MySQL server has gone away during upgrade #8863
Comments
|
Googling |
|
@rullzer @nickvergessen I really don't like that we more often get the "MySQL server has gone away" messages ... it seem to be not reasonable that all the mysql servers to stop working right in the middle of the updates :/ This looks really 🐟y |
MorrisJobke
added
bug
0. Needs triage
|
@MorrisJobke It seem to be getting more indeed. Maybe recent mysql/mariadb updates shortend the timout of connections? |
|
@splitt3r I have not yet ruled out the possibility that of it being related to NCP, though I haven't seen the exact same issues reported over there. @rullzer in terms of timeouts this is what I have:
These "MySQL server has gone away" errors are causing me problems during normal operation also, see |
|
@leavenotracks What is the |
|
That means 16 MB ... 🤔 ... @rullzer @nickvergessen could this be a reason? But 16 MB is really a lot. :/ |
|
in theory sure. but in practice 16MB should be more than enough. Also it seems to fail on setting a config value. I find it very unlikely that this results in a packet of more than 16MB. |
|
Not to redirect away from the app, but raspberry pi's use sd cards for storage by default, correct? If this user is using an sdcard, couldn't this be caused by various io issues? Assuming the db is on the pi. |
|
I am on debian testing here, mariadb 10.1.29, on a machine which is comparable in slowness to the raspi, and I upgraded to 13.0.1 with the web updater just fine without mariadb crashing. |
|
@user23498723452 The nc database sits on a USB HD, along with the data. @LollenJumplan I was able to upgrade to 13.0 from 12.0.5 via the web updater without any issues. |
|
I tried to upgrade from 13.0.0 to 13.0.1 via the web updater. But the update did not finish completely. When I open Nextcloud in the browser, I get this message: I get this exception when trying to set the maintenance mode to off: MariaDB version: 5.5.57-MariaDB |
|
I made a small modification to the code to fully throw the original stack trace. Here is part of it: |
|
Further on the issue I had, it has been solved now. I manually set When I accessed Nextcloud in the browser after changing this setting the update was triggered again, and went smoothly. Everything seems to work fine again. Concluding: somehow the value of |
|
#9065 had this:
and
|
MorrisJobke
changed the title
|
@IPlayForHer @janvlug @leavenotracks @LollenJumplan @user23498723452 @splitt3r Do you use SSL for the mysql connection? (see #3709) Maybe this only affects SSL mysql connections and this is the reason why we can't reproduce it. |
|
No I Turn SSL off
… On 4 Apr 2018, at 10:11 pm, Morris Jobke ***@***.*** ***@***.***>> wrote:
@IPlayForHer <https://github.com/IPlayForHer> @janvlug <https://github.com/janvlug> @leavenotracks <https://github.com/leavenotracks> @LollenJumplan <https://github.com/LollenJumplan> @user23498723452 <https://github.com/user23498723452> @splitt3r <https://github.com/splitt3r> Do you use SSL for the mysql connection? (see #3709 <#3709>) Maybe this only affects SSL mysql connections and this is the reason why we can't reproduce it.
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub <#8863 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AUSXH1NKY2SJVsXLkMWPFcJFl1zs6HHiks5tlLhfgaJpZM4Sumsj>.
|
|
I got this error while using the "download as zip" feature of a large folder. In that case I was able to fix it by increasing |
|
Hello. I tried to upgrade from 13.0.6 to 14.0.3. When I read the other postings correctly: there's still no solution to this, right? I ran updater.phar and it "told me" it did a backup. As I'm new to Nextcloud: where did updater.phar put the backup, so I can restore it, like described here? |
The updater only backs up the code itself but not the whole data nor database. But you could rerun "occ upgrade" and check if it works then. |
|
When I issue I don't know if this is correct, since I don't know the SQL language by heart. |
|
@icewind1991 @nickvergessen @rullzer Any idea what could be the problem here and why MySQL drops the connection? |
|
Do you also see this in your DB log? |
|
Yes, I see it around the time I was trying to update nextcloud: |
is there anything else that maybe indicate what the crash cause was? |
hmm... nothing I can think of. I updated Nextcloud before, but only in the range of version 13.
|
Maybe check for "mariadb" as well? |
|
I got this: Don't know if it's helpful... |
That one looks weird -.- https://discuss.elastic.co/t/es-node-killed-with-6-abrt-how-to-find-out-why-and-by-whom/24108/3 Could you maybe check your system log? syslog, journalctl or dmesg at this time maybe? |
|
This is Maybe I should consider a fresh installation of 14.03... |
|
occ upgrade triggered and 3 seconds later mariadb crashes :/ |
|
Running the latest stable version of Nextcloud and this error started popping up out of blue. Particularly when I click on the gallery button or on a file name that is a picture. |
|
Here my is MySQL 5.7 config: Keep receiving this error when navigating folders with pictures. Very small folders. Only handful files and each file is about ~1 mb. Any thoughts? |
|
From what I see this is the main thread for this "Server has gone away" issue caused by MySQL SSL connection. To me, this is obviously a bug in the nextcloud code and here is why I think so. I have a few months old owncloud instance next to the nextcloud. Configuration in config/config.php is exactly the same (except obvious things like dbname). At the time writing it stores over 100GB of files in total for 3 users. Owncloud doesn't report any "server has gone away" nor "SSL_shutdown:shutdown while in init" errors. Same Ubuntu 18.04, PHP, nginx, OpenSSL, MariaDB server, MariaDB certificates, config/config.php config. But different codebase (from some point as I understand). Hope this will reopen the discussion a bit :) In my case when the issue appears I always get at least 4 errors in the following order. First 3 from APP = PHP: And then failed commands. Usually up to two. One from App = core and second from App = cron: |
|
Hello. I tried to upgrade on WebUI and has error when upload file on Web: An exception occurred while executing 'SELECT * FROM Help me solve it! |
|
I also get this error since weeks. I'm using NC 16.0.6.1. |
|
change the MySQL config, resolved. |
|
As I cannot reproduce the original issue anymore, I will close this ticket. If this is still happening please make sure to upgrade to the latest version. After that, feel free to reopen. |
|
@skjnldsv I found this issue updating from 19.0.8 to 20.0.6, as suggested by Nextcloud itself. The server is a KVM virtual machine with 4 GB. It runs PHP 7.4 as FPM, also updated to the latest version available at Debian Stretch, with memory_limit=512M. After upgrading to Debian Buster, I was able to finish the Nextcloud upgrade. |
Steps to reproduce
Expected behaviour
Updater should complete without error.
Actual behaviour
Update is not successful:
NextCloudPi diagnostics
Nextcloud configuration
HTTPd logs
Database logs
Nextcloud logs
The text was updated successfully, but these errors were encountered: