Make ldmsd's caching hostname behavior configurable #2386
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| # Test overview: | |
| # 1 sampler (v4.3.3, named samp-4.3.3) | |
| # | | |
| # V | |
| # L1 agg (v4.3.3, named agg-4.3.3) | |
| # | | |
| # V | |
| # L2 agg (OVIS-4, named agg-4) | |
| # | |
| # Scenario | |
| # - samp-4.3.3 started (1 set: meminfo) | |
| # - agg-4.3.3 started, collecting sets from samp-4.3.3 | |
| # - agg-4 started, collecting sets from agg-4.3.3 | |
| # - ldms_ls to agg-4.3.3 & verify that it is not empty | |
| # - ldms_ls to agg-4 & verify that it is the same as agg-4.3.3 | |
| # - kill samp-4.3.3 | |
| # - ldms_ls to agg-4.3.3 & verify that dir result is empty | |
| # - ldms_ls to agg-4 & verify that dir result is empty | |
| name: Compatibility test with OVIS-4.3.3 | |
| on: | |
| push: | |
| branches: | |
| - 'OVIS-4' | |
| - main | |
| - 'b[0-9]+.[0-9]+' | |
| pull_request: | |
| branches: | |
| - 'OVIS-4' | |
| - main | |
| - 'b[0-9]+.[0-9]+' | |
| defaults: | |
| run: | |
| shell: bash | |
| jobs: | |
| test: | |
| runs-on: ubuntu-latest | |
| container: | |
| image: ovishpc/ovis-compat-centos7 | |
| steps: | |
| - uses: actions/checkout@v1 | |
| - run: sh autogen.sh | |
| - name: build and install | |
| run: | | |
| set -e | |
| mkdir -p build | |
| pushd build | |
| ../configure --prefix=/opt/ovis-4 --enable-etc --enable-munge | |
| make | |
| make install | |
| - name: test preparation | |
| run: | | |
| set -e | |
| mkdir -p /test/logs | |
| SRCDIR=${PWD} | |
| pushd /test/ | |
| ln -s ${SRCDIR}/.github/compat/list_samp.py | |
| ln -s ${SRCDIR}/.github/compat/list_samp.sh | |
| ln -s /opt/ovis-4.3.3/sbin/ldmsd ldmsd-4.3.3 | |
| ln -s /opt/ovis-4/sbin/ldmsd ldmsd-4 | |
| cat > ldmsd-4.3.3.sh << EOF | |
| . /opt/ovis-4.3.3/etc/profile.d/set-ovis-variables.sh | |
| ./ldmsd-4.3.3 \$@ | |
| EOF | |
| chmod 755 ldmsd-4.3.3.sh | |
| cat > ldmsd-4.sh << EOF | |
| . /opt/ovis-4/etc/profile.d/set-ovis-variables.sh | |
| ./ldmsd-4 \$@ & | |
| EOF | |
| chmod 755 ldmsd-4.sh | |
| cat > ldms_ls.sh << EOF | |
| . /opt/ovis-4/etc/profile.d/set-ovis-variables.sh | |
| ldms_ls \$@ | |
| EOF | |
| chmod 755 ldms_ls.sh | |
| cat > ldms_ls-4.3.3.sh << EOF | |
| . /opt/ovis-4.3.3/etc/profile.d/set-ovis-variables.sh | |
| ldms_ls \$@ | |
| EOF | |
| chmod 755 ldms_ls-4.3.3.sh | |
| cat > samp-4.3.3.conf << EOF | |
| load name=meminfo | |
| config name=meminfo producer=ovis-4.3.3 instance=ovis-4.3.3/meminfo | |
| start name=meminfo interval=1000000 offset=0 | |
| EOF | |
| cat > agg-4.3.3.conf << EOF | |
| prdcr_add name=samp type=active host=localhost port=10000 xprt=sock interval=1000000 | |
| prdcr_start name=samp | |
| updtr_add name=updtr interval=1000000 offset=100000 | |
| updtr_prdcr_add name=updtr regex=.* | |
| updtr_start name=updtr | |
| EOF | |
| cat > agg-4.conf << EOF | |
| prdcr_add name=samp type=active host=127.0.0.1 port=10001 xprt=sock interval=1000000 | |
| prdcr_start name=samp | |
| updtr_add name=updtr interval=1000000 offset=200000 | |
| updtr_prdcr_add name=updtr regex=.* | |
| updtr_start name=updtr | |
| EOF | |
| - name: compatibility test | |
| if: ${{ success() }} # all previous steps are good | |
| run: | | |
| error() { | |
| echo "ERROR:" $@ | |
| exit -1 | |
| } | |
| ps_state() { | |
| _PID=$1 | |
| if [[ -d /proc/${_PID} ]]; then | |
| _STATE=$(cat /proc/${_PID}/stat | cut -f 3 -d ' ') | |
| [[ ${_STATE} = "Z" ]] && echo "zombie" || echo "alive" | |
| else | |
| echo "dead" | |
| fi | |
| } | |
| ps_check() { | |
| local _NAME=$1 | |
| local _PID=$2 | |
| echo -n "Checking $_NAME ($_PID) ..." | |
| local _S=$(ps_state $_PID) | |
| echo "$_S" | |
| [[ "$_S" = "alive" ]] || error "process $_NAME ($_PID) is not alive" | |
| } | |
| cd /test | |
| # samp | |
| echo "starting samp-4.3.3" | |
| ./ldmsd-4.3.3.sh -c samp-4.3.3.conf -l logs/samp-4.3.3.log -v INFO -x sock:10000 | |
| echo "starting Python-based ovis-4 sampler with ldms_list" | |
| ./list_samp.sh & | |
| sleep 10 | |
| # agg1 | |
| echo "starting agg-4.3.3" | |
| ./ldmsd-4.3.3.sh -c agg-4.3.3.conf -l logs/agg-4.3.3.log -v INFO -x sock:10001 | |
| sleep 10 | |
| # agg2 | |
| echo "starting agg-4" | |
| ./ldmsd-4.sh -c agg-4.conf -l logs/agg-4.log -v INFO -x sock:10002 | |
| sleep 10 | |
| # check that they're running | |
| SAMP_433_PID=$(pgrep -f samp-4.3.3.conf) | |
| AGG_433_PID=$(pgrep -f agg-4.3.3.conf) | |
| AGG_4_PID=$(pgrep -f agg-4.conf) | |
| LIST_SAMP_4_PID=$(pgrep -f list_samp.py) | |
| echo "--- ldmsd's ---" | |
| pgrep -a ldmsd | |
| echo "---------------" | |
| [[ -n "${SAMP_433_PID}" ]] || error "samp-4.3.3 is not running" | |
| [[ -n "${AGG_433_PID}" ]] || error "agg-4.3.3 is not running" | |
| [[ -n "${AGG_4_PID}" ]] || error "agg-4 is not running" | |
| [[ -n "${LIST_SAMP_4_PID}" ]] || error "list_samp.py is not running" | |
| # ldms_ls-4 to agg-4.3.3 | |
| echo -n "ldms_ls agg-4.3.3 ... " | |
| D0=$( ./ldms_ls.sh -x sock -p 10001 -h 127.0.0.1 ) | |
| echo "result: ${D0}" | |
| # ldms_ls-4 to agg-4 | |
| echo -n "ldms_ls agg-4 ... " | |
| D1=$( ./ldms_ls.sh -x sock -p 10002 ) | |
| echo "result: ${D1}" | |
| # ldms_ls-4.3.3 to agg-4 | |
| echo -n "ldms_ls(4.3.3) agg-4 ... " | |
| D2=$( ./ldms_ls-4.3.3.sh -x sock -p 10002 ) | |
| echo "result: ${D2}" | |
| [[ "${D0}" = "ovis-4.3.3/meminfo" ]] || error "unexpected ldms_ls agg-4.3.3 result" | |
| [[ "${D1}" = "ovis-4.3.3/meminfo" ]] || error "unexpected ldms_ls agg-4 result" | |
| [[ "${D2}" = "ovis-4.3.3/meminfo" ]] || error "unexpected ldms_ls(4.3.3) agg-4 result" | |
| # ldms_ls-4 to agg-4.3.3 -l | |
| echo -n "ldms_ls -l agg-4.3.3 ... " | |
| D0=$( ./ldms_ls.sh -l -x sock -p 10001 -h 127.0.0.1 | grep MemTotal | sed 's/\s\+/ /g' ) | |
| echo "${D0}" | |
| [[ -n "${D0}" ]] || error "cannot get MemTotal from ldms_ls -l" | |
| # ldms_ls-4 to agg-4 -l | |
| echo -n "ldms_ls -l agg-4 ... " | |
| D1=$( ./ldms_ls.sh -l -x sock -p 10002 | grep MemTotal | sed 's/\s\+/ /g' ) | |
| echo "${D1}" | |
| [[ -n "${D1}" ]] || error "cannot get MemTotal from ldms_ls -l" | |
| # ldms_ls-4.3.3 to agg-4 -l | |
| echo -n "ldms_ls-4.3.3 -l agg-4 ... " | |
| D2=$( ./ldms_ls-4.3.3.sh -l -x sock -p 10002 | grep MemTotal | sed 's/\s\+/ /g' ) | |
| echo "${D2}" | |
| [[ -n "${D2}" ]] || error "cannot get MemTotal from ldms_ls -l" | |
| # check if they're the same | |
| [[ "${D0}" == "${D1}" ]] || error "agg-4.3.3 MemTotal != agg-4 MemTotal" | |
| [[ "${D1}" == "${D2}" ]] || error "ldms_ls-4 MemTotal != ldms_ls-4.3.3 MemTotal" | |
| # ldms_ls-4.3.3 to list_samp.py | |
| echo -n "ldms_ls-4.3.3 -l ..." | |
| ./ldms_ls-4.3.3.sh -x sock -p 10003 -l | tee list.txt | |
| RC=$? | |
| echo "Checking results ..." | |
| [[ ${RC} == 0 ]] || error "ldms_ls-4.3.3 crashed" | |
| D3=$( grep "ovis4/list: consistent" list.txt ) | |
| [[ -n "${D3}" ]] || error "bad ldms_ls result" | |
| echo "OK" | |
| # kill samp-4.3.3 so that the set disappeared from agg-4.3.3 | |
| echo "Killing samp-4.3.3" | |
| kill ${SAMP_433_PID} | |
| sleep 4 | |
| echo "--- ldmsd's ---" | |
| pgrep -a ldmsd | |
| echo "---------------" | |
| echo "--- ldmsd's ---" | |
| ps ax | grep ldmsd | |
| echo "---------------" | |
| # ldms_ls to agg-4.3.3 | |
| echo -n "ldms_ls ... " | |
| D0=$(./ldms_ls.sh -x sock -p 10001 -h 127.0.0.1) | |
| [[ -z "${D0}" ]] && echo "dir(agg-4.3.3) is empty (good)" || error "ERROR: dir(agg-4.3.3) is not empty" | |
| # check aggregators | |
| ps_check agg-4.3.3 ${AGG_433_PID} | |
| ps_check agg-4 ${AGG_4_PID} | |
| # ldms_ls to agg-4 | |
| echo -n "ldms_ls ... " | |
| D1=$(./ldms_ls.sh -x sock -p 10002) | |
| [[ -z "${D1}" ]] && echo "dir(agg-4) is empty (good)" || error "ERROR: dir(agg-4) is not empty" | |
| kill ${AGG_433_PID} | |
| kill ${AGG_4_PID} | |
| kill ${LIST_SAMP_4_PID} | |
| echo "DONE!" | |
| - name: compatibility test with munge | |
| if: ${{ success() }} # all previous steps are good | |
| run: | | |
| error() { | |
| echo "ERROR:" $@ | |
| exit -1 | |
| } | |
| ps_state() { | |
| _PID=$1 | |
| if [[ -d /proc/${_PID} ]]; then | |
| _STATE=$(cat /proc/${_PID}/stat | cut -f 3 -d ' ') | |
| [[ ${_STATE} = "Z" ]] && echo "zombie" || echo "alive" | |
| else | |
| echo "dead" | |
| fi | |
| } | |
| ps_check() { | |
| local _NAME=$1 | |
| local _PID=$2 | |
| echo -n "Checking $_NAME ($_PID) ..." | |
| local _S=$(ps_state $_PID) | |
| echo "$_S" | |
| [[ "$_S" = "alive" ]] || error "process $_NAME ($_PID) is not alive" | |
| } | |
| sudo -u munge munged | |
| cd /test | |
| # samp | |
| echo "starting samp-4.3.3 w/munge" | |
| ./ldmsd-4.3.3.sh -c samp-4.3.3.conf -l logs/samp-4.3.3.log -v INFO -x sock:10000 -a munge | |
| # agg1 | |
| echo "starting agg-4.3.3 w/munge" | |
| ./ldmsd-4.3.3.sh -c agg-4.3.3.conf -l logs/agg-4.3.3.log -v INFO -x sock:10001 -a munge | |
| sleep 10 | |
| # agg2 | |
| echo "starting agg-4 w/munge" | |
| ./ldmsd-4.sh -c agg-4.conf -l logs/agg-4.log -v INFO -x sock:10002 -a munge | |
| sleep 10 | |
| # check that they're running | |
| SAMP_433_PID=$(pgrep -f samp-4.3.3.conf) | |
| AGG_433_PID=$(pgrep -f agg-4.3.3.conf) | |
| AGG_4_PID=$(pgrep -f agg-4.conf) | |
| echo "--- ldmsd's ---" | |
| pgrep -a ldmsd | |
| echo "---------------" | |
| [[ -n "${SAMP_433_PID}" ]] || error "samp-4.3.3 is not running" | |
| [[ -n "${AGG_433_PID}" ]] || error "agg-4.3.3 is not running" | |
| [[ -n "${AGG_4_PID}" ]] || error "agg-4 is not running" | |
| # ldms_ls-4 to agg-4.3.3 | |
| echo -n "ldms_ls agg-4.3.3 -a munge ... " | |
| D0=$( ./ldms_ls.sh -x sock -p 10001 -a munge -h 127.0.0.1 ) | |
| echo "result: ${D0}" | |
| # ldms_ls-4 to agg-4 | |
| echo -n "ldms_ls agg-4 -a munge ... " | |
| D1=$( ./ldms_ls.sh -x sock -p 10002 -a munge ) | |
| echo "result: ${D1}" | |
| [[ "${D0}" = "ovis-4.3.3/meminfo" ]] || error "unexpected ldms_ls agg-4.3.3 result" | |
| [[ "${D1}" = "ovis-4.3.3/meminfo" ]] || error "unexpected ldms_ls agg-4 result" | |
| # ldms_ls-4.3.3 to agg-4 | |
| echo -n "ldms_ls(4.3.3) agg-4 -a munge ... " | |
| D2=$( ./ldms_ls-4.3.3.sh -x sock -p 10002 -a munge ) | |
| echo "result: ${D2}" | |
| [[ "${D2}" = "ovis-4.3.3/meminfo" ]] || error "unexpected ldms_ls(4.3.3) agg-4 result" | |
| # ldms_ls-4 to agg-4.3.3 -l | |
| echo -n "ldms_ls -l agg-4.3.3 -a munge ... " | |
| D0=$( ./ldms_ls.sh -l -x sock -p 10001 -a munge -h 127.0.0.1 | grep MemTotal | sed 's/\s\+/ /g' ) | |
| echo "${D0}" | |
| [[ -n "${D0}" ]] || error "cannot get MemTotal from ldms_ls -l" | |
| # ldms_ls-4 to agg-4 -l | |
| echo -n "ldms_ls -l agg-4 -a munge ... " | |
| D1=$( ./ldms_ls.sh -l -x sock -p 10002 -a munge | grep MemTotal | sed 's/\s\+/ /g' ) | |
| echo "${D1}" | |
| [[ -n "${D1}" ]] || error "cannot get MemTotal from ldms_ls -l" | |
| # ldms_ls-4.3.3 to agg-4 -l | |
| echo -n "ldms_ls-4.3.3 -l agg-4 -a munge ... " | |
| D2=$( ./ldms_ls-4.3.3.sh -l -x sock -p 10002 -a munge | grep MemTotal | sed 's/\s\+/ /g' ) | |
| echo "${D2}" | |
| [[ -n "${D2}" ]] || error "cannot get MemTotal from ldms_ls -l" | |
| # check if they're the same | |
| [[ "${D0}" == "${D1}" ]] || error "agg-4.3.3 MemTotal != agg-4 MemTotal" | |
| [[ "${D1}" == "${D2}" ]] || error "ldms_ls-4 MemTotal != ldms_ls-4.3.3 MemTotal" | |
| # kill samp-4.3.3 so that the set disappeared from agg-4.3.3 | |
| echo "Killing samp-4.3.3" | |
| kill ${SAMP_433_PID} | |
| sleep 4 | |
| echo "--- ldmsd's ---" | |
| pgrep -a ldmsd | |
| echo "---------------" | |
| echo "--- ldmsd's ---" | |
| ps ax | grep ldmsd | |
| echo "---------------" | |
| # ldms_ls to agg-4.3.3 | |
| echo -n "ldms_ls agg-4.3.3 -a munge ... " | |
| D0=$(./ldms_ls.sh -x sock -p 10001 -a munge -h 127.0.0.1 ) | |
| [[ -z "${D0}" ]] && echo "dir(agg-4.3.3) is empty (good)" || error "ERROR: dir(agg-4.3.3) is not empty" | |
| # check aggregators | |
| ps_check agg-4.3.3 ${AGG_433_PID} | |
| ps_check agg-4 ${AGG_4_PID} | |
| # ldms_ls to agg-4 | |
| echo -n "ldms_ls agg-4 -a munge ... " | |
| D1=$(./ldms_ls.sh -x sock -p 10002 -a munge ) | |
| [[ -z "${D1}" ]] && echo "dir(agg-4) is empty (good)" || error "ERROR: dir(agg-4) is not empty" | |
| echo "DONE!" | |
| - name: enum compatibility check | |
| if: ${{ success() }} # all previous steps are good | |
| run: | | |
| set -e | |
| pushd .github/enum-check | |
| for V in 4.3.{3..6}; do | |
| ./enumsym.py /root/ovis-ldms-${V}/ -e ovis-${V}-enum.list > /test/${V}.enum | |
| ./enumsym.py ../../ -e ovis-${V}-enum.list > /test/HEAD-${V}.enum | |
| done | |
| for V in 4.3.{3..6}; do | |
| diff -u /test/${V}.enum /test/HEAD-${V}.enum > /test/enum-${V}.diff | |
| done |