pulling any data at all. The container logs don't have any errors. |
Beta Was this translation helpful? Give feedback.
-
|
Could someone help me figure this out, please? |
Beta Was this translation helpful? Give feedback.
-
|
I'm not the original author but this is the response from /api/v2.0/system/info |
Beta Was this translation helpful? Give feedback.
-
|
and the one from /api/v2.0/pool |
Beta Was this translation helpful? Give feedback.
-
|
the widget does not care about possible raid settings which reduce the pool capacity. It simply adds free and allocated which is the capacity of the pool without raid |
Beta Was this translation helpful? Give feedback.
-
|
the correct formula (for raidz1) would be pool size = (number of disks - 1) * ((free + allocated) / number of disks) all needed infos are available in the response from truenas in some form |
Beta Was this translation helpful? Give feedback.
-
|
adding something like this into pool.jsx in the truenas widget works (only tested for RAIDZ1). however the actually used space is still wrong. I can't figure out if the problem in that area is a problem in the truenas api, the truenas ui or the calculation of the homepage widget does someone have an idea? export default function Pool({ name, free, allocated, healthy, data, nasType }) {
let total = 0;
let totalDriveCapacity = free + allocated;
let numberOfDrives = data[0].children.length;
if (nasType === "scale") {
if (data[0].type === "RAID0") {
total = totalDriveCapacity;
} else if (data[0].type === "RAID1") {
total = totalDriveCapacity / numberOfDrives;
} else if (data[0].type === "RAID5") {
total = totalDriveCapacity * (1 - 1 / numberOfDrives);
} else if (data[0].type === "RAID6") {
total = totalDriveCapacity * (1 - 2 / numberOfDrives);
} else if (data[0].type === "RAID10") {
total = totalDriveCapacity / 2;
} else if (data[0].type === "RAIDZ1") {
total = totalDriveCapacity * (1 - 1 / numberOfDrives);
} else if (data[0].type === "RAIDZ2") {
total = totalDriveCapacity * (1 - 2 / numberOfDrives);
} else if (data[0].type === "RAIDZ3") {
total = totalDriveCapacity * (1 - 3 / numberOfDrives);
} else {
total = free + allocated;
} |
Beta Was this translation helpful? Give feedback.
-
|
can back this up. have the same problem |
Beta Was this translation helpful? Give feedback.
-
|
I found something even better. However my javascript and my understandig of this service are not good enough to implement it cleanly: you can get the name and the health status of all pools like it's done now. however you can then send a second api request to note: the api values are in bit. so maybe a conversion would be good I can recommend trying the following curl command: this will return only the needed values for the pool size of that id |
Beta Was this translation helpful? Give feedback.
-
|
Having this issue as well. |
Beta Was this translation helpful? Give feedback.
-
|
I have the same issue and I have a stripe pool.. and even if I add the reported disk sizes it's still off by 200GB |
Beta Was this translation helpful? Give feedback.
This comment has been hidden.
This comment has been hidden.
-
I thought I did: isn't that "usable capacity" in my reply above what you're looking for? This is what TrueNAS reports as the total size of the pool. |
Beta Was this translation helpful? Give feedback.
-
|
In a reply above @seekermarcel pointed out a way to get the values easily and I can confirm:
{
"id": "data",
"type": "FILESYSTEM",
"name": "data",
"pool": "data",
"encrypted": false,
"encryption_root": null,
"key_loaded": false,
"used": {
"parsed": 10448456291712,
"rawvalue": "10448456291712",
"value": "9.50T",
"source": "NONE",
"source_info": null
},
"available": {
"parsed": 20801055061824,
"rawvalue": "20801055061824",
"value": "18.9T",
"source": "NONE",
"source_info": null
}
} |
Beta Was this translation helpful? Give feedback.
-
|
Correct. In my view that would be an easy way to implement this feature @zkvvoob I tried it myself but my js was too bad |
Beta Was this translation helpful? Give feedback.
-
|
Yes I saw that but it could potentially be a bunch of unnecessary API calls. We aggressively try to keep the amount of requests low. If I can’t easily figure it out with some math then perhaps we will have to look at doing it that way |
Beta Was this translation helpful? Give feedback.
-
|
Yea it seems like the truenas calculation has access to some information about usable space that it does not provide via the So this is close, but still a bit off: Taken from Calculated from If someone is interested in asking over in e.g. a truenas forum, perhaps they can clarify |
Beta Was this translation helpful? Give feedback.
This comment has been hidden.
This comment has been hidden.
-
|
Has there been able to be any more progress on this? |
Beta Was this translation helpful? Give feedback.
-
|
Was mostly waiting on an answer to this:
|
Beta Was this translation helpful? Give feedback.
-
|
I want to take a moment to share my understanding of TrueNAS, which may help with this issue. Disclaimer: I am not affiliated with TrueNAS, nor am I an expert with TrueNAS; this is just my understanding. I believe /api/v2.0/pool is analogous to "zpool list", which provides a list of the pools. These numbers are in TiB units; they are based on raw disk size and do not take into account mirroring, RAID, etc. In the example below, I have 16 2TB drives, so SIZE = 16 * 2TB = 32TB; 32TB * 1000^4 / 1024^4 = 29TiB I believe /api/v2.0/pool/dataset is analogous to "zfs list", which provides a list of the datasets. These numbers are also in TiB units; but they are based on actual usable disk space (i.e., with mirroring, RAID, etc. taken into consideration). I haven't done programming in years so I can't assist with that, but I think you can get the info you need via /api/v2.0/pool/dataset. Ignore everything except the top-level datasets (i.e., the ones that don't have slashes in them). Take the "used" and "available" raw values for the top-level datasets, then do the GiB calculation so you can display it however you want (e.g., 12T, 11.7T, 11.70T, etc.) rather than relying on the TrueNAS pre-formatted values. Size = Used + Available. In the little bit of testing I did, the API numbers match the numbers from the "zfs list" command which match the numbers on the TrueNAS Dashboard. In my opinion, the dataset values would be far more useful on HomePage than the pool numbers that are currently being displayed, In my case HomePage gives the appearance that I have 13.7TB free, when in reality I only have 8.66T free due to RAID. I would also recommend using the XiB units rather than XB units so that it aligns with the way TrueNAS displays it. Personally, I'd prefer to see Available / Total (xx%) rather than Used / Total (xx%); it saves me from doing the math to find out how much space I have left, which is the number that I really care about. Also, as a side note, ZFS RAIDZ# is not RAID-# (e.g., ZFS RAIDZ1 is not RAID-5). Functionally they are similar (e.g., both ZFS RAIDZ1 and RAID-5 can survive the loss of a single drive), however ZFS uses a more complex method of handling the parity bits. Because of this, you can't do the calculation by simply looking at the number of drives and the RAID configuration, then subtracting the appropriate number of parity drives from the total; it doesn't work that way with ZFS. |
Beta Was this translation helpful? Give feedback.
-
Interesting. Can you post the actual api output for that endpoint? The other question is, will, by definition, all users of TrueNAS pools be able to use the |
Beta Was this translation helpful? Give feedback.
-
|
Is there a particular piece you're wanting to see? I have a lot of datasets so the output is over 33k lines and I would need to sanitize it before posting it. I can't answer your question definitively, but I would think so. I don't have any pools that don't have any datasets defined, but even in that scenario I would think you would still get the top-level dataset associated with the pool. If you don't have pools I would presume you would get nothing, which is what I would expect from /api/v2.0/pool as well. |
Beta Was this translation helpful? Give feedback.
-
|
Ha, yea that’s a lot. I would just need enough of the output to be able to emulate the api to work on the changes. I guess that would be at least one or two “top-level” datasets and yes please redact whatever of course, all without changing the structure of the overall json returned by the api |
Beta Was this translation helpful? Give feedback.
-
|
First, a quick heads up regarding some info I ran across while digging into this. TrueNAS Core is currently at v13.3; TrueNAS Scale is currently at v24.10. Documentation for both products can be found on the TrueNAS Documentation Hub at https://www.truenas.com/docs/. If you expand the "TrueNAS" section of the Documentation Hub, you can find documentation for various versions, and under each version is an associated API section. Starting with v24.10 (and including development v25.04), there is a REST API Deprecation Notice dated 11/15/2024: The TrueNAS REST API is deprecated in TrueNAS 25.04 and replaced by the TrueNAS Websocket client. Full removal of the REST API is planned for a future release. This new API Client is not the deprecated TrueNAS CLI (midcli). The API Client is integrated in TrueNAS 25.04 onwards. It provides the midclt command-line tool, and the means to easily communicate with middleware using Python to make calls through the Websocket API. This API client allows for better integration of TrueNAS into third-party solutions. Use this as a reference for projects that require direct TrueNAS integration. TrueNAS API documentation remains available in the UI and through the TrueNAS Documentation Hub. This is a bit over my head, and I don't know if you're using REST or Websocket, but this might be something you want to take into consideration as it looks like the future direction for TrueNAS API access. There are query-filters and query-options features that might make retrieval of the information needed easier and quicker, but they are only available via Websocket, not via REST. In particular, query-options.extra.flat provides a hierarchical response where only the top level datasets are returned in the list. From the API documentation for pool/dataset: At present, via REST, the flat structure looks like this (I have stripped out most of the data so you can see the overall structure: p01 is the pool As you can see, each dataset is listed as a separate object, and each object also contains all of the data for it's children, so there is a lot of repetitious data, thus it's slower. I have around 30 datasets and my /api/v2.0/pool/dataset response is 32k+ lines (665 KB). Unfortunately without using Websocket, I don't think there is a way to retrieve the data needed in a more streamlined way (i.e., with the aid of query-options.extra.flat). So, continuing with REST, you're basically looking for the top level pool "id"s (i.e., the ones that don't have any slashes in them) such as The original JSON was produced by curl -s -X GET -H "Authorization: Bearer 1-XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX" "https://(IPADDRESS)/api/v2.0/pool/dataset" This is the beginning of the JSON output, which starts with p01, the first of two pools. This shows the iocage dataset on p01. This shows p02, the second pool. To make it valid JSON, get rid of that last comma and add a closing brace ( ] ) instead. You just need the "id", "used" and "available" raw values from the top level pool ids; size=used+available. In theory, you could provide monitoring per-dataset if you wanted. The "used" value for each dataset includes the dataset itself plus all of it's children. "Available" is provided for each dataset, but it will be the same for all datasets on a given pool unless per-dataset quotas are in effect. I only use quotas on a couple of datasets so this isn't a huge need for me, but others may find it helpful. Regardless, you'll always ignore the children and just look for the top level datasets as each dataset will have a top level object in the JSON. |
Beta Was this translation helpful? Give feedback.
-
|
Thanks, I’m aware of the api docs, they don’t provide examples for most endpoints. Screenshots of the json makes it pretty cumbersome to copy / paste. If you (or anyone else) are willing to share a copy / pasted text (or on a gist) of the actual output that would be helpful. I suspect there isn’t actually a ton of sensitive information and it can be easily sanitized, but of course I don’t know for sure. |
Beta Was this translation helpful? Give feedback.
-
|
Oh, and homepage has no mechanism for websocket support, so that’s a bummer |
Beta Was this translation helpful? Give feedback.
-
|
I figured it was quicker to spin up a VM with TrueNAS than to read through 32k lines of JSON and potentially miss something that should have been redacted, so... Here is the API output from /api/v2.0/pool/dataset from a new TrueNAS server. The server has two pools, pool1 and pool2. Each pool has a single dataset, dataset1 on pool1 and dataset2 on pool2. curl -s -X GET -H "Authorization: Bearer 1-XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX" "http://(IPADDRESS)/api/v2.0/pool/dataset" |
Beta Was this translation helpful? Give feedback.
-
|
That unfortunate about websocket support in homepage. Considering that it's being deprecated in 25.04 which isn't even out yet, I suspect it will be a bit before it's removed, but at some point it will be. |
Beta Was this translation helpful? Give feedback.
-
|
Awesome thanks, I’ll play with it |
Beta Was this translation helpful? Give feedback.
-
|
Annoyingly the dataset endpoint doesnt include the healthy property, so I've combined the two. Im curious if anyone would like to test the docker image tagged It would also be important to test on Truenas scale vs core (whatever the heck the difference is) 
|
Beta Was this translation helpful? Give feedback.
-
|
I copied my existing homepage data to a new homepage-test and created a new container for it. That worked fine, so I changed the tag to grab the new version. Now I get this when I try to open homepage. Tried clearing browser cache. Did I miss something to cause this? Something went wrong. |
Beta Was this translation helpful? Give feedback.
-
|
Regarding TrueNAS Core vs TrueNAS Scale (at a very high level)... |
Beta Was this translation helpful? Give feedback.
-
|
Thanks for the explanation, I just mean no real idea if that will matter for this stuff. A new image is building, should be about ~20-30 minutes, whenever you have the chance to re-test, thanks ( https://github.com/gethomepage/homepage/actions/runs/12979837367 ) |
Beta Was this translation helpful? Give feedback.
-
|
Ok. Don't know if I'll get to it tonight. Also, I'm spinning up a TrueNAS Scale so I should be able to test both platforms. |
Beta Was this translation helpful? Give feedback.
-
|
Cool, yea no rush at all |
Beta Was this translation helpful? Give feedback.
-
|
That fixed the semver issue. Thanks! I'm seeing something odd that doesn't make sense to me. The Used, Total, and Percentages for zfs list, the API output, and Homepage are matching for both Core and Scale, except for one: The Total (and thus the %) for the first pool on TrueNAS Core does not match, and I cannot seem to figure out where Homepage is getting the Total for the first pool from; it doesn't seem to match anything. On my test server, it should be 15.9GB according to zfs list and /api/v2.0/pool/dataset, but Homepage is saying 28.5GB. What's wierd is that the total for the second pool on Core is correct, and the totals for both pools on Scale are correct.
|
Beta Was this translation helpful? Give feedback.
-
|
Is that your real server? If its a test one maybe you can send me the api output again =) The logic is pretty simple, for each pool in Perhaps some of the children need to be filtered out? Or should it just be the |
Beta Was this translation helpful? Give feedback.
-
|
Oh yea I think I misunderstood your post above, it should just be the top level |
Beta Was this translation helpful? Give feedback.
-
|
The docker image is re-building with that change: https://github.com/gethomepage/homepage/actions/runs/12984000399 (again, takes ~20 min, no rush) |
Beta Was this translation helpful? Give feedback.
-
|
Yes...ignore the children. The pool dataset's used and available values already include the children. You have it correct now. Looks great! Everything matches perfectly! Here are some screenshots. Excel spreadsheet with the details. Top section is based on byte counts obtained from zfs list command of TrueNAS; middle section from the API; bottom section from the revised Homepage. The numbers are presented in both TB/GB (the old way) and TiB/Gib (the new way that matches how TrueNAS presents them). The OLD Homepage; incorrect values, used TB/GB instead of TiB/GiB. The NEW Homepage: correct values, uses TiB/GiB which matches TrueNAS values. Corresponding displays from TrueNAS UI. Bonus screenshot showing degraded pools. |
Beta Was this translation helpful? Give feedback.
-
|
The new version (tag = feature-truenas-dataset-3493) looks good on my test and production TrueNAS systems. Can OP confirm? The only other suggestion I have on this would be to display Available / Total rather than Used / Total. TrueNAS Core UI displays available space; TrueNAS Scale UI displays both used and available. Both display the percentage as %Used. I suspect that is because performance can start to degrade once you reach 80% Used so TrueNAS wants you to see that percentage. For me, when I glance at status screens like Homepage, seeing how much space I have remaining (i.e. Available) is more useful than knowing how much I've used. Pool sizes can vary, but empty is always zero, so the closer Available gets to zero, the more urgent doing something becomes vs having to look at Used vs Total to make that determination. The Used percentage is helpful, but it's relative based on the size of the pool (e.g., a large pool could be 90% Used but still have 10s of TiB available). Just my opinion, but I'd be interested in hearing others' perspectives on this. I supposed you could display both as TrueNAS Scale is doing, but I think that adds clutter, and with long pool names it could cause display issues with the limited amount of space available in the widget. |
Beta Was this translation helpful? Give feedback.
-
|
Awesome, glad it works. Yes it would be good if as many people can confirm as are able. I dont feel strongly about the change free vs used, we can see if others do but I will say from a UI perspective looking at the progress bar it makes more sense to me to show used / total, and then have the text and % match that Also, tbh, I've spent a lot of time on this and would like to focus on the main issue about stats being incorrect before debating this little thing too much, but again, happy for others to chime in |
Beta Was this translation helpful? Give feedback.
-
|
Agreed. Sorry, I didn't mean to introduce scope creep. I believe the issue at hand has been resolved pending confirmation by others. If Available vs Used is something folks would like to see changed, I'd suggest we submit it as a feature request for implementation when priorities allow. |
Beta Was this translation helpful? Give feedback.
-
Description
Hello,
I noticed that the TrueNAS widget is not reporting the size and used space of my pools correctly. Here's what I have in Homepage:
And here are snapshots of the two pools from within TrueNAS's interface:
My
settings.yamlcontains this:and I'm running TrueNAS-SCALE-23.10.0.1.
Do you know if this is caused by their API or is it a bug on Homepage's end?
Thanks!
homepage version
v0.8.13 (986a181, May 6, 2024)
Installation method
Docker
Configuration
No response
Container Logs
No response
Browser Logs
No response
Troubleshooting
N/A
Beta Was this translation helpful? Give feedback.
All reactions