Re: [Gluster-devel] [Gluster-users] CentOS 7: Gluster Test Framework testcases failure
I forgot to mention other issues I faced and temporary fixes 1) Most of the testcases were failing due to /var/run/gluster directory not found. grep: /var/run/gluster/: No such file or directory ls: cannot access /var/run/gluster: No such file or directory I created the directory mkdir /var/run/gluster and rerun the tests and testcases passed where they were failing earlier. Justin, Is CentOS 7 considered for regression testing or not ? How about Fedora 19/20, which one do you recommend ? 2) ./tests/bugs/../include.rc: line 123: setfattr: command not found Setfattr command is part of attr package, # yum install attr Thanks, Kiran. On Thu, Sep 25, 2014 at 7:01 PM, Kiran Patil kirantpa...@gmail.com wrote: The below testcases failing are related to xfs, cluster and others.. The hardcoded ones I have fixed temporarily by providing the absolute pathname. Testcase /tests/bugs/bug-767095.t is fixed by changing awk parameter $5 to $4. Testcase tests/bugs/bug-861542.t is failing at EXPECT N/A port_field $V0 '0'; # volume status If I change its value to to '1' it passes, is that correct ? I will keep posting the new findings and possible fixes. Test Summary Report ./tests/bugs/bug-802417.t (Wstat: 0 Tests: 39 Failed: 2) Failed tests: 28, 31 ./tests/bugs/bug-821056.t (Wstat: 0 Tests: 21 Failed: 1) Failed test: 15 ./tests/bugs/bug-861542.t (Wstat: 0 Tests: 13 Failed: 1) Failed test: 10 ./tests/bugs/bug-908146.t (Wstat: 0 Tests: 10 Failed: 2) Failed tests: 8-9 ./tests/bugs/bug-913555.t (Wstat: 0 Tests: 11 Failed: 4) Failed tests: 4-6, 9 ./tests/bugs/bug-948686.t (Wstat: 0 Tests: 19 Failed: 8) Failed tests: 5-7, 9, 11, 13-15 ./tests/bugs/bug-948729/bug-948729-force.t (Wstat: 0 Tests: 35 Failed: 14) Failed tests: 15, 17, 19, 21, 24-27, 29-31, 33-35 ./tests/bugs/bug-948729/bug-948729-mode-script.t (Wstat: 0 Tests: 35 Failed: 14) Failed tests: 15, 17, 19, 21, 24-27, 29-31, 33-35 ./tests/bugs/bug-948729/bug-948729.t(Wstat: 0 Tests: 23 Failed: 4) Failed tests: 12, 15, 19, 23 Files=124, Tests=2031, 4470 wallclock secs ( 1.57 usr 0.24 sys + 259.31 cusr 220.81 csys = 481.93 CPU) Result: FAIL Thanks, Kiran. On Thu, Sep 25, 2014 at 2:17 PM, Niels de Vos nde...@redhat.com wrote: On Thu, Sep 25, 2014 at 12:49:21PM +0530, Kiran Patil wrote: I installed the 'psmisc' package and it installed killall command and reverted pkill to killall in include.rc file. Testcases started executing properly and will send tests failure report soon. Thanks, I've added 'psmisc' to the list of packages in the wiki: - http://www.gluster.org/community/documentation/index.php/Using_the_Gluster_Test_Framework#Preparation_steps_for_CentOS_7_.28only.29 Niels Thanks, Kiran. On Thu, Sep 25, 2014 at 12:32 PM, Niels de Vos nde...@redhat.com wrote: On Thu, Sep 25, 2014 at 10:45:49AM +0530, Kiran Patil wrote: pkill expects only one pattern, so I did as below in tests/include.rc file and test cases started working fine. pkill glusterfs 2/dev/null || true; pkill glusterfsd 2/dev/null || true; pkill glusterd 2/dev/null || true; Sorry, I'm a little late to the party, but the 'killall' command should be available for CentOS-7 too. It seems to be part of the 'psmisc' package. I guess we should add this as a dependency on the wiki page. Could you check if that works for you too? If not, and you are interested, I'll help you posting a patch to make the pkill change. Thanks, Niels On Wed, Sep 24, 2014 at 6:48 PM, Justin Clift jus...@gluster.org wrote: On 24/09/2014, at 2:07 PM, Kiran Patil wrote: Some of the reasons I have found so far are as below, 1. Cleanup operation does not work since killall is not part of CentOS 7 2. I used pkill and still testcases fail at first step Ex: TEST glusterd 3. Subsequent running of testcases does not proceed and hangs at the first testcase (tests/basic/bd.t) This sounds like there could be a few challenges then. I'm setting up a new Fedora 20 (or 21 alpha) VM in Rackspace for running btrfs regression tests on. Guessing that will experience these same problems as your CentOS 7 test run, so I'm definitely interested in this too. + Justin -- GlusterFS - http://www.gluster.org An open source, distributed file system scaling to several petabytes, and handling thousands of clients. My personal twitter: twitter.com/realjustinclift ___ Gluster-users mailing list gluster-us...@gluster.org
Re: [Gluster-devel] [Gluster-users] CentOS 7: Gluster Test Framework testcases failure
On 26/09/2014, at 7:16 AM, Kiran Patil wrote: snip Justin, Is CentOS 7 considered for regression testing or not ? How about Fedora 19/20, which one do you recommend ? The only OS that I know for sure regression testing works on is CentOS 6.5. We _want_ regression testing to work on CentOS 7 and all recent Fedora's, but I don't know that anyone has yet put serious time into it. :( For btrfs testing, we're going to have to, as CentOS 6.5 isn't recent enough for that to work. Please keep testing CentOS 7... if you have the time/inclination to delve into fixing the failures. Otherwise you'll probably just get amazingly frustrated. :( That being said, once you've discovered and resolved all of the dependency issues, there might not be that many failures after all. :) + Justin -- GlusterFS - http://www.gluster.org An open source, distributed file system scaling to several petabytes, and handling thousands of clients. My personal twitter: twitter.com/realjustinclift ___ Gluster-devel mailing list Gluster-devel@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-devel
Re: [Gluster-devel] [Gluster-users] CentOS 7: Gluster Test Framework testcases failure
On 09/26/2014 04:22 PM, Justin Clift wrote: On 26/09/2014, at 7:16 AM, Kiran Patil wrote: snip Justin, Is CentOS 7 considered for regression testing or not ? How about Fedora 19/20, which one do you recommend ? The only OS that I know for sure regression testing works on is CentOS 6.5. We _want_ regression testing to work on CentOS 7 and all recent Fedora's, but I don't know that anyone has yet put serious time into it. :( I certainly do not expect to see a lot of core problems on CentOS 7 with regression tests. The chief problem is that regression tests are not very portable across systems and we need to put in some incremental effort to clean that up. FWIW, I run regression tests on my F20 laptop and most tests usually pass (apart from known offenders like mgmt-v3-locks.t ;-)). Please keep testing CentOS 7... if you have the time/inclination to delve into fixing the failures. Kiran - thanks for your report. Would it be possible to determine what's causing the tests to fail in your setup? Running tests with DEBUG=1 or set -x in the failing testcases will help us understand the problem better. Thanks, Vijay ___ Gluster-devel mailing list Gluster-devel@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-devel
Re: [Gluster-devel] [Gluster-users] CentOS 7: Gluster Test Framework testcases failure
Please keep testing CentOS 7... if you have the time/inclination to delve into fixing the failures. Kiran - thanks for your report. Would it be possible to determine what's causing the tests to fail in your setup? Running tests with DEBUG=1 or set -x in the failing testcases will help us understand the problem better. Vijay - where should I upload each testcase output with DEBUG=1 enabled ? Thanks, Kiran. ___ Gluster-devel mailing list Gluster-devel@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-devel
Re: [Gluster-devel] [Gluster-users] CentOS 7: Gluster Test Framework testcases failure
I installed the 'psmisc' package and it installed killall command and reverted pkill to killall in include.rc file. Testcases started executing properly and will send tests failure report soon. Thanks, Kiran. On Thu, Sep 25, 2014 at 12:32 PM, Niels de Vos nde...@redhat.com wrote: On Thu, Sep 25, 2014 at 10:45:49AM +0530, Kiran Patil wrote: pkill expects only one pattern, so I did as below in tests/include.rc file and test cases started working fine. pkill glusterfs 2/dev/null || true; pkill glusterfsd 2/dev/null || true; pkill glusterd 2/dev/null || true; Sorry, I'm a little late to the party, but the 'killall' command should be available for CentOS-7 too. It seems to be part of the 'psmisc' package. I guess we should add this as a dependency on the wiki page. Could you check if that works for you too? If not, and you are interested, I'll help you posting a patch to make the pkill change. Thanks, Niels On Wed, Sep 24, 2014 at 6:48 PM, Justin Clift jus...@gluster.org wrote: On 24/09/2014, at 2:07 PM, Kiran Patil wrote: Some of the reasons I have found so far are as below, 1. Cleanup operation does not work since killall is not part of CentOS 7 2. I used pkill and still testcases fail at first step Ex: TEST glusterd 3. Subsequent running of testcases does not proceed and hangs at the first testcase (tests/basic/bd.t) This sounds like there could be a few challenges then. I'm setting up a new Fedora 20 (or 21 alpha) VM in Rackspace for running btrfs regression tests on. Guessing that will experience these same problems as your CentOS 7 test run, so I'm definitely interested in this too. + Justin -- GlusterFS - http://www.gluster.org An open source, distributed file system scaling to several petabytes, and handling thousands of clients. My personal twitter: twitter.com/realjustinclift ___ Gluster-users mailing list gluster-us...@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-users ___ Gluster-devel mailing list Gluster-devel@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-devel
Re: [Gluster-devel] [Gluster-users] CentOS 7: Gluster Test Framework testcases failure
On Thu, Sep 25, 2014 at 12:49:21PM +0530, Kiran Patil wrote: I installed the 'psmisc' package and it installed killall command and reverted pkill to killall in include.rc file. Testcases started executing properly and will send tests failure report soon. Thanks, I've added 'psmisc' to the list of packages in the wiki: - http://www.gluster.org/community/documentation/index.php/Using_the_Gluster_Test_Framework#Preparation_steps_for_CentOS_7_.28only.29 Niels Thanks, Kiran. On Thu, Sep 25, 2014 at 12:32 PM, Niels de Vos nde...@redhat.com wrote: On Thu, Sep 25, 2014 at 10:45:49AM +0530, Kiran Patil wrote: pkill expects only one pattern, so I did as below in tests/include.rc file and test cases started working fine. pkill glusterfs 2/dev/null || true; pkill glusterfsd 2/dev/null || true; pkill glusterd 2/dev/null || true; Sorry, I'm a little late to the party, but the 'killall' command should be available for CentOS-7 too. It seems to be part of the 'psmisc' package. I guess we should add this as a dependency on the wiki page. Could you check if that works for you too? If not, and you are interested, I'll help you posting a patch to make the pkill change. Thanks, Niels On Wed, Sep 24, 2014 at 6:48 PM, Justin Clift jus...@gluster.org wrote: On 24/09/2014, at 2:07 PM, Kiran Patil wrote: Some of the reasons I have found so far are as below, 1. Cleanup operation does not work since killall is not part of CentOS 7 2. I used pkill and still testcases fail at first step Ex: TEST glusterd 3. Subsequent running of testcases does not proceed and hangs at the first testcase (tests/basic/bd.t) This sounds like there could be a few challenges then. I'm setting up a new Fedora 20 (or 21 alpha) VM in Rackspace for running btrfs regression tests on. Guessing that will experience these same problems as your CentOS 7 test run, so I'm definitely interested in this too. + Justin -- GlusterFS - http://www.gluster.org An open source, distributed file system scaling to several petabytes, and handling thousands of clients. My personal twitter: twitter.com/realjustinclift ___ Gluster-users mailing list gluster-us...@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-users ___ Gluster-devel mailing list Gluster-devel@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-devel
Re: [Gluster-devel] [Gluster-users] CentOS 7: Gluster Test Framework testcases failure
The below testcases failing are related to xfs, cluster and others.. The hardcoded ones I have fixed temporarily by providing the absolute pathname. Testcase /tests/bugs/bug-767095.t is fixed by changing awk parameter $5 to $4. Testcase tests/bugs/bug-861542.t is failing at EXPECT N/A port_field $V0 '0'; # volume status If I change its value to to '1' it passes, is that correct ? I will keep posting the new findings and possible fixes. Test Summary Report ./tests/bugs/bug-802417.t (Wstat: 0 Tests: 39 Failed: 2) Failed tests: 28, 31 ./tests/bugs/bug-821056.t (Wstat: 0 Tests: 21 Failed: 1) Failed test: 15 ./tests/bugs/bug-861542.t (Wstat: 0 Tests: 13 Failed: 1) Failed test: 10 ./tests/bugs/bug-908146.t (Wstat: 0 Tests: 10 Failed: 2) Failed tests: 8-9 ./tests/bugs/bug-913555.t (Wstat: 0 Tests: 11 Failed: 4) Failed tests: 4-6, 9 ./tests/bugs/bug-948686.t (Wstat: 0 Tests: 19 Failed: 8) Failed tests: 5-7, 9, 11, 13-15 ./tests/bugs/bug-948729/bug-948729-force.t (Wstat: 0 Tests: 35 Failed: 14) Failed tests: 15, 17, 19, 21, 24-27, 29-31, 33-35 ./tests/bugs/bug-948729/bug-948729-mode-script.t (Wstat: 0 Tests: 35 Failed: 14) Failed tests: 15, 17, 19, 21, 24-27, 29-31, 33-35 ./tests/bugs/bug-948729/bug-948729.t(Wstat: 0 Tests: 23 Failed: 4) Failed tests: 12, 15, 19, 23 Files=124, Tests=2031, 4470 wallclock secs ( 1.57 usr 0.24 sys + 259.31 cusr 220.81 csys = 481.93 CPU) Result: FAIL Thanks, Kiran. On Thu, Sep 25, 2014 at 2:17 PM, Niels de Vos nde...@redhat.com wrote: On Thu, Sep 25, 2014 at 12:49:21PM +0530, Kiran Patil wrote: I installed the 'psmisc' package and it installed killall command and reverted pkill to killall in include.rc file. Testcases started executing properly and will send tests failure report soon. Thanks, I've added 'psmisc' to the list of packages in the wiki: - http://www.gluster.org/community/documentation/index.php/Using_the_Gluster_Test_Framework#Preparation_steps_for_CentOS_7_.28only.29 Niels Thanks, Kiran. On Thu, Sep 25, 2014 at 12:32 PM, Niels de Vos nde...@redhat.com wrote: On Thu, Sep 25, 2014 at 10:45:49AM +0530, Kiran Patil wrote: pkill expects only one pattern, so I did as below in tests/include.rc file and test cases started working fine. pkill glusterfs 2/dev/null || true; pkill glusterfsd 2/dev/null || true; pkill glusterd 2/dev/null || true; Sorry, I'm a little late to the party, but the 'killall' command should be available for CentOS-7 too. It seems to be part of the 'psmisc' package. I guess we should add this as a dependency on the wiki page. Could you check if that works for you too? If not, and you are interested, I'll help you posting a patch to make the pkill change. Thanks, Niels On Wed, Sep 24, 2014 at 6:48 PM, Justin Clift jus...@gluster.org wrote: On 24/09/2014, at 2:07 PM, Kiran Patil wrote: Some of the reasons I have found so far are as below, 1. Cleanup operation does not work since killall is not part of CentOS 7 2. I used pkill and still testcases fail at first step Ex: TEST glusterd 3. Subsequent running of testcases does not proceed and hangs at the first testcase (tests/basic/bd.t) This sounds like there could be a few challenges then. I'm setting up a new Fedora 20 (or 21 alpha) VM in Rackspace for running btrfs regression tests on. Guessing that will experience these same problems as your CentOS 7 test run, so I'm definitely interested in this too. + Justin -- GlusterFS - http://www.gluster.org An open source, distributed file system scaling to several petabytes, and handling thousands of clients. My personal twitter: twitter.com/realjustinclift ___ Gluster-users mailing list gluster-us...@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-users ___ Gluster-devel mailing list Gluster-devel@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-devel