**** BEGIN LOGGING AT Fri Nov 14 03:00:00 2014 Nov 14 11:45:06 build #668 of au1000 is complete: Failure [failed compile_8] Build details are at http://buildbot.openwrt.org:8010/builders/au1000/builds/668 Nov 14 15:23:42 just to confirm, on BB, I should not be using _any_ luci-reload/ucitrack files should I? I should _only_ be using procd's support for watching uci config? Nov 14 15:24:06 karlp: there is still no complete ucitrack replacement Nov 14 15:24:13 luci does not use config-reload either, yet Nov 14 15:24:55 ok, I'm seeing sort of partial stuff, where somethings reload some don't, possibly my problems still too, moving things to BB, but it definitely isn't just "working like it did in AA, but there's a better way available" Nov 14 15:25:48 for luci nothing changed actually, it still uses ucitrack which in turn triggers /etc/init.d/foo reload Nov 14 15:26:06 however procd now filters reloads - if it thinks the runtime config did not change it will not relaod the service Nov 14 15:26:41 I also had issues with config file tracking in procd (file based reload triggers) - not sure if this works properly by now Nov 14 15:28:01 hrmm. will have to go over them all and recheck the whole lot Nov 14 15:28:24 by "no complete ucitrack" you mean the depends stuff right? Nov 14 15:28:34 simple 1 uci file == 1 process works ok? Nov 14 15:28:57 I didn't understand your last question Nov 14 15:30:05 you said that there's no complete ucitrack replacement, Nov 14 15:30:20 is that only for the depends features in ucitrack, that could reload other services? Nov 14 15:30:36 well in theory there is Nov 14 15:30:57 if you change an arbritary uci file and call "config-reload" it should restart the services affected by the change Nov 14 15:31:19 if it does not then either the init script is not converted to procd or there's a bug in procd Nov 14 15:31:34 so the pile of entries in /e/c/ucitrack are mostly because of pakcages that aren't converted to procd then? Nov 14 15:31:59 ucitrack is maintained by luci and luci has not been converted to procd's reload facilities yet Nov 14 15:32:44 because the last time I took a look (sometime during summer iirc) it didn't behave deterministically for me Nov 14 15:32:53 it = the procd reload stuff Nov 14 15:33:07 so I opted for the "keep things as they're working now" way Nov 14 15:33:14 fair enough :) Nov 14 15:33:21 to be re-evaluated $later Nov 14 15:33:52 I've got a few luci controllers that currently call "luci-reload blah" and it feels slike some of that isn' towkring. Nov 14 15:34:06 but it might also be that some of the apps aren't behaving well on BB, need to look at it further Nov 14 15:34:16 some of them should be simply to just port the init script to procd though Nov 14 15:34:24 you should see if /etc/init.d/service reload behaves as intended Nov 14 15:34:35 because thats what its boilding down to in the end Nov 14 15:34:40 -d Nov 14 15:35:13 e.g. if the assembled commandline did not change then prcd things there is no need for further action and will not touch the running instance Nov 14 15:35:14 right, but ucitrack is currently handling depends for those Nov 14 15:35:26 ok, that's going to suck for procd then, Nov 14 15:35:41 I've got apps that use uci config files themselves, rather than having init scripts turn uci config into arguments Nov 14 15:35:46 this can be problematic if the calculated commandline is indeed the same but some auxilary file did change (e.g. a lease file) Nov 14 15:35:59 so the command line is just "mlifter" but mlifter uses libuci to read it it's config Nov 14 15:36:02 for this there is the facility to register file trigger Nov 14 15:36:41 http://wiki.openwrt.org/inbox/procd-init-scripts#procd.triggers.on.config.file.network.interface.changes Nov 14 15:38:04 so maybe it is enough for your mlifter init to include "procd_add_reload_trigger mlifter" (assuming it reads /etc/config/mlifter) Nov 14 15:38:15 yeah, that sounds like it will be sufficient. Nov 14 15:38:31 if the md5sum of /etc/config/mlifter changed since the service spawning it will get restarted on reload requests Nov 14 15:38:52 oh, its already dokumented in the wiki Nov 14 15:38:58 I'm actually surprised that it is Nov 14 15:39:17 heh, I added half of that page, Nov 14 15:39:24 probably last time I was trying to understand this :) Nov 14 15:40:14 so how's that service trigger different from "procd_set_param file /etc/config/blah" ? Nov 14 15:40:44 iirc the file param is for arbritary files Nov 14 15:40:49 non-uci ones Nov 14 15:40:55 does it matter though? Nov 14 15:41:50 oh, it's for "reload_config" vs auto restarting when the file changes Nov 14 15:45:29 there's more to it Nov 14 15:45:47 procd_add_reload_trigger will also react on "config.change" events broadcasted via ubus Nov 14 15:46:25 while "procd_set_param file" is simply a compare criterion to compare runtime state with new state (similar to how calculated cmdlines are compared) Nov 14 15:46:37 config.change events are for example triggered by rpcd Nov 14 15:48:21 the /sbin/reload_config file is more of a helper program to manually trigger broadcasting of config.change events Nov 14 15:48:39 similar to how "udevtrigger" synthesizes udev events which are otherwise emitted by the kernel Nov 14 15:49:23 why is that page under "few adjustments (shiva)" ? Nov 14 15:49:24 in the future, with luci2 or any other hypothetical frontend, all uci manipulation will happen via the ubus-uci api Nov 14 15:49:50 and this one will trigger the events automatically Nov 14 15:50:07 one could think about adding the event broadcasting to the uci cli too Nov 14 15:50:24 e.g. "uci commit" will connect to ubus and trigger config.change Nov 14 15:51:24 karlp: because someone renamed "Inbox" to that Nov 14 15:51:43 bleh, editor errors again, *tries another browser* Nov 14 15:52:00 I should certainly hope that uci commit would do the right thing :) Nov 14 15:52:08 also simply editing the file Nov 14 15:52:15 that requires inotify though Nov 14 15:52:31 and the problem with auto-reloading on edits is to decide when to trigger the reload Nov 14 15:52:57 only for apps that have registered for it? Nov 14 15:53:40 I certainly do not want ssh to reload immediately after ":w" in vi after I just entered an invalid ip Nov 14 15:53:56 or disabled password auth and not yet uploaded keys Nov 14 16:19:03 build #793 of orion is complete: Failure [failed compile_8] Build details are at http://buildbot.openwrt.org:8010/builders/orion/builds/793 Nov 14 16:21:30 anyone seen this sort of thing? http://pastebay.net/1522474 Nov 14 16:21:38 "jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x00040018: 0xec66 instead" Nov 14 16:22:04 started seeing this today, doing sysupgrade from a file in /tmp Nov 14 16:23:04 wrong image for sysupgrade? Nov 14 16:23:32 happens if you sysupgrade with a factory one iirc Nov 14 16:26:40 nah, same method and files as always Nov 14 16:27:03 will hope it goes away for now. Nov 14 16:28:49 what's dropbearkey doing? is that new? I don't remember seeing this run at such high cpu for so long before Nov 14 16:29:15 it generates the host keys Nov 14 16:29:19 if there are none yet Nov 14 16:33:31 I'm on a fresh insatll, so telnet is still active, and I've not run passwd yet. Nov 14 16:33:41 it's chewing up huge cpu and just spinning on it Nov 14 16:34:14 ok, finally finished now, Nov 14 16:34:20 weird, never noticed that before :| Nov 14 16:41:23 matteo r43241 trunk/package/network/services/ppp/ files/ppp.sh patches/520-uniq.patch * pppd: add option to set custom host-uniq pppoe tag Nov 14 16:53:49 blogic r43242 trunk/target/linux/ramips/patches-3.14/0032-PCI-MIPS-adds-mt7620a-pcie-driver.patch * ralink: drop some unused code from pci-mt7620a Nov 14 16:53:53 blogic r43243 trunk/target/linux/ramips/files/drivers/net/ethernet/ralink/soc_mt7620.c * ralink: mt7621 has a different base addr for PVID Nov 14 16:53:59 blogic r43244 trunk/target/linux/ramips/ (7 files in 6 dirs) * ralink: add support for WSR-600/1166 Nov 14 16:54:05 blogic r43245 trunk/target/linux/ramips/files/drivers/net/ethernet/ralink/gsw_mt7620a.c * ralink: improve mt7530 support Nov 14 16:54:10 blogic r43246 trunk/target/linux/ramips/files/drivers/net/ethernet/ralink/gsw_mt7620a.c * ralink: fix the 10mbit bug on mt7621 Nov 14 16:54:16 blogic r43247 trunk/target/linux/ramips/files/drivers/net/ethernet/ ralink/soc_rt305x.c ralink/esw_rt3052.c * ralink: update ethernet driver to use new ralink_soc variable Nov 14 16:54:26 blogic r43248 trunk/target/linux/ramips/patches-3.14/0060-soc_type.patch * ralink: add a new global ralink_soc variable Nov 14 16:54:32 blogic r43249 trunk/target/linux/ramips/ dts/mt7620a.dtsi patches-3.14/0032-PCI-MIPS-adds-mt7620a-pcie-driver.patch dts/mt7621.dtsi dts/mt7628an.dtsi * ralink: update pcie driver to load ranges from dts Nov 14 16:54:38 blogic r43250 trunk/target/linux/ramips/patches-3.14/0053-mmc-MIPS-ralink-add-sdhci-for-mt7620a-SoC.patch * ralink: update sdhci driver Nov 14 16:54:44 blogic r43251 trunk/target/linux/ramips/patches-3.14/0012-MIPS-ralink-add-MT7621-support.patch * ralink: fix broken pinmux function naming inside mt7621.c Nov 14 16:54:52 blogic r43252 trunk/target/linux/ramips/mt7621/config-3.14 * ralink: enable the proper drivers in the mt7621 config Nov 14 16:54:57 blogic r43253 trunk/target/linux/ramips/patches-3.14/0048-GPIO-ralink-add-mt7621-gpio-controller.patch * ralink: fix bank detection in mt7621 gpio driver Nov 14 16:55:05 blogic r43254 trunk/target/linux/ramips/ patches-3.14/999-no-pm_poweroff.patch mt7621/config-3.14 * ralink: dont define pm_poweroff Nov 14 16:55:11 blogic r43255 trunk/target/linux/lantiq/base-files/etc/uci-defaults/01_leds Nov 14 16:55:11 lantiq: set default led configuration from /sys entry later, after per board settings Nov 14 16:55:16 blogic r43256 trunk/target/linux/lantiq/base-files/etc/uci-defaults/01_leds * lantiq: vgv7519 - set some defaults for leds Nov 14 16:55:22 blogic r43257 trunk/package/system/procd/Makefile * procd: update to latest git HEAD Nov 14 16:55:31 blogic r43258 trunk/target/ (5 files in 4 dirs) * ar71xx: add kernel support for the OpenMesh MR900 board Nov 14 16:55:40 blogic r43259 trunk/target/ linux/ar71xx/base-files/etc/diag.sh linux/ar71xx/base-files/etc/uci-defaults/01_leds linux/ar71xx/base-files/lib/ar71xx.sh linux/ar71xx/base-files/etc/uci-defaults/02_network * ar71xx: add user-space support for the OpenMesh MR900 board Nov 14 16:55:46 blogic r43260 trunk/scripts/om-fwupgradecfg-gen.sh * scripts/om-fwupgradecfg-gen.sh: add support for the MR900 Nov 14 16:55:51 blogic r43261 trunk/target/ linux/ar71xx/base-files/lib/upgrade/openmesh.sh linux/ar71xx/base-files/lib/upgrade/platform.sh * ar71xx: enable sysupgrade for the OpenMesh MR900 Nov 14 16:56:02 blogic r43262 trunk/package/kernel/om-watchdog/files/om-watchdog.init * package/om-watchdog: add OpenMesh MR900 support Nov 14 16:56:08 blogic r43263 trunk/package/boot/uboot-envtools/files/ar71xx * package/uboot-envtools: add OpenMesh MR900 support Nov 14 16:56:14 blogic r43264 trunk/target/ linux/ar71xx/generic/profiles/openmesh.mk linux/ar71xx/image/Makefile * ar71xx: create profile and build image for the OpenMesh MR900 board Nov 14 16:56:21 blogic r43265 trunk/target/ linux/ar71xx/patches-3.10/731-MIPS-ath79-add-mr900v2-support.patch linux/ar71xx/files/arch/mips/ath79/mach-mr900.c linux/ar71xx/patches-3.14/731-MIPS-ath79-add-mr900v2-support.patch * ar71xx: add kernel support for the OpenMesh MR900v2 Nov 14 16:56:27 blogic r43266 trunk/target/ linux/ar71xx/base-files/etc/diag.sh linux/ar71xx/base-files/etc/uci-defaults/01_leds linux/ar71xx/base-files/lib/ar71xx.sh linux/ar71xx/base-files/etc/uci-defaults/02_network * ar71xx: add user-space support for OpenMesh MR900v2 Nov 14 16:56:37 blogic r43267 trunk/target/ linux/ar71xx/base-files/lib/upgrade/openmesh.sh linux/ar71xx/base-files/lib/upgrade/platform.sh * ar71xx: add sysupgrade support for OpenMesh MR900v2 Nov 14 16:56:43 blogic r43268 trunk/package/kernel/om-watchdog/files/om-watchdog.init * package/om-watchdog: add OpenMesh MR900v2 support Nov 14 16:56:50 blogic r43269 trunk/package/boot/uboot-envtools/files/ar71xx * package/uboot-envtools: add OpenMesh MR900v2 support Nov 14 16:56:57 blogic r43270 trunk/target/linux/ar71xx/generic/profiles/openmesh.mk * ar71xx: adjust MR900 profile for the OpenMesh MR900v2 board Nov 14 16:59:52 blogic r43271 trunk/target/ linux/ar71xx/files/arch/mips/ath79/mach-el-mini.c linux/ar71xx/base-files/etc/diag.sh linux/ar71xx/files/arch/mips/ath79/mach-el-m150.c * ar71xx: Fix the sys LEDs on the EasyLink M150 and mini Nov 14 17:07:27 blogic r43272 trunk/target/linux/lantiq/patches-3.14/0160-owrt-lantiq-multiple-flash.patch * lantiq: Concatenate multiple flash chips for lantiq-flash Nov 14 17:07:32 blogic r43273 trunk/target/linux/generic/files/drivers/net/phy/ar8216.c * ar8216: fix "unused variable" compiler warning Nov 14 18:29:53 jow_laptop: this is what I mean with luci-reload not doign anything: http://pastebay.net/1522570 Nov 14 18:30:23 /etc/init.d/mlifter reload does the right thing, luci-reload mlifter just returns immediately and has no impact and nothing visible in the logs or console Nov 14 18:53:04 blogic r43274 trunk/target/linux/ramips/ patches-3.14/0060-soc_type.patch modules.mk * ralink: make the sdhci module build again Nov 15 02:19:21 build #722 of xburst is complete: Failure [failed compile_8] Build details are at http://buildbot.openwrt.org:8010/builders/xburst/builds/722 **** ENDING LOGGING AT Sat Nov 15 03:00:00 2014