Turris OS packages issueshttps://gitlab.nic.cz/turris/os/packages/-/issues2023-07-18T12:54:28+02:00https://gitlab.nic.cz/turris/os/packages/-/issues/790resolver-conf problem with both hotplug scripts and the wan6 interface2023-07-18T12:54:28+02:00Christoph Metzresolver-conf problem with both hotplug scripts and the wan6 interfaceI have two issues with the hotplug scripts restarting the resolver on every interface update. Currently there seems to be a bug / issue in OpenWrt in resulting in frequent interface updates on my wan6 interface, also there seems to be no...I have two issues with the hotplug scripts restarting the resolver on every interface update. Currently there seems to be a bug / issue in OpenWrt in resulting in frequent interface updates on my wan6 interface, also there seems to be no change in the ip addresses at all, even the temporary addresses are still the same, but i did not investigate this any further. I just added a simple nested if to it. Some similar issues where also mentioned in the forum https://forum.turris.cz/t/every-3-secs-in-log-kresd-hard-limit-for-number-of-file-descriptors-any-help:
**/etc/hotplug.d/iface/40-ip-resolver-reload**
```
if [ "$ACTION" = "ifupdate" -o "$ACTION" = "ifup" ]; then
if [ "$IFUPDATE_ADDRESSES" = "1" -o "$IFUPDATE_PREFIXES" = "1" ]; then
if [ "$INTERFACE" != "wan6" ]; then <------- added line
logger -t hotplug "Reload resolver because of interface address update"
/etc/init.d/resolver reload
fi
fi
fi
```
I also recognized that the second script /etc/hotplug.d/iface/40-resolver-reload is also triggerd, there is some matching against an previous md5, but which never got set on my box. So i added storing the md5 in the script as well, i'm not sure if some other script should set this:
**/etc/hotplug.d/iface/40-resolver-reload**
```
if /etc/init.d/resolver enabled && \
[ "$MD5" != "$PREVIOUS" ] && \
[ "$DO_FORWARD" = "1" ] ; then
/etc/init.d/resolver reload
echo "$MD5" > /tmp/resolv.conf.auto.last.md5 <-------- added line
fi
```
maybe this whole stuff could be refactored in a single script, i have no clue why there is a need for two different hotplug scripts at all.
**Update**:
i did some further investigations and it seems the "netifd" is doing a prefix update event also when only the preferred_until / valid_until values are modified.
**https://git.openwrt.org/?p=project/netifd.git;a=blob;f=interface-ip.c**
```
if (node_old && node_new) {
/* Move assignments and refresh addresses to update valid times */
list_splice(&prefix_old->assignments, &prefix_new->assignments);
list_for_each_entry(c, &prefix_new->assignments, head)
if ((iface = vlist_find(&interfaces, c->name, iface, node)))
interface_set_prefix_address(c, prefix_new, iface, true);
if (prefix_new->preferred_until != prefix_old->preferred_until ||
prefix_new->valid_until != prefix_old->valid_until)
ip->iface->updated |= IUF_PREFIX; <--- ALSO RESULTING IN AN PREFIX_UPDATE EVENT
```Turris OS 5.3.0https://gitlab.nic.cz/turris/os/packages/-/issues/789add turris-timezone as foris-controller dependency2023-08-16T14:54:09+02:00Štěpán Henekadd turris-timezone as foris-controller dependencyOtherwise this breaks the builds once newer version of foris controller is issued > "2.1.0"Otherwise this breaks the builds once newer version of foris controller is issued > "2.1.0"Štěpán HenekŠtěpán Henekhttps://gitlab.nic.cz/turris/os/packages/-/issues/788turris.local/netdata (without trailing slash) doesn't load JS and CSS from th...2023-08-16T14:56:50+02:00Michal Vasilekturris.local/netdata (without trailing slash) doesn't load JS and CSS from the proper locationsGoing to turris.local/netdata (without the trailing slash) tries to load resources from the root - /style.css, instead of /netdata/style.css. This is not a big problem, because all links use /netdata/ (with the trailing slash).
The bug ...Going to turris.local/netdata (without the trailing slash) tries to load resources from the root - /style.css, instead of /netdata/style.css. This is not a big problem, because all links use /netdata/ (with the trailing slash).
The bug was reported in netdata - https://github.com/netdata/netdata/issues/11415, but we could fix it in the lighttpd config in the meantime.https://gitlab.nic.cz/turris/os/packages/-/issues/787tos3to4 : migration don't finish (omnia)2023-08-16T14:35:44+02:00Ghost Usertos3to4 : migration don't finish (omnia)turris_version : 3.11.23
[updater-log.tar.xz](/uploads/a136c5c9326e2c11b4839a30d3ac40cd/updater-log.tar.xz)
I tried to do the migration, but here the output of notifications:
```
Error from 2021/08/04 12:59:40
Updater approvals were dea...turris_version : 3.11.23
[updater-log.tar.xz](/uploads/a136c5c9326e2c11b4839a30d3ac40cd/updater-log.tar.xz)
I tried to do the migration, but here the output of notifications:
```
Error from 2021/08/04 12:59:40
Updater approvals were deactivated to prevent problems during migration to the latest major release of Turris OS!
×
Update from 2021/08/04 12:59:58
• Installed version 1.0.3-1 of package tos3to4
×
Error from 2021/08/04 13:02:06
Updater selhal:
[string "transaction"]:327: [string "transaction"]:151: Collisions:
• /etc/crontabs/.keep: cronie (new-file), vixie-cron (existing-file)
• /usr/bin/crontab: cronie (new-file), vixie-cron (existing-file)
• /usr/bin/updater-supervisor: updater-supervisor (new-file), updater-ng-supervisor (existing-file)
• /etc/updater/hook_postupdate/99_approvals_cleanup: updater-supervisor (new-file), updater-ng-supervisor (existing-file)
• /etc/cron.d/.keep: cronie (new-file), vixie-cron (existing-file)
• /etc/cron.d/updater: updater-supervisor (new-file), updater-ng-supervisor (existing-file)
• /usr/share/updater/l10n_supported: l10n_supported (existing-file), l10n-supported (new-file)
• /etc/init.d/cron: cronie (new-file), vixie-cron (existing-file)
×
Update from 2021/08/04 13:02:06
• Installed version 7.5.0-2 of package libgcc
• Installed version 1.1.24-2 of package libc
• Installed version 0.12.1-3.1 of package libjson-c
• Installed version 2020-05-25-66195aee-1 of package libubox
• Installed version 2018-02-04-c7e938d6-1 of package jsonfilter
• Installed version 1.30.1-8 of package busybox
• Installed version 20200601-1 of package ca-certificates
• Installed version 1.1.1k-1 of package libopenssl
• Installed version 2.2.9-1 of package libexpat
• Installed version 1.0.8-1 of package libbz2
• Installed version 1.1.24-2 of package libpthread
• Installed version 5.2.5-1 of package liblzma
• Installed version 1.2.11-3 of package zlib
• Installed version 3.4.2-1 of package libarchive
• Installed version 0.9.4-2 of package liburiparser
[updater-log.txt](/uploads/2fd885d07a31a6034c05bd83e1791c1f/updater-log.txt)• Installed version 20200601-1 of package ca-bundle
• Installed version 7.66.0-3 of package libcurl
• Installed version 2.1.11-1 of package libevent2
• Installed version 5.1.5-3 of package liblua
• Installed version 2019-09-01-415f9e48-4 of package libuci
• Installed version 2019-09-01-415f9e48-4 of package libuci-lua
• Installed version 69.1.3-1 of package updater-ng
• Installed version 5.1.5-3 of package lua
• Installed version 3.0-rc1-20130909-5 of package luasocket
• Installed version 0.9.54-1 of package json4lua
• Installed version 1.6.3-1 of package pkglists
• Installed version 7.66.0-3 of package curl
• Installed version 2020-06-17-51e16ebf-1 of package libuclient
• Installed version 2020-06-17-51e16ebf-1 of package uclient-fetch
• Installed version 2021-01-31-c5dccea9-1 of package opkg
• Installed version 8.30-2 of package coreutils
• Installed version 8.30-2 of package coreutils-sort
• Installed version 2019-09-01-415f9e48-4 of package uci
• Installed version 2.6.0-3 of package switch-branch
• Installed version 2.2.1-3 of package tos3to4-early
• Installed version 1-1 of package fix-pkglists-options
• Installed version 1-1 of package fix-pkglists-hardening-options
×
Error from 2021/08/04 13:52:59
Updater selhal:
[string "transaction"]:327: [string "transaction"]:151: Collisions:
• /etc/crontabs/.keep: cronie (new-file), vixie-cron (existing-file)
• /usr/bin/crontab: cronie (new-file), vixie-cron (existing-file)
• /usr/bin/updater-supervisor: updater-supervisor (new-file), updater-ng-supervisor (existing-file)
• /etc/updater/hook_postupdate/99_approvals_cleanup: updater-supervisor (new-file), updater-ng-supervisor (existing-file)
• /etc/cron.d/.keep: cronie (new-file), vixie-cron (existing-file)
• /etc/cron.d/updater: updater-supervisor (new-file), updater-ng-supervisor (existing-file)
• /usr/share/updater/l10n_supported: l10n_supported (existing-file), l10n-supported (new-file)
• /etc/init.d/cron: cronie (new-file), vixie-cron (existing-file)
```
Since, I can't acces to the router with ssh.
Seems that I'm now in the middle of the migration, but it don't finish.Turris OS 5.2.5 (formal release)https://gitlab.nic.cz/turris/os/packages/-/issues/786Update Knot Resolver to version 5.4.02023-08-16T14:54:11+02:00Josef SchlehoferUpdate Knot Resolver to version 5.4.0Recently, there was released a new version 5.4.0 of Knot Resolver.
Changelog: https://www.knot-resolver.cz/2021-07-29-knot-resolver-5.4.0.html
Would it be possible to take a look? This should be send to `develop` branch of this reposito...Recently, there was released a new version 5.4.0 of Knot Resolver.
Changelog: https://www.knot-resolver.cz/2021-07-29-knot-resolver-5.4.0.html
Would it be possible to take a look? This should be send to `develop` branch of this repository, so we can include it in minor release.
**Knot Resolver** runs on `Turris Omnia` and `Turris MOX` (Shield). Once you do run time testing, then it needs to be send to the upstream repository. In this case, I'm talking about `packages` feed.Turris OS 5.3.0https://gitlab.nic.cz/turris/os/packages/-/issues/785[MOVED] Wireless not automatically re-connecting when connection was lost (up...2021-08-06T01:59:27+02:00Marek Ľach[MOVED] Wireless not automatically re-connecting when connection was lost (update watchcat, mwan3) config by default in reForisThe WiFi sometimes disconnects in mid-day, and TurrisOS does not seem to be able to reboot/restart such a once lost connection immediately on its own, without the need for a lenghty manual intervention.
The `OpenWRT` packages `mwan3` an...The WiFi sometimes disconnects in mid-day, and TurrisOS does not seem to be able to reboot/restart such a once lost connection immediately on its own, without the need for a lenghty manual intervention.
The `OpenWRT` packages `mwan3` and also `watchcat` seem to be helpful in these situations, so **TurrisOS** could have them included, and active by default in its own _reForis_ interface, to make it even more convenient, useful and user-friendly.
~~Also, `ppp` should be set to `keepalive` out-of-the gate.~~
It’d be worth it implementing these plugins into the latest iteration of TurrisOS directly... for laymen like myself :-)Marek ĽachMarek Ľachhttps://gitlab.nic.cz/turris/os/packages/-/issues/784turris-maintain: remove option to set message bus2021-08-12T22:35:50+02:00Martin Matějekturris-maintain: remove option to set message busFollowup of commit c5428642
Set mqtt as the only message bus, because we no longer use another message
buses on regular router setup.Followup of commit c5428642
Set mqtt as the only message bus, because we no longer use another message
buses on regular router setup.https://gitlab.nic.cz/turris/os/packages/-/issues/783turris-maintain: fix reading of IP addresses2021-08-12T22:36:19+02:00Martin Matějekturris-maintain: fix reading of IP addressesIssue turris/foris-controller/foris-controller#146 introduced new way of storing IP address and
netmask for lan, to be more aligned with the way LuCI stores it.
However it breaks maintain-network-restart script, which in turn won't
sen...Issue turris/foris-controller/foris-controller#146 introduced new way of storing IP address and
netmask for lan, to be more aligned with the way LuCI stores it.
However it breaks maintain-network-restart script, which in turn won't
send notification to re/foris to show the "reloading network..."
spinner.https://gitlab.nic.cz/turris/os/packages/-/issues/782Could not upgrade from TOS 5.0.42021-08-03T09:13:05+02:00Lukas JelinekCould not upgrade from TOS 5.0.4MOX with TOS 5.0.4 could not upgrade itself to newer versions of Turris OS. It displayed in Foris:
```
Updater failed: Called uri_path on URI of scheme: https
```
The next attempt succeeded and the MOX has been upgraded to TOS 5.2.3.MOX with TOS 5.0.4 could not upgrade itself to newer versions of Turris OS. It displayed in Foris:
```
Updater failed: Called uri_path on URI of scheme: https
```
The next attempt succeeded and the MOX has been upgraded to TOS 5.2.3.https://gitlab.nic.cz/turris/os/packages/-/issues/781SyntaxWarning in dhcp_host_domain while booting HBD2021-09-27T13:58:19+02:00Josef SchlehoferSyntaxWarning in dhcp_host_domain while booting HBDI plugged UART cable to the Turris Omnia router which has HBD branch and I noticed this warning while it was booting:
```
[ 9.785704] device lan4 entered promiscuous mode
[ 9.826746] mv88e6085 f1072004.mdio-mii:10: p5: already a me...I plugged UART cable to the Turris Omnia router which has HBD branch and I noticed this warning while it was booting:
```
[ 9.785704] device lan4 entered promiscuous mode
[ 9.826746] mv88e6085 f1072004.mdio-mii:10: p5: already a member of VLAN 1
[ 9.854006] mvneta f1034000.ethernet eth2: PHY [f1072004.mdio-mii:01] driver [Marvell 88E1510] (irq=POLL)
[ 9.863860] mvneta f1034000.ethernet eth2: configuring for phy/sgmii link mode
[ 9.930339] mv88e6085 f1072004.mdio-mii:10 lan3: Link is Up - 1Gbps/Full - flow control rx/tx
[ 9.940642] br-lan: port 4(lan3) entered blocking state
[ 9.945917] br-lan: port 4(lan3) entered forwarding state
[ 9.967543] IPv6: ADDRCONF(NETDEV_CHANGE): br-lan: link becomes ready
[ 15.042215] mvneta f1034000.ethernet eth2: Link is Up - 1Gbps/Full - flow control rx/tx
[ 15.050261] IPv6: ADDRCONF(NETDEV_CHANGE): eth2: link becomes ready
/etc/resolver/dhcp_host_domain_ng.py:50: SyntaxWarning: "is" with a literal. Did you mean "=="?
if output is "syslog":
```Turris OS 6.0https://gitlab.nic.cz/turris/os/packages/-/issues/780Restarting kresd or resolver leads to no such file or directory for resolv.co...2022-07-02T08:55:14+02:00Josef SchlehoferRestarting kresd or resolver leads to no such file or directory for resolv.conf.vpnThis is happening on Turris OS 6.0.
```
root@omnia:~# /etc/init.d/kresd stop
root@omnia:~# /etc/init.d/kresd start
ls: /etc/resolv.conf.vpn.*: No such file or directory
job 4 at Sun Jul 18 09:12:00 2021
root@omnia:~# /etc/init.d/kresd s...This is happening on Turris OS 6.0.
```
root@omnia:~# /etc/init.d/kresd stop
root@omnia:~# /etc/init.d/kresd start
ls: /etc/resolv.conf.vpn.*: No such file or directory
job 4 at Sun Jul 18 09:12:00 2021
root@omnia:~# /etc/init.d/kresd stop
root@omnia:~# /etc/init.d/kresd start
ls: /etc/resolv.conf.vpn.*: No such file or directory
job 5 at Sun Jul 18 09:12:00 2021
root@omnia:~# /etc/init.d/kresd stop
root@omnia:~# /etc/init.d/kresd start
ls: /etc/resolv.conf.vpn.*: No such file or directory
job 6 at Sun Jul 18 09:12:00 2021
```Turris OS 5.3.11https://gitlab.nic.cz/turris/os/packages/-/issues/779Freshly flashed system is sometimes corrupted2022-07-29T13:34:20+02:00Lukas JelinekFreshly flashed system is sometimes corruptedI've encountered some cases in which a freshly flashed Turris OS 5.2.x was corrupted. These cases occured with both flash ways (from USB drives and from the repository server) and on Omnias, MOXes and Shields. The corrupted systems had m...I've encountered some cases in which a freshly flashed Turris OS 5.2.x was corrupted. These cases occured with both flash ways (from USB drives and from the repository server) and on Omnias, MOXes and Shields. The corrupted systems had malformed initial setup guides (didn't started correctly and/or couldn't proceed) or even didn't start at all.
May be related to #778 and https://gitlab.nic.cz/turris/reforis/reforis/-/issues/340.
![mox-tos-5-2-3](/uploads/aa1558c7219cf984047c344b907c8fa9/mox-tos-5-2-3.mp4)Turris OS 6.0https://gitlab.nic.cz/turris/os/packages/-/issues/777Occasionally unable to log in to LuCI with correct username and password (sav...2021-07-24T11:35:40+02:00Jakub KákonaOccasionally unable to log in to LuCI with correct username and password (saved in browser)Hello, I have encountered a bug that prohibits a log-in to the LuCI web interface with a browser-saved login name and password.
The browser window is simply reloaded to the login page again:
![image](/uploads/fe8710aaab7704571b71ed77d...Hello, I have encountered a bug that prohibits a log-in to the LuCI web interface with a browser-saved login name and password.
The browser window is simply reloaded to the login page again:
![image](/uploads/fe8710aaab7704571b71ed77dc0c944f/image.png)
LuCI branch (git-21.160.68865-15ca915)
ssh works normally at the same time. Therefore I found this message:
root@omnia:~# cat /var/log/lighttpd/error.log
2021-07-06 19:42:13: ../src/server.c.1513) server started (lighttpd/1.4.59)
2021-07-06 19:56:39: ../src/mod_fastcgi.c.487) FastCGI-stderr:[2021-07-06 19:56:39,507] WARNING in flask_seasurf: Forbidden (CSRF token missing or incorrect.): /login
root@omnia:~#
After that, I found that a "private browsing window" of Firefox browser allows me to login in situation where the bug is active.https://gitlab.nic.cz/turris/os/packages/-/issues/776Package python3-flask requires package python3-click that is not available2021-06-28T14:24:01+02:00Lukas JelinekPackage python3-flask requires package python3-click that is not availableAfter updating Turris Omnia to TOS 5.2.3 (in HBK), it displayed this error message twice:
```
Updater failed:
inconsistent: Package python3-flask requires package python3-click that is not available.
```After updating Turris Omnia to TOS 5.2.3 (in HBK), it displayed this error message twice:
```
Updater failed:
inconsistent: Package python3-flask requires package python3-click that is not available.
```https://gitlab.nic.cz/turris/os/packages/-/issues/775Mariadb memory tuning should respect device total RAM2023-08-16T14:54:12+02:00Vojtech MyslivecMariadb memory tuning should respect device total RAMCurrently, we provide one static `my.cnf` mariadb config for the `mysqld` server.
There are some hardcoded values that should be tuned according to a device total available RAM. i.e. on a 512 MB MOX, there should be lesser limits than ...Currently, we provide one static `my.cnf` mariadb config for the `mysqld` server.
There are some hardcoded values that should be tuned according to a device total available RAM. i.e. on a 512 MB MOX, there should be lesser limits than on a 2 GB Omnia.
Also, `mysqld` parameters should be tunable by a user, preferably through uci config.Turris OS 6.0https://gitlab.nic.cz/turris/os/packages/-/issues/774Nextcloud: fails to build in HBL2021-09-21T12:26:31+02:00Karel KociNextcloud: fails to build in HBL```
Applying ./patches/0001-Replace-gzip-by-identity.patch using plaintext:
patching file lib/private/App/AppStore/Fetcher/Fetcher.php
Hunk #1 FAILED at 98.
1 out of 1 hunk FAILED -- saving rejects to file lib/private/App/AppStore/Fetch...```
Applying ./patches/0001-Replace-gzip-by-identity.patch using plaintext:
patching file lib/private/App/AppStore/Fetcher/Fetcher.php
Hunk #1 FAILED at 98.
1 out of 1 hunk FAILED -- saving rejects to file lib/private/App/AppStore/Fetcher/Fetcher.php.rej
Patch failed! Please fix ./patches/0001-Replace-gzip-by-identity.patch!
make[2]: *** [Makefile:100: /home/beast/beast/workspace/turris-os-packages-lions-turris1x/build/build_dir/target-powerpc_8540_musl/nextcloud/.prepared_8a9b8cce925916fa45a368c2535da49d_6664517399ebbbc92a37c5bb081b5c53] Error 1
time: package/feeds/turrispackages/nextcloud/compile#31.85#3.71#46.14
```Turris OS 5.3.0https://gitlab.nic.cz/turris/os/packages/-/issues/773updater-ng: version with new hash file list name supported2022-03-02T10:07:15+01:00Karel Kociupdater-ng: version with new hash file list name supported* turris/updater/updater#320* turris/updater/updater#320Turris OS 6.02022-01-16https://gitlab.nic.cz/turris/os/packages/-/issues/771updater-ng: version 69.1.32023-08-16T14:35:46+02:00Karel Kociupdater-ng: version 69.1.3With improved error message.
* turris/updater/updater!289With improved error message.
* turris/updater/updater!289Turris OS 5.2.3https://gitlab.nic.cz/turris/os/packages/-/issues/770foris-storage-plugin: fix version 3.0.12023-08-16T14:35:47+02:00Martin Matějekforis-storage-plugin: fix version 3.0.1Fix setup of nextcloud accountFix setup of nextcloud accountTurris OS 5.2.3https://gitlab.nic.cz/turris/os/packages/-/issues/768Minor bug in logging of /etc/resolver/dhcp_host_domain_ng.py2023-08-24T13:57:40+02:00ktetzlaffMinor bug in logging of /etc/resolver/dhcp_host_domain_ng.pyThe `filename. line` arguments in last line of:
```python
def _remove_hints_hosts(self, filename):
with open(filename, "r") as handle:
for line in handle:
line = line.strip()
if no...The `filename. line` arguments in last line of:
```python
def _remove_hints_hosts(self, filename):
with open(filename, "r") as handle:
for line in handle:
line = line.strip()
if not line or line.startswith("#"):
continue
try:
host = line.strip().split()[1]
self._call_kresd("hints.del('%s')" % host)
except:
log("Wrong host format '%s' in host file %s " %
(filename, line), LOG_ERR)
```
need to be switched to `line, filename`:
```python
log("Wrong host format '%s' in host file %s " %
(line, filename), LOG_ERR)
```