BIBO on wp289b

[zhihao]

(08-18-2021 08:31 PM)lingu Wrote:  

(08-18-2021 08:27 PM)zhihao Wrote:  install sage on sage_helper(wp289b-1) successfully

better specify the commit id so that we know what version you installed.

ok, added the commit id

[zhihao]

run "auntie list" on sage_helper and get log. The origin log is 91MB, so I only copy the last few lines.

[zhihao]

install think on wp289b-6, failed on test

Code:

[**Test basic mechanisms]
[Wed Aug 18 21:40:05 CST 2021] null.sh => null.c PASSED
[Wed Aug 18 21:40:10 CST 2021] newr.sh => newr.c PASSED
[Wed Aug 18 21:40:14 CST 2021] strlen.sh => strlen.c PASSED
[Wed Aug 18 21:40:18 CST 2021] strnlen.sh => strnlen.c
PASSED
[Wed Aug 18 21:40:22 CST 2021] asm.sh => asm.c PASSED
[Wed Aug 18 21:40:26 CST 2021] bmsg.sh => sum-multiple-languages FAIL
Time used: 0:20.45
make: *** [test] Error 1
Connection to 143.64.66.176 closed.
make: *** [test] Error 2
/thinker/globe/udata/root/Limbo/Packages/wp289b/think.tar.gz failed with exit status 2
(END)

install sage on wp289b-6, failed on test

Code:

====== configure is done ====== [0818-21:52:49]
          lockstep mark /thinker/globe/.think/lockstep//wp289b-6/sage
          lockstep wait wp289b-6/sage
Check whether wp289b-6/sage is ready...
wp289b-6/sage is not yet ready. Wait for a while and check again...(Wed Aug 18 21:52:49 CST 2021)
install completed for sage
set -o errexit
if [[ "True" == "True" ]]; then
ssh sage@143.64.66.176 "cd ~/sage; make test"
else
echo "I am not sage_portal (143.64.66.176). Skip testing."
fi
lockstep mark "$(hostname)/sage/tested"
lockstep mark "$(hostname)/sage/decent/workdone"
make -C correctness test
make[1]: Entering directory `/home/sage/sage/correctness'
=== Begin Sage Service Test Set ===
./sage_service/sage-service-start.sh
timeout: failed to run command ‘/bin/sage’: No such file or directory
make[1]: Leaving directory `/home/sage/sage/correctness'
make[1]: *** [sage-tests] Error 127
make: *** [test] Error 2
make: *** [test] Error 2
/thinker/globe/udata/root/Limbo/Packages/wp289b/sage.tar.gz failed with exit status 2
(END)

I insert a case, and query fails.

try list cases, it is

Code:

dep.grid letter 792

forget update bibo on all nodes, updated now and make bibo-ext.pcf contain only one line "bibo_doc_cnt: 0"

[zhihao]

戚经理早上提到相较于删除功能，搜索时间缩短更为重要一些，删除功能平时他们不怎么用，而搜索时间13秒左右不够快。

[lingu]

(08-19-2021 02:53 PM)zhihao Wrote:  戚经理早上提到相较于删除功能，搜索时间缩短更为重要一些，删除功能平时他们不怎么用，而搜索时间13秒左右不够快。

Yes that's one important thing to improve. Pls add a todo.

Another important thing is concurrent query support.

[lingu]

20220717/lingu visits tiejian and find wp289b bibo down. think test fails on null and think reports that it could not contact containers.

i find the iptables rules on wp289b-6 drop all incoming packets except for a few ports. but all 6 wp289b nodes are added to be "ACCEPT" nodes.

i find the routing table uses 143.64.66.254 as the default router on wp289b-6, and packets go from wp289b-6 to 143.64.66.254 then 143.64.64.254 then wp289b-1 (143.64.64.145).

so the forwarding rule may be a factor in this setting.

however, even wp289b-5 (143.64.66.190) cannot tcp to wp289b-6:30480 with iptables service running on wp289b-6. Without the iptables service, the tcp to wp289b-6:30480 is fine.

In /etc/sysconfig/iptables on wp289b-6, there is a line '-A INPUT -j REJECT --reject-with-icmp-host-prohibited' before the ACCEPT rules for the wp289b IPs. iptables-save does not show this line on wp289b-1 or wp289b-3, but wp289b-5 has it. I move the line to the end of the filter block. now tcp from wp289b-5 to wp289b-6 is successful.

i went through all iptables files and restarted all iptables services on 6 nodes.

then think test works.

[zhihao]

(08-20-2021 12:02 AM)lingu Wrote:  

(08-19-2021 02:53 PM)zhihao Wrote:  戚经理早上提到相较于删除功能，搜索时间缩短更为重要一些，删除功能平时他们不怎么用，而搜索时间13秒左右不够快。

Yes that's one important thing to improve. Pls add a todo.

Another important thing is concurrent query support.

ok, add two TODOs to D info http://rar.shufangkeji.com:60380/showthr...p?tid=9162