The following has be reported by IBM LTC: Benchmarking(ab) failed in RHR2 testing Hardware Environment: Hostname: plinuxt8.cn.ibm.com Model: p630 HMC : plinuxt1.cn.ibm.com Firmware Environment: RM3R041029 Software Environment: RHEL4 Beta2 isos. kernel : 2.6.9-1.648_EL rhr2 packages: rhr2-rhel4-0.9-14.9b.noarch.rpm Steps to Reproduce: 1. Install rhr2-rhel4-0.9-14.9b.noarch.rpm 2. setup rhr2 test env 3. do NETWORK sub-test by redhat-ready. Actual Results: NETWORK test fail Expected Results: NETWORK test pass Additional Information: [root@plinuxt8 14]# cat error.log 12+0 records in 12+0 records out apr_poll: The timeout specified has expired (70007) [root@plinuxt8 14]# uname -r 2.6.9-1.648_EL Jun, Please provide more infomation on setting the env. and how the test was run. Mark, What was the resolution for bug 11829? I can not get anything out of that bug. Was it a testcase or kernel issue? Thanks NETWORK requires a remote machine to talk to. On this remote machine a user with ssh, apache bench, and mount permissions is required. The default user is "root"; this may be overridden in /etc/rhr/tests.conf. and also we should privode the remote machine to be contacted in /etc/rhr/tests.conf. Please do RHR configuration under directory /etc/rhr. Jun, is this recreatable? and how often?
----- Additional Comments From bujunhui.com 2004-11-18 20:29 EDT ------- Yes, this is recreatable. see following function in NETWORK script. tcp_test(){ service httpd start scp $httptestfile $RHRUSER@$SERVER:~/httptest.file ssh -l $RHRUSER -x $SERVER "ab -c 128 -k -n 256 -p ~/httptest.file $ipaddr/httptest.file" } ab is a tool for benchmarking Apache (HTTP) server. option "-n" -- Number of requests to perform for the benchmarking session. If we decrease number of requests, we can pass this benchmark. otherwise if we use default(256) number, it failed exactly.
----- Additional Comments From bujunhui.com 2004-11-19 05:19 EDT ------- Recreate this bug on plinuxt6 which is a 630(Power4).
----- Additional Comments From bujunhui.com 2004-11-19 05:25 EDT ------- As well recreate this bug on castorlp1.upt.austin.ibm.com which is a SF2(Power5)
Changing product.
In discussion with rlandry, it appears this is a known issue (dealing with apr_pool timeout issues). Rob, are there any workarounds to the problem?
still work issue; still effects 0.9-16-pre-rc1.1 package.
----- Additional Comments From bujunhui.com 2004-12-27 04:00 EDT ------- Recreate this bug with rhr2-rhel4-0.9-14.9i.noarch.rpm on rhel4-pre-rc2. Fyi. [root@romelp1 tests]# tail /var/log/rhr/tests/NETWORK/1/output.log Starting httpd: [ OK ] + scp /var/www/html/httptest.file 'root.111.249:~/httptest.file' + ssh -l root -x 9.3.111.249 'ab -c 128 -k -n 256 -p ~/httptest.file 9.3.111.153/httptest.file' This is ApacheBench, Version 2.0.41-dev <$Revision: 1.141 $> apache-2.0 Copyright (c) 1996 Adam Twiss, Zeus Technology Ltd, http://www.zeustech.net/ Copyright (c) 1998-2002 The Apache Software Foundation, http://www.apache.org/ Benchmarking 9.3.111.153 (be patient) apr_poll: The timeout specified has expired (70007) 196878881320399768977964849610609286410
---- Additional Comments From khoa.com 2005-01-18 10:21 EST ------- I'd like to move RHR bugs to Ship Issue as this will impact the RH certification of our power hardware. Thanks.
Rob and IBM, is this still an issue with the latest RC rhr2 shipped last week ?
---- Additional Comments From marksmit.com 2005-01-18 13:22 EST ------- CSDL will test the latest rhr2 on RC bits this weekend. results by 1/24.
---- Additional Comments From bujunhui.com 2005-01-20 06:04 EST ------- Fixed in latest RHR RC.
this is closed on the IBM side. This can be closed on the Red Hat side
closing based on markwiz comment on 2/24