I'm impacted by this issue as well and looking for alternative to InfluxDB.
VictoriaMetrics seems to be a suitable candidate, but it's written in go as well. Does this issue with syscalls impacts all applications written in golang? They all do some home made syscalls stuff bypassing libc. Cheers, Chris On Sun, 2023-04-16 at 19:11 +0200, Landry Breuil wrote: > Hi, > > spent some days fixin' influxdb which currently badly crashes at > startup > on 7.3 with SIGSYS, due to go and syscalls *sigh*. > > diffs attached: > - influxdb-1.8.10-stable.diff is for 7.3 and 'only' updates x/sys to > 0.5.0, i've it running fine on my 7.3/amd64 vm receiving metrics > from all the > telegraf instances running on various boxes. That's the 'minimal' > fix. > > - influxdb-1.11.0.diff updates it to the latest version of the 1.11.0 > branch, backporting the x/sys update to 0.5.0 extracted from > https://github.com/influxdata/influxdb/commit/906c090a48c807bc1799f62c849ce30793566e76 > - note that this port is now using go *and* rust, so i've kept the > build system > in control of go, just letting it run the cargo build commands to > build the > 'libflux' bits from > https://github.com/influxdata/flux/tree/master/libflux - > feedback welcome on the porting bits, some might be horrible (note > the > overloading of a 'pkg-config.sh' script..among others) > > - and while here i've ported the latest version of the 2 branch, > 2.7.0, also > using go and rust. v2 is a bit different, doesnt ship a default > config file > nor the 'influx' cli, requires authentification, doesnt have the udp > listener > that collectd/prometheus/graphite could use to directly send metrics > to > influxdb, those have to go through telegraf now (!?). see > https://docs.influxdata.com/influxdb/v2.7/upgrade/v1-to-v2/automatic-upgrade/#supported-protocols > for details, but i'd be wary of force-updating everyone using > influxdb v1 to > v2. Maybe it'd make sense to have both ports coexist ? feedback from > actual > influxdb users welcome. > > to be able to set EXTRACT_ONLY in the last two ports, i have this > local patch: > > Index: go.port.mk > =================================================================== > RCS file: /cvs/ports/lang/go/go.port.mk,v > retrieving revision 1.55 > diff -u -r1.55 go.port.mk > --- go.port.mk 28 Feb 2023 15:44:54 -0000 1.55 > +++ go.port.mk 16 Apr 2023 16:58:39 -0000 > @@ -76,7 +76,7 @@ > ALL_TARGET ?= ${MODGO_MODNAME} > MODGO_FLAGS += -modcacherw -trimpath > DISTFILES += > ${DISTNAME_ESC}{${MODGO_VERSION}}${EXTRACT_SUFX} > -EXTRACT_ONLY = ${DISTNAME_ESC}${EXTRACT_SUFX} > +EXTRACT_ONLY ?= ${DISTNAME_ESC}${EXTRACT_SUFX} > MASTER_SITES ?= > ${MASTER_SITE_ATHENS}${MODGO_MODNAME_ESC}/@v/ > . for _modpath _modver in ${MODGO_MODULES} > DISTFILES += > ${MODGO_DIST_SUBDIR}/{}${_modpath}/@v/${_modver}.zip:${MODGO_MASTER_S > ITESN} > > this will run through a bulk. jsing@ is okay with it, still being > discussed > wether it should be += or ?=. > > oks and feedback on all this welcome. Phew. -- +44 7502 415 180 (Phone, Signal, WhatsApp) @ezaquarii:etacassiopeiae.net (Matrix)
