Discussion:
[Linux-ha-dev] [Patch:crmsh] Correction of the mistake of the processing to transfer comment.
r***@ybb.ne.jp
2014-01-10 07:27:47 UTC
Permalink
Hi Dejan,

I send a patch of crmsh1.2.5.
Similar correction seems to be necessary for latest crmsh.

Best Regards,
Hideo Yamauchi
Kristoffer Grönlund
2014-01-10 12:54:46 UTC
Permalink
On Fri, 10 Jan 2014 16:27:47 +0900 (JST)
Post by r***@ybb.ne.jp
Hi Dejan,
I send a patch of crmsh1.2.5.
Similar correction seems to be necessary for latest crmsh.
Best Regards,
Hideo Yamauchi
Hello,

Thank you for the patch!

I tried applying the patch to the latest crmsh, but when running the
regression test suite, I got an error. I think the patch is fixing a
bug, but unfortunately it seems to reveal a different problem.

Maybe you can help me figure out what is going wrong!

Failing test case output included below:

[ 77s] Fri Jan 10 12:38:47 UTC 2014: BEGIN testcase edit
[ 77s] ----------
[ 77s] testcase edit failed
[ 77s] output is in crmtestout/edit.out
[ 77s] diff (from crmtestout/edit.diff):
[ 77s] --- /usr/share/crmsh/tests/testcases/edit.exp 2014-01-10 12:38:36.000000000 +0000
[ 77s] +++ - 2014-01-10 12:38:53.149599264 +0000
[ 77s] @@ -84,4 +84,8 @@
[ 77s] .TRY configure rsc_defaults $id="rsc_options" failure-timeout=10m
[ 77s] +INFO: object loc-d1 cannot be represented in the CLI notation
[ 77s] .TRY configure filter "sed 's/2m/60s/'" cib-bootstrap-options
[ 77s] +INFO: object loc-d1 cannot be represented in the CLI notation
[ 77s] +INFO: object loc-d1 cannot be represented in the CLI notation
[ 77s] .TRY configure show rsc_options
[ 77s] +INFO: object loc-d1 cannot be represented in the CLI notation
[ 77s] rsc_defaults $id="rsc_options" \
[ 77s] @@ -89,3 +93,6 @@
[ 77s] .TRY configure property stonith-enabled=true
[ 77s] +INFO: object loc-d1 cannot be represented in the CLI notation
[ 77s] +INFO: object loc-d1 cannot be represented in the CLI notation
[ 77s] .TRY configure show cib-bootstrap-options
[ 77s] +INFO: object loc-d1 cannot be represented in the CLI notation
[ 77s] property $id="cib-bootstrap-options" \
[ 77s] @@ -94,4 +101,8 @@
[ 77s] .TRY configure filter 'sed "s/stonith-enabled=.true.//"'
[ 77s] +INFO: object loc-d1 cannot be represented in the CLI notation
[ 77s] +ERROR: 13: syntax: Unknown command near <xml> parsing 'xml <rsc_location id=loc-d1 rsc=d1> <!--# --> <rule id=r1 score=-INFINITY boolean-op=or> <expression operation=not_defined attribute=webserver id=loc-d1-expression/> <expression attribute=mem type=number operation=lte value=0 id=loc-d1-expression-3/> </rule> <rule id=loc-d1-rule score=-INFINITY> <expression operation=not_defined attribute=a2 id=loc-d1-expression-2/> </rule> <rule id=r2 score-attribute=webserver> <expression operation=defined attribute=webserver id=loc-d1-expression-0/> </rule> <!--# --> </rsc_location>'
[ 77s] .TRY configure show cib-bootstrap-options
[ 77s] +INFO: object loc-d1 cannot be represented in the CLI notation
[ 77s] property $id="cib-bootstrap-options" \
[ 77s] - default-action-timeout="60s"
[ 77s] + default-action-timeout="60s" \
[ 77s] + stonith-enabled="true"
[ 77s] ----------
[ 77s] Fri Jan 10 12:38:53 UTC 2014: END testcase edit
--
// Kristoffer Grönlund
// ***@suse.com
r***@ybb.ne.jp
2014-01-12 04:15:39 UTC
Permalink
Hi Kristoffer,

Thank you for comment.

I will look the day after tomorrow.

Best Regards,
Hideo Yamauchi.
Post by Kristoffer Grönlund
On Fri, 10 Jan 2014 16:27:47 +0900 (JST)
Post by r***@ybb.ne.jp
Hi Dejan,
I send a patch of crmsh1.2.5.
Similar correction seems to be necessary for latest crmsh.
Best Regards,
Hideo Yamauchi
Hello,
Thank you for the patch!
I tried applying the patch to the latest crmsh, but when running the
regression test suite, I got an error. I think the patch is fixing a
bug, but unfortunately it seems to reveal a different problem.
Maybe you can help me figure out what is going wrong!
[   77s] Fri Jan 10 12:38:47 UTC 2014: BEGIN testcase edit
[   77s] ----------
[   77s] testcase edit failed
[   77s] output is in crmtestout/edit.out
[   77s] --- /usr/share/crmsh/tests/testcases/edit.exp    2014-01-10 12:38:36.000000000 +0000
[   77s] +++ -    2014-01-10 12:38:53.149599264 +0000
[   77s]  .TRY configure rsc_defaults $id="rsc_options" failure-timeout=10m
[   77s] +INFO: object loc-d1 cannot be represented in the CLI notation
[   77s]  .TRY configure filter "sed 's/2m/60s/'" cib-bootstrap-options
[   77s] +INFO: object loc-d1 cannot be represented in the CLI notation
[   77s] +INFO: object loc-d1 cannot be represented in the CLI notation
[   77s]  .TRY configure show rsc_options
[   77s] +INFO: object loc-d1 cannot be represented in the CLI notation
[   77s]  rsc_defaults $id="rsc_options" \
[   77s]  .TRY configure property stonith-enabled=true
[   77s] +INFO: object loc-d1 cannot be represented in the CLI notation
[   77s] +INFO: object loc-d1 cannot be represented in the CLI notation
[   77s]  .TRY configure show cib-bootstrap-options
[   77s] +INFO: object loc-d1 cannot be represented in the CLI notation
[   77s]  property $id="cib-bootstrap-options" \
[   77s]  .TRY configure filter 'sed "s/stonith-enabled=.true.//"'
[   77s] +INFO: object loc-d1 cannot be represented in the CLI notation
[   77s] +ERROR: 13: syntax: Unknown command near <xml> parsing 'xml <rsc_location id=loc-d1 rsc=d1> <!--# --> <rule id=r1 score=-INFINITY boolean-op=or> <expression operation=not_defined attribute=webserver id=loc-d1-expression/> <expression attribute=mem type=number operation=lte value=0 id=loc-d1-expression-3/> </rule> <rule id=loc-d1-rule score=-INFINITY> <expression operation=not_defined attribute=a2 id=loc-d1-expression-2/> </rule> <rule id=r2 score-attribute=webserver> <expression operation=defined attribute=webserver id=loc-d1-expression-0/> </rule> <!--# --> </rsc_location>'
[   77s]  .TRY configure show cib-bootstrap-options
[   77s] +INFO: object loc-d1 cannot be represented in the CLI notation
[   77s]  property $id="cib-bootstrap-options" \
[   77s] -    default-action-timeout="60s"
[   77s] +    default-action-timeout="60s" \
[   77s] +    stonith-enabled="true"
[   77s] ----------
[   77s] Fri Jan 10 12:38:53 UTC 2014: END testcase edit
--
// Kristoffer Grönlund
r***@ybb.ne.jp
2014-01-14 03:16:53 UTC
Permalink
Hi Krstoffer,

As for the problem, contents of rpm of crmsh1.2.5 which we used seem to have a problem somehow or other.
The problem did not occur in crm which I made from a source code of crmsh1.2.5.

The application of the patch which I donated seems to be unnecessary.

I confirm the details and contact me again.

Best Regards,
Hideo Yamauchi.
Post by r***@ybb.ne.jp
Hi Kristoffer,
Thank you for comment.
I will look the day after tomorrow.
Best Regards,
Hideo Yamauchi.
Post by Kristoffer Grönlund
On Fri, 10 Jan 2014 16:27:47 +0900 (JST)
Post by r***@ybb.ne.jp
Hi Dejan,
I send a patch of crmsh1.2.5.
Similar correction seems to be necessary for latest crmsh.
Best Regards,
Hideo Yamauchi
Hello,
Thank you for the patch!
I tried applying the patch to the latest crmsh, but when running the
regression test suite, I got an error. I think the patch is fixing a
bug, but unfortunately it seems to reveal a different problem.
Maybe you can help me figure out what is going wrong!
[   77s] Fri Jan 10 12:38:47 UTC 2014: BEGIN testcase edit
[   77s] ----------
[   77s] testcase edit failed
[   77s] output is in crmtestout/edit.out
[   77s] --- /usr/share/crmsh/tests/testcases/edit.exp    2014-01-10 12:38:36.000000000 +0000
[   77s] +++ -    2014-01-10 12:38:53.149599264 +0000
[   77s]  .TRY configure rsc_defaults $id="rsc_options" failure-timeout=10m
[   77s] +INFO: object loc-d1 cannot be represented in the CLI notation
[   77s]  .TRY configure filter "sed 's/2m/60s/'" cib-bootstrap-options
[   77s] +INFO: object loc-d1 cannot be represented in the CLI notation
[   77s] +INFO: object loc-d1 cannot be represented in the CLI notation
[   77s]  .TRY configure show rsc_options
[   77s] +INFO: object loc-d1 cannot be represented in the CLI notation
[   77s]  rsc_defaults $id="rsc_options" \
[   77s]  .TRY configure property stonith-enabled=true
[   77s] +INFO: object loc-d1 cannot be represented in the CLI notation
[   77s] +INFO: object loc-d1 cannot be represented in the CLI notation
[   77s]  .TRY configure show cib-bootstrap-options
[   77s] +INFO: object loc-d1 cannot be represented in the CLI notation
[   77s]  property $id="cib-bootstrap-options" \
[   77s]  .TRY configure filter 'sed "s/stonith-enabled=.true.//"'
[   77s] +INFO: object loc-d1 cannot be represented in the CLI notation
[   77s] +ERROR: 13: syntax: Unknown command near <xml> parsing 'xml <rsc_location id=loc-d1 rsc=d1> <!--# --> <rule id=r1 score=-INFINITY boolean-op=or> <expression operation=not_defined attribute=webserver id=loc-d1-expression/> <expression attribute=mem type=number operation=lte value=0 id=loc-d1-expression-3/> </rule> <rule id=loc-d1-rule score=-INFINITY> <expression operation=not_defined attribute=a2 id=loc-d1-expression-2/> </rule> <rule id=r2 score-attribute=webserver> <expression operation=defined attribute=webserver id=loc-d1-expression-0/> </rule> <!--# --> </rsc_location>'
[   77s]  .TRY configure show cib-bootstrap-options
[   77s] +INFO: object loc-d1 cannot be represented in the CLI notation
[   77s]  property $id="cib-bootstrap-options" \
[   77s] -    default-action-timeout="60s"
[   77s] +    default-action-timeout="60s" \
[   77s] +    stonith-enabled="true"
[   77s] ----------
[   77s] Fri Jan 10 12:38:53 UTC 2014: END testcase edit
--
// Kristoffer Grönlund
_______________________________________________________
http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
Home Page: http://linux-ha.org/
r***@ybb.ne.jp
2014-01-14 03:31:29 UTC
Permalink
Hi Kristoffer,

In addition, the error did not happen on the edit test.
I passed the test of edit even if I did not apply my patch even if I applied a patch.

Did you execute the command of what kind of test?

* on crmsh-7cd5688c164d.tar(tip)
(snip)
[***@rh64-2744 test]# ./regression.sh
confbasic. checking... PASS
confbasic-xml. checking... PASS
edit........ checking... PASS
(snip)

* on crmsh-ef3f08547688(1.2.5)
(snip)
[***@rh64-2744 test]# ./regression.sh
confbasic. checking... PASS
confbasic-xml. checking... FAIL
edit. checking... PASS
(snip)

Best Regards,
Hideo Yamauchi.
Post by r***@ybb.ne.jp
Hi Krstoffer,
As for the problem, contents of rpm of crmsh1.2.5 which we used seem to have a problem somehow or other.
The problem did not occur in crm which I made from a source code of crmsh1.2.5.
The application of the patch which I donated seems to be unnecessary.
I confirm the details and contact me again.
Best Regards,
Hideo Yamauchi.
Post by r***@ybb.ne.jp
Hi Kristoffer,
Thank you for comment.
I will look the day after tomorrow.
Best Regards,
Hideo Yamauchi.
Post by Kristoffer Grönlund
On Fri, 10 Jan 2014 16:27:47 +0900 (JST)
Post by r***@ybb.ne.jp
Hi Dejan,
I send a patch of crmsh1.2.5.
Similar correction seems to be necessary for latest crmsh.
Best Regards,
Hideo Yamauchi
Hello,
Thank you for the patch!
I tried applying the patch to the latest crmsh, but when running the
regression test suite, I got an error. I think the patch is fixing a
bug, but unfortunately it seems to reveal a different problem.
Maybe you can help me figure out what is going wrong!
[   77s] Fri Jan 10 12:38:47 UTC 2014: BEGIN testcase edit
[   77s] ----------
[   77s] testcase edit failed
[   77s] output is in crmtestout/edit.out
[   77s] --- /usr/share/crmsh/tests/testcases/edit.exp    2014-01-10 12:38:36.000000000 +0000
[   77s] +++ -    2014-01-10 12:38:53.149599264 +0000
[   77s]  .TRY configure rsc_defaults $id="rsc_options" failure-timeout=10m
[   77s] +INFO: object loc-d1 cannot be represented in the CLI notation
[   77s]  .TRY configure filter "sed 's/2m/60s/'" cib-bootstrap-options
[   77s] +INFO: object loc-d1 cannot be represented in the CLI notation
[   77s] +INFO: object loc-d1 cannot be represented in the CLI notation
[   77s]  .TRY configure show rsc_options
[   77s] +INFO: object loc-d1 cannot be represented in the CLI notation
[   77s]  rsc_defaults $id="rsc_options" \
[   77s]  .TRY configure property stonith-enabled=true
[   77s] +INFO: object loc-d1 cannot be represented in the CLI notation
[   77s] +INFO: object loc-d1 cannot be represented in the CLI notation
[   77s]  .TRY configure show cib-bootstrap-options
[   77s] +INFO: object loc-d1 cannot be represented in the CLI notation
[   77s]  property $id="cib-bootstrap-options" \
[   77s]  .TRY configure filter 'sed "s/stonith-enabled=.true.//"'
[   77s] +INFO: object loc-d1 cannot be represented in the CLI notation
[   77s] +ERROR: 13: syntax: Unknown command near <xml> parsing 'xml <rsc_location id=loc-d1 rsc=d1> <!--# --> <rule id=r1 score=-INFINITY boolean-op=or> <expression operation=not_defined attribute=webserver id=loc-d1-expression/> <expression attribute=mem type=number operation=lte value=0 id=loc-d1-expression-3/> </rule> <rule id=loc-d1-rule score=-INFINITY> <expression operation=not_defined attribute=a2 id=loc-d1-expression-2/> </rule> <rule id=r2 score-attribute=webserver> <expression operation=defined attribute=webserver id=loc-d1-expression-0/> </rule> <!--# --> </rsc_location>'
[   77s]  .TRY configure show cib-bootstrap-options
[   77s] +INFO: object loc-d1 cannot be represented in the CLI notation
[   77s]  property $id="cib-bootstrap-options" \
[   77s] -    default-action-timeout="60s"
[   77s] +    default-action-timeout="60s" \
[   77s] +    stonith-enabled="true"
[   77s] ----------
[   77s] Fri Jan 10 12:38:53 UTC 2014: END testcase edit
--
// Kristoffer Grönlund
_______________________________________________________
http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
Home Page: http://linux-ha.org/
_______________________________________________________
http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
Home Page: http://linux-ha.org/
Kristoffer Grönlund
2014-01-15 14:14:06 UTC
Permalink
On Tue, 14 Jan 2014 12:31:29 +0900 (JST)
Post by r***@ybb.ne.jp
Hi Kristoffer,
In addition, the error did not happen on the edit test.
I passed the test of edit even if I did not apply my patch even if I applied a patch.
Did you execute the command of what kind of test?
Sorry, I should have mentioned that I applied the patch to the
development version, not to 1.2.5, when testing.

I suspect that the difference is that in older versions, comments were
stripped completely from the configuration, but in newer versions,
comments are kept. However, it seems that with this patch there are
comments generated in the XML code that the CLI syntax cannot
represent.

I have not had time to completely investigate. I will look into the
problem further and let you know what I find.

Thank you,
Post by r***@ybb.ne.jp
* on crmsh-7cd5688c164d.tar(tip)
(snip)
confbasic. checking... PASS
confbasic-xml. checking... PASS
edit........ checking... PASS
(snip)
* on crmsh-ef3f08547688(1.2.5)
(snip)
confbasic. checking... PASS
confbasic-xml. checking... FAIL
edit. checking... PASS
(snip)
Best Regards,
Hideo Yamauchi.
--
// Kristoffer Grönlund
// ***@suse.com
r***@ybb.ne.jp
2014-01-16 01:19:00 UTC
Permalink
Hi Kristoffer,
Post by Kristoffer Grönlund
Sorry, I should have mentioned that I applied the patch to the
development version, not to 1.2.5, when testing.
I suspect that the difference is that in older versions, comments were
stripped completely from the configuration, but in newer versions,
comments are kept. However, it seems that with this patch there are
comments generated in the XML code that the CLI syntax cannot
represent.
I have not had time to completely investigate. I will look into the
problem further and let you know what I find.
It was recognized that the patch which I donated was unnecessary after all.
rpm which we used somehow or other seemed to have a problem.

I withdraw a patch.

Best Regards,
Hideo Yamauchi.
Post by Kristoffer Grönlund
On Tue, 14 Jan 2014 12:31:29 +0900 (JST)
Post by r***@ybb.ne.jp
Hi Kristoffer,
In addition, the error did not happen on the edit test.
I passed the test of edit even if I did not apply my patch even if I
applied a patch.
Did you execute the command of what kind of test?
Sorry, I should have mentioned that I applied the patch to the
development version, not to 1.2.5, when testing.
I suspect that the difference is that in older versions, comments were
stripped completely from the configuration, but in newer versions,
comments are kept. However, it seems that with this patch there are
comments generated in the XML code that the CLI syntax cannot
represent.
I have not had time to completely investigate. I will look into the
problem further and let you know what I find.
Thank you,
Post by r***@ybb.ne.jp
  * on crmsh-7cd5688c164d.tar(tip)
(snip)
confbasic. checking... PASS
confbasic-xml. checking... PASS
edit........ checking... PASS
(snip)
  * on crmsh-ef3f08547688(1.2.5)
(snip)
confbasic. checking... PASS
confbasic-xml. checking... FAIL
edit. checking... PASS
(snip)
Best Regards,
Hideo Yamauchi.
--
// Kristoffer Grönlund
Loading...