Cerbo GX Timezone display wrong, reboot issue with SD card

I just spent 15 minutes trying to find a contact where to submit bug reports. A community forum is not responsible to deal with product bugs.

This morning (switch to european summer time) I noticed that the time on the gui-v2 is wrong (fw 3.71). In gui-v1 it is displayed correctly and the device’s time settings are also correct. Timezone is set correctly, and location in VRM portal is also set correctly.

Furthermore, when rebooting the Cerbo GX to see if this problem “resolves itself”, the Cerbo GX didn’t come back up again and I lost power to the whole compound. It turns out that it would not boot with the SD card inserted.

How do I report bugs?

It is getting crazier and crazier. When opening the Cerbo page on my mobile I get some random time 11h20m ahead. This is clearly broken. Opening the page from my computer shows the 1h difference. gui-v1 still shows the correct time.

Hi,

GUI v2 timezone presentation depends on browser setting.

You can check browser based timezone by test page like (need JavaScript):

Yes, my timezone is perfectly correct, according to this test page. Still gui-v2 is displaying the wrong time (-1h).

On the mobile it is now showing the correct time. What the..

First, verify that UTC time is correct. (Settings → General → Date & Time → Date/Time UTC)

In GUI v1 timezone from Cerbo GX setup is used. ( Settings → General → Date & Time → Timezone)

In GUI v2 “header” time is formed by browser timezone information (which is future proof as every (time) presentation should be handled on client side)

The time shown as UTC in gui-v2 and gui-v1 is one hour behind, which is correct for DST and my timezone (Lisbon).

So the timezone calculation is broken. I am using librewolf 148.0.2 and I even restarted it now to see if it might have used some cached values from before the summertime switch, but it still shows the same offset. Chromium/Android shows random garbage (see screenshot above) or the correct time, Edge on Windows shows the correct time.

Besides: Why would I want to see my local time on a remote control? All activities, logs etc. are in the system’s local time. This doesn’t make sense.

Regarding the other two topics:

  1. How do I report bugs? Here in the forum?

  2. The Cerbo GX does not boot with a SD card inserted. I am using a SD card I had lying around from an old drone, it still contains a DCIM and MISC folder. With this SD card inserted, the Cerbo GX does NOT BOOT. If I insert the card during operation, data is stored on it as expected. When trying to reboot the Cerbo GX, I lost power to the whole compound until I figured out what was going on, this is not funny. No matter what is in the SD card slot, the Cerbo GX has to boot. I have made an image copy of this SD card for debugging. If I insert a SD card and it works, I don’t expect the Cerbo to fail booting weeks/months later.

Still broken in fw 3.72.

I have to say, finding out after buying 3 Multiplus-II 10k, one RS Solar, a series of Lynx devices and one MPPT that there is no way to submit bugs and no support is a bit unsettling. I have to re-evaluate Victron products for future purchase decisions.

Considering the sheer number of users, and that no one is joining in with you, it may be worth considering what is unique to your installation/devices?
We have seen the occasional time bug over the years, but for the most part, it hasn’t been an issue recently.

I am almost certain Librewolf is not one of the supported browsers.
What version android are you using? What device?

As for the SD, also thousands of those in use, the system is probably expecting a clean formatted card with no other folders. SD’s can be used for recovery as well so whatever else is on there may be confusing it. The easy solution is to reformat it as per the documented specs.

We have seen USB thumb-drives cause this behaviour, or rather, delay startup.

As for support that is provided via the distributor → supplier → installer chain.
Tickets can be logged via Victron’s support link documented in our community guidelines.

For the most part when an issue appears that we can see affects multiple users then we can try raise awareness but unfortunately, too many bugs tend to have external factors.

Librewolf is basically Firefox and it worked fine until sunday. Except for the Cerbo GX there is not a single issue anywhere, and as mentioned above the Geolocation test page confirms the correct settings. This is not a critical, but a mere cosmetical issue, but I also question the idea behind it to map device time to local time. This is done nowhere else in IT, and for good reasons.

Regarding the SD card: It doesn’t matter what the system expects, it cannot STOP BOOTING. This is a critical issue as it causes a system failure days/weeks/months after inserting the SD card.

What spec are you referring to? The manual just states

“The internal buffer can hold logs for several days. To extend this period, insert a microSD card or USB stick. Check internal storage status via the settings menu. When external storage is inserted, internally stored logs automatically transfer to it, ensuring no data loss”, and

“Storage Device Requirements
• Supported file systems: FAT (12, 16, 32), exFAT, ext3, and ext4.
• MicroSD cards (SD and SDHC types) up to 32 GB typically come pre-formatted with FAT12, FAT16, or FAT32 and can be used
immediately. Avoid reformatting them to unsupported file systems”

I don’t see any requirement for it not to contain a DCIM folder. It was recognized and used by the software, so it is fine. I know that I can reformat the card, that is not the issue. The issue is that the system FAILS. If it was a delay, it was longer than 10 minutes, because that’s when I started cold booting the Cerbo GX.

I have seen the support chain on the Victron web page. I am not reporting a bug regarding the handling of SD cards to a spanish solar shop, this is not an installation or hardware issue. I checked the support page, and I didn’t find anything to open a ticket or submit a bug. Googling coughed up nothing. That’s why I wrote in my first paragraph that this is the wrong place to submit this issue. I have just re-read the community guidelines, there is not a single mention of tickets.

Don’t get me wrong, I appreciate you guys replying, but there is clearly something wrong here if I am not able to find the spot to place a ticket. It might be me, of course :slight_smile:

While I don’t disagree with you, reformatting it cleanly so it doesn’t have folders causing it to stop, is a quick easy fix, vs trying to get a dev’s attention to work out why it stopped.

Can you just confirm you are on a current-ish version of Android?
There must be a reason why yours is misbehaving and others are not. Eliminating stale OS’s (there has been a lot of change) is one of those steps.

Sorry, that’s the wrong procedure. The Cerbo boot process has to be made more robust, because if it happens here, it will happen anywhere else. Anybody can insert a card that still contains some random data or somebody with malice intentions can prepare a card in such a way that a system will run into a fail days/weeks later. This is basically a Denial-of-service attack (CVE 4.0-6.9 due to local access requirements according to CVSS).

I come from high availability industrial embedded device development, and when I see something like this it makes my head explode.

@AndroidAndroidAndroidAndroid: It was an S23 with latest Samsung fw, unfortunately there was an update yesterday followed by a reboot, so I don’t have the previous version anymore. There is no stale software anywhere here, rather the opposite - bleeding edge everywhere because as developer I want to see bugs before my clients/users do. Just tested, at this moment it is displaying the correct time.

I poked a bit around in the gui-v2 JS code, and in the websocket stream in one of the initializing packets I see N/c061XXXXXXXX/settings/0/Settings/System/TimeZone{“value”:“Europe/London”}, so at least that is correct (what is this, MQTT over Websocket? :). Couldn’t find anything regarding offset, DST etc. No errors in the console. No idea where the wrong information is coming from.

Update: chromium-147.0.7727.24 also displays the time one hour off.

Just to provide a counterpoint, most logging and SIEM tools do just that: They ingest and store time in UTC and display it in the user’s browser timezone.