5:52 p.m.
This is a multi-part message in MIME format.
--------------040303020501050109020008
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Hello,
I'm starting to know the engine code. I chose a little unstandardized
behaviour to follow through the devel process. I have a patch and I'd
like to know if you fell relevant to correct this issue:
- Description: Adding a LOCAL storage [1], webadmin does not validate
path against regex, sendind the invalid path (with final slash) to vdsm
[2] [3]. But, adding a NFS storage, the path is validated before
contacting vdsm [4] avoiding extra vdsm processing and quickly/clearly
informing user about what's wrong.
- Expected result: Same behaviour to NFS and LOCALFS storage path
validation. Validate LOCALFS path in webadmin before send it to vdsm [5].
- Newbie doubt: Wouldn't be better to validate the both local and nfs
path on the backend, achieving all user interfaces/APIs?
[1] -
https://picasaweb.google.com/lh/photo/FWNiou2Y12GZO3AjfCH6K7QAv8cs6edaj3f...
[2] -
https://picasaweb.google.com/lh/photo/Pof6Z8ohgQAkRTDpEJKG-LQAv8cs6edaj3f...
[3] - https://gist.github.com/2762656
[4] -
https://picasaweb.google.com/lh/photo/Fd3zWegWE0T5C2tDo_tPZrQAv8cs6edaj3f...
[5] -
https://picasaweb.google.com/lh/photo/PgzYrZHkkvm-WtFk_UFZLrQAv8cs6edaj3f...
I look forward to hearing your comments.
Best Regards,
--
Pahim
--------------040303020501050109020008
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
<html>
<head>
<meta http-equiv="content-type" content="text/html;
charset=ISO-8859-1">
</head>
<body bgcolor="#FFFFFF" text="#000000">
Hello,<br>
<br>
I'm starting to know the engine code. I chose a little
unstandardized behaviour to follow through the devel process. I have
a patch and I'd like to know if you fell relevant to correct this
issue:<br>
<br>
- Description: Adding a LOCAL storage [1], webadmin does not
validate path against regex, sendind the invalid path (with final
slash) to vdsm [2] [3]. But, adding a NFS storage, the path is
validated before contacting vdsm [4] avoiding extra vdsm processing
and quickly/clearly informing user about what's wrong.<br>
<br>
- Expected result: Same behaviour to NFS and LOCALFS storage path
validation. Validate LOCALFS path in webadmin before send it to vdsm
[5].<br>
<br>
- Newbie doubt: Wouldn't be better to validate the both local and
nfs path on the backend, achieving all user interfaces/APIs?<br>
<br>
[1] -
<a class="moz-txt-link-freetext"
href="https://picasaweb.google.com/lh/photo/FWNiou2Y12GZO3AjfCH6K7QA...
[2] -
<a class="moz-txt-link-freetext"
href="https://picasaweb.google.com/lh/photo/Pof6Z8ohgQAkRTDpEJKG-LQA...
[3] - <a class="moz-txt-link-freetext"
href="https://gist.github.com/2762656">https://gist.github.c...
[4] -
<a class="moz-txt-link-freetext"
href="https://picasaweb.google.com/lh/photo/Fd3zWegWE0T5C2tDo_tPZrQA...
[5] -
<a class="moz-txt-link-freetext"
href="https://picasaweb.google.com/lh/photo/PgzYrZHkkvm-WtFk_UFZLrQA...
<br>
<meta http-equiv="content-type" content="text/html;
charset=ISO-8859-1">
I look forward to hearing your comments.<br>
<br>
Best Regards,<br>
--<br>
Pahim<br>
</body>
</html>
--------------040303020501050109020008--
9:37 p.m.
Hi Pahim,
Thanks for the input!
Comments inline.
Thanks,
Gilad.
----- Original Message -----
From: "Amador pahim" <apahim(a)redhat.com>
To: engine-devel(a)ovirt.org
Sent: Monday, May 21, 2012 5:52:34 PM
Subject: [Engine-devel] LOCALFS path validation
Hello,
I'm starting to know the engine code. I chose a little unstandardized
behaviour to follow through the devel process. I have a patch and
I'd like to know if you fell relevant to correct this issue:
- Description: Adding a LOCAL storage [1], webadmin does not validate
path against regex, sendind the invalid path (with final slash) to
vdsm [2] [3]. But, adding a NFS storage, the path is validated
before contacting vdsm [4] avoiding extra vdsm processing and
quickly/clearly informing user about what's wrong.
- Expected result: Same behaviour to NFS and LOCALFS storage path
validation. Validate LOCALFS path in webadmin before send it to vdsm
[5].
you may and should send a patch :)
- Newbie doubt: Wouldn't be better to validate the both local and nfs
path on the backend, achieving all user interfaces/APIs?
Because we have a rich client app (gwt), we can perform the validation also in the client
side very easily,
we do that to avoid unnecessary calls to the backend side, and to have a better &
responsive ui
(client side validation is performed instantly - without the need to wait).
Anyway, every validation performed in the client side needs to be performed also in
backend side (for api, and other reasons (security?)).
[1] -
https://picasaweb.google.com/lh/photo/FWNiou2Y12GZO3AjfCH6K7QAv8cs6edaj3f...
[2] -
https://picasaweb.google.com/lh/photo/Pof6Z8ohgQAkRTDpEJKG-LQAv8cs6edaj3f...
[3] - https://gist.github.com/2762656
[4] -
https://picasaweb.google.com/lh/photo/Fd3zWegWE0T5C2tDo_tPZrQAv8cs6edaj3f...
[5] -
https://picasaweb.google.com/lh/photo/PgzYrZHkkvm-WtFk_UFZLrQAv8cs6edaj3f...
I look forward to hearing your comments.
Best Regards,
--
Pahim
_______________________________________________
Engine-devel mailing list
Engine-devel(a)ovirt.org
http://lists.ovirt.org/mailman/listinfo/engine-devel
6:39 p.m.
On 05/21/2012 09:37 PM, Gilad Chaplik wrote:
Hi Pahim,
Thanks for the input!
Comments inline.
Thanks,
Gilad.
----- Original Message -----
> From: "Amador pahim"<apahim(a)redhat.com>
> To: engine-devel(a)ovirt.org
> Sent: Monday, May 21, 2012 5:52:34 PM
> Subject: [Engine-devel] LOCALFS path validation
>
>
> Hello,
>
> I'm starting to know the engine code. I chose a little unstandardized
> behaviour to follow through the devel process. I have a patch and
> I'd like to know if you fell relevant to correct this issue:
>
> - Description: Adding a LOCAL storage [1], webadmin does not validate
> path against regex, sendind the invalid path (with final slash) to
> vdsm [2] [3]. But, adding a NFS storage, the path is validated
> before contacting vdsm [4] avoiding extra vdsm processing and
> quickly/clearly informing user about what's wrong.
>
> - Expected result: Same behaviour to NFS and LOCALFS storage path
> validation. Validate LOCALFS path in webadmin before send it to vdsm
> [5].
you may and should send a patch :)
>
> - Newbie doubt: Wouldn't be better to validate the both local and nfs
> path on the backend, achieving all user interfaces/APIs?
Because we have a rich client app (gwt), we can perform the validation also in the client
side very easily,
we do that to avoid unnecessary calls to the backend side, and to have a better&
responsive ui
(client side validation is performed instantly - without the need to wait).
Anyway, every validation performed in the client side needs to be performed also in
backend side (for api, and other reasons (security?)).
the client side can validate the format of the path, not if it really
exists.
do you mean local storage via the wizard, or just adding another storage
domain?
the 'configure local storage' wizard requires the host to be in
maintenance, which may or may not be a problem to implement this type of
validation.
8:16 p.m.
On 06/01/2012 12:39 PM, Itamar Heim wrote:
On 05/21/2012 09:37 PM, Gilad Chaplik wrote:
> Hi Pahim,
>
> Thanks for the input!
> Comments inline.
>
> Thanks,
> Gilad.
>
> ----- Original Message -----
>> From: "Amador pahim"<apahim(a)redhat.com>
>> To: engine-devel(a)ovirt.org
>> Sent: Monday, May 21, 2012 5:52:34 PM
>> Subject: [Engine-devel] LOCALFS path validation
>>
>>
>> Hello,
>>
>> I'm starting to know the engine code. I chose a little unstandardized
>> behaviour to follow through the devel process. I have a patch and
>> I'd like to know if you fell relevant to correct this issue:
>>
>> - Description: Adding a LOCAL storage [1], webadmin does not validate
>> path against regex, sendind the invalid path (with final slash) to
>> vdsm [2] [3]. But, adding a NFS storage, the path is validated
>> before contacting vdsm [4] avoiding extra vdsm processing and
>> quickly/clearly informing user about what's wrong.
>>
>> - Expected result: Same behaviour to NFS and LOCALFS storage path
>> validation. Validate LOCALFS path in webadmin before send it to vdsm
>> [5].
>
> you may and should send a patch :)
>
>>
>> - Newbie doubt: Wouldn't be better to validate the both local and nfs
>> path on the backend, achieving all user interfaces/APIs?
>
> Because we have a rich client app (gwt), we can perform the
> validation also in the client side very easily,
> we do that to avoid unnecessary calls to the backend side, and to
> have a better& responsive ui
> (client side validation is performed instantly - without the need to
> wait).
>
> Anyway, every validation performed in the client side needs to be
> performed also in backend side (for api, and other reasons (security?)).
the client side can validate the format of the path, not if it really
exists.
do you mean local storage via the wizard, or just adding another
storage domain?
the 'configure local storage' wizard requires the host to be in
maintenance, which may or may not be a problem to implement this type
of validation.
I mean just adding a LOCALFS storage domain with an invalid path, like this:
https://picasaweb.google.com/lh/photo/FWNiou2Y12GZO3AjfCH6K7QAv8cs6edaj3f...
Currently we have this:
https://picasaweb.google.com/lh/photo/Pof6Z8ohgQAkRTDpEJKG-LQAv8cs6edaj3f...
I sent a patch...
http://gerrit.ovirt.org/4857
... in order to do this:
https://picasaweb.google.com/lh/photo/PgzYrZHkkvm-WtFk_UFZLrQAv8cs6edaj3f...
Just like "New NFS Domain" currently does:
https://picasaweb.google.com/lh/photo/Fd3zWegWE0T5C2tDo_tPZrQAv8cs6edaj3f...
11:04 a.m.
On 06/01/2012 08:16 PM, Amador Pahim wrote:
On 06/01/2012 12:39 PM, Itamar Heim wrote:
> On 05/21/2012 09:37 PM, Gilad Chaplik wrote:
>> Hi Pahim,
>>
>> Thanks for the input!
>> Comments inline.
>>
>> Thanks,
>> Gilad.
>>
>> ----- Original Message -----
>>> From: "Amador pahim"<apahim(a)redhat.com>
>>> To: engine-devel(a)ovirt.org
>>> Sent: Monday, May 21, 2012 5:52:34 PM
>>> Subject: [Engine-devel] LOCALFS path validation
>>>
>>>
>>> Hello,
>>>
>>> I'm starting to know the engine code. I chose a little unstandardized
>>> behaviour to follow through the devel process. I have a patch and
>>> I'd like to know if you fell relevant to correct this issue:
>>>
>>> - Description: Adding a LOCAL storage [1], webadmin does not validate
>>> path against regex, sendind the invalid path (with final slash) to
>>> vdsm [2] [3]. But, adding a NFS storage, the path is validated
>>> before contacting vdsm [4] avoiding extra vdsm processing and
>>> quickly/clearly informing user about what's wrong.
>>>
>>> - Expected result: Same behaviour to NFS and LOCALFS storage path
>>> validation. Validate LOCALFS path in webadmin before send it to vdsm
>>> [5].
>>
>> you may and should send a patch :)
>>
>>>
>>> - Newbie doubt: Wouldn't be better to validate the both local and nfs
>>> path on the backend, achieving all user interfaces/APIs?
>>
>> Because we have a rich client app (gwt), we can perform the
>> validation also in the client side very easily,
>> we do that to avoid unnecessary calls to the backend side, and to
>> have a better& responsive ui
>> (client side validation is performed instantly - without the need to
>> wait).
>>
>> Anyway, every validation performed in the client side needs to be
>> performed also in backend side (for api, and other reasons (security?)).
>
> the client side can validate the format of the path, not if it really
> exists.
> do you mean local storage via the wizard, or just adding another
> storage domain?
> the 'configure local storage' wizard requires the host to be in
> maintenance, which may or may not be a problem to implement this type
> of validation.
I mean just adding a LOCALFS storage domain with an invalid path, like
this:
https://picasaweb.google.com/lh/photo/FWNiou2Y12GZO3AjfCH6K7QAv8cs6edaj3f...
Currently we have this:
https://picasaweb.google.com/lh/photo/Pof6Z8ohgQAkRTDpEJKG-LQAv8cs6edaj3f...
I sent a patch...
http://gerrit.ovirt.org/4857
... in order to do this:
https://picasaweb.google.com/lh/photo/PgzYrZHkkvm-WtFk_UFZLrQAv8cs6edaj3f...
Just like "New NFS Domain" currently does:
https://picasaweb.google.com/lh/photo/Fd3zWegWE0T5C2tDo_tPZrQAv8cs6edaj3f...
ok - sounds basic enough of a change. I'd just note in the patch you are
re-using same REGEX used for the previous validation happening in the
backend (and actually try to re-use it, to not maintain a change in too
many places0
2:50 p.m.
This is a multi-part message in MIME format.
--------------010009060709070503080105
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
On 06/03/2012 05:04 AM, Itamar Heim wrote:
On 06/01/2012 08:16 PM, Amador Pahim wrote:
> On 06/01/2012 12:39 PM, Itamar Heim wrote:
>> On 05/21/2012 09:37 PM, Gilad Chaplik wrote:
>>> Hi Pahim,
>>>
>>> Thanks for the input!
>>> Comments inline.
>>>
>>> Thanks,
>>> Gilad.
>>>
>>> ----- Original Message -----
>>>> From: "Amador pahim"<apahim(a)redhat.com>
>>>> To: engine-devel(a)ovirt.org
>>>> Sent: Monday, May 21, 2012 5:52:34 PM
>>>> Subject: [Engine-devel] LOCALFS path validation
>>>>
>>>>
>>>> Hello,
>>>>
>>>> I'm starting to know the engine code. I chose a little
unstandardized
>>>> behaviour to follow through the devel process. I have a patch and
>>>> I'd like to know if you fell relevant to correct this issue:
>>>>
>>>> - Description: Adding a LOCAL storage [1], webadmin does not validate
>>>> path against regex, sendind the invalid path (with final slash) to
>>>> vdsm [2] [3]. But, adding a NFS storage, the path is validated
>>>> before contacting vdsm [4] avoiding extra vdsm processing and
>>>> quickly/clearly informing user about what's wrong.
>>>>
>>>> - Expected result: Same behaviour to NFS and LOCALFS storage path
>>>> validation. Validate LOCALFS path in webadmin before send it to vdsm
>>>> [5].
>>>
>>> you may and should send a patch :)
>>>
>>>>
>>>> - Newbie doubt: Wouldn't be better to validate the both local and
nfs
>>>> path on the backend, achieving all user interfaces/APIs?
>>>
>>> Because we have a rich client app (gwt), we can perform the
>>> validation also in the client side very easily,
>>> we do that to avoid unnecessary calls to the backend side, and to
>>> have a better& responsive ui
>>> (client side validation is performed instantly - without the need to
>>> wait).
>>>
>>> Anyway, every validation performed in the client side needs to be
>>> performed also in backend side (for api, and other reasons
>>> (security?)).
>>
>> the client side can validate the format of the path, not if it really
>> exists.
>> do you mean local storage via the wizard, or just adding another
>> storage domain?
>> the 'configure local storage' wizard requires the host to be in
>> maintenance, which may or may not be a problem to implement this type
>> of validation.
>
> I mean just adding a LOCALFS storage domain with an invalid path, like
> this:
> https://picasaweb.google.com/lh/photo/FWNiou2Y12GZO3AjfCH6K7QAv8cs6edaj3f...
>
>
> Currently we have this:
> https://picasaweb.google.com/lh/photo/Pof6Z8ohgQAkRTDpEJKG-LQAv8cs6edaj3f...
>
>
> I sent a patch...
> http://gerrit.ovirt.org/4857
> ... in order to do this:
> https://picasaweb.google.com/lh/photo/PgzYrZHkkvm-WtFk_UFZLrQAv8cs6edaj3f...
>
>
> Just like "New NFS Domain" currently does:
> https://picasaweb.google.com/lh/photo/Fd3zWegWE0T5C2tDo_tPZrQAv8cs6edaj3f...
>
>
ok - sounds basic enough of a change. I'd just note in the patch you
are re-using same REGEX used for the previous validation happening in
the backend (and actually try to re-use it, to not maintain a change
in too many places0
There is no validation in backend. I mean, the error
"Error while
executing action AddLocalStorageDomain: Remote path is illegal" is
coming from VDSM:
https://gist.github.com/2762656
And is not caused by regex. VDSM is just comparing path /test/ with
os.path.abspath(), /test:
https://gist.github.com/2867874
--------------010009060709070503080105
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
<html>
<head>
<meta content="text/html; charset=ISO-8859-1"
http-equiv="Content-Type">
</head>
<body bgcolor="#FFFFFF" text="#000000">
On 06/03/2012 05:04 AM, Itamar Heim wrote:
<blockquote cite="mid:4FCB1A94.8080201@redhat.com"
type="cite">On
06/01/2012 08:16 PM, Amador Pahim wrote:
<br>
<blockquote type="cite">On 06/01/2012 12:39 PM, Itamar Heim wrote:
<br>
<blockquote type="cite">On 05/21/2012 09:37 PM, Gilad Chaplik
wrote:
<br>
<blockquote type="cite">Hi Pahim,
<br>
<br>
Thanks for the input!
<br>
Comments inline.
<br>
<br>
Thanks,
<br>
Gilad.
<br>
<br>
----- Original Message -----
<br>
<blockquote type="cite">From: "Amador
pahim"<a class="moz-txt-link-rfc2396E"
href="mailto:apahim@redhat.com"><apahim@redhat.com></a>
<br>
To: <a class="moz-txt-link-abbreviated"
href="mailto:engine-devel@ovirt.org">engine-devel@ovirt.org</a>
<br>
Sent: Monday, May 21, 2012 5:52:34 PM
<br>
Subject: [Engine-devel] LOCALFS path validation
<br>
<br>
<br>
Hello,
<br>
<br>
I'm starting to know the engine code. I chose a little
unstandardized
<br>
behaviour to follow through the devel process. I have a
patch and
<br>
I'd like to know if you fell relevant to correct this
issue:
<br>
<br>
- Description: Adding a LOCAL storage [1], webadmin does
not validate
<br>
path against regex, sendind the invalid path (with final
slash) to
<br>
vdsm [2] [3]. But, adding a NFS storage, the path is
validated
<br>
before contacting vdsm [4] avoiding extra vdsm processing
and
<br>
quickly/clearly informing user about what's wrong.
<br>
<br>
- Expected result: Same behaviour to NFS and LOCALFS
storage path
<br>
validation. Validate LOCALFS path in webadmin before send
it to vdsm
<br>
[5].
<br>
</blockquote>
<br>
you may and should send a patch :)
<br>
<br>
<blockquote type="cite">
<br>
- Newbie doubt: Wouldn't be better to validate the both
local and nfs
<br>
path on the backend, achieving all user interfaces/APIs?
<br>
</blockquote>
<br>
Because we have a rich client app (gwt), we can perform the
<br>
validation also in the client side very easily,
<br>
we do that to avoid unnecessary calls to the backend side,
and to
<br>
have a better& responsive ui
<br>
(client side validation is performed instantly - without the
need to
<br>
wait).
<br>
<br>
Anyway, every validation performed in the client side needs
to be
<br>
performed also in backend side (for api, and other reasons
(security?)).
<br>
</blockquote>
<br>
the client side can validate the format of the path, not if it
really
<br>
exists.
<br>
do you mean local storage via the wizard, or just adding
another
<br>
storage domain?
<br>
the 'configure local storage' wizard requires the host to be
in
<br>
maintenance, which may or may not be a problem to implement
this type
<br>
of validation.
<br>
</blockquote>
<br>
I mean just adding a LOCALFS storage domain with an invalid
path, like
<br>
this:
<br>
<a class="moz-txt-link-freetext"
href="https://picasaweb.google.com/lh/photo/FWNiou2Y12GZO3AjfCH6K7QA...
<br>
<br>
Currently we have this:
<br>
<a class="moz-txt-link-freetext"
href="https://picasaweb.google.com/lh/photo/Pof6Z8ohgQAkRTDpEJKG-LQA...
<br>
<br>
I sent a patch...
<br>
<a class="moz-txt-link-freetext"
href="http://gerrit.ovirt.org/4857">http://gerrit.ovirt.org/...
<br>
... in order to do this:
<br>
<a class="moz-txt-link-freetext"
href="https://picasaweb.google.com/lh/photo/PgzYrZHkkvm-WtFk_UFZLrQA...
<br>
<br>
Just like "New NFS Domain" currently does:
<br>
<a class="moz-txt-link-freetext"
href="https://picasaweb.google.com/lh/photo/Fd3zWegWE0T5C2tDo_tPZrQA...
<br>
<br>
</blockquote>
<br>
ok - sounds basic enough of a change. I'd just note in the patch
you are re-using same REGEX used for the previous validation
happening in the backend (and actually try to re-use it, to not
maintain a change in too many places0
<br>
</blockquote>
There is no validation in backend. I mean, the error "Error while
executing action AddLocalStorageDomain: Remote path is illegal" is
coming from VDSM:<br>
<meta http-equiv="content-type" content="text/html;
charset=ISO-8859-1">
<a
href="https://gist.github.com/2762656">https://gist.github.c...
And is not caused by regex. VDSM is just comparing path /test/ with
os.path.abspath(), /test:<br>
<meta http-equiv="content-type" content="text/html;
charset=ISO-8859-1">
<a
href="https://gist.github.com/2867874">https://gist.github.c...
</body>
</html>
--------------010009060709070503080105--
3:08 p.m.
On 06/04/2012 02:50 PM, Amador Pahim wrote:
On 06/03/2012 05:04 AM, Itamar Heim wrote:
> On 06/01/2012 08:16 PM, Amador Pahim wrote:
>> On 06/01/2012 12:39 PM, Itamar Heim wrote:
>>> On 05/21/2012 09:37 PM, Gilad Chaplik wrote:
>>>> Hi Pahim,
>>>>
>>>> Thanks for the input!
>>>> Comments inline.
>>>>
>>>> Thanks,
>>>> Gilad.
>>>>
>>>> ----- Original Message -----
>>>>> From: "Amador pahim"<apahim(a)redhat.com>
>>>>> To: engine-devel(a)ovirt.org
>>>>> Sent: Monday, May 21, 2012 5:52:34 PM
>>>>> Subject: [Engine-devel] LOCALFS path validation
>>>>>
>>>>>
>>>>> Hello,
>>>>>
>>>>> I'm starting to know the engine code. I chose a little
unstandardized
>>>>> behaviour to follow through the devel process. I have a patch and
>>>>> I'd like to know if you fell relevant to correct this issue:
>>>>>
>>>>> - Description: Adding a LOCAL storage [1], webadmin does not
validate
>>>>> path against regex, sendind the invalid path (with final slash) to
>>>>> vdsm [2] [3]. But, adding a NFS storage, the path is validated
>>>>> before contacting vdsm [4] avoiding extra vdsm processing and
>>>>> quickly/clearly informing user about what's wrong.
>>>>>
>>>>> - Expected result: Same behaviour to NFS and LOCALFS storage path
>>>>> validation. Validate LOCALFS path in webadmin before send it to vdsm
>>>>> [5].
>>>>
>>>> you may and should send a patch :)
>>>>
>>>>>
>>>>> - Newbie doubt: Wouldn't be better to validate the both local and
nfs
>>>>> path on the backend, achieving all user interfaces/APIs?
>>>>
>>>> Because we have a rich client app (gwt), we can perform the
>>>> validation also in the client side very easily,
>>>> we do that to avoid unnecessary calls to the backend side, and to
>>>> have a better& responsive ui
>>>> (client side validation is performed instantly - without the need to
>>>> wait).
>>>>
>>>> Anyway, every validation performed in the client side needs to be
>>>> performed also in backend side (for api, and other reasons
>>>> (security?)).
>>>
>>> the client side can validate the format of the path, not if it really
>>> exists.
>>> do you mean local storage via the wizard, or just adding another
>>> storage domain?
>>> the 'configure local storage' wizard requires the host to be in
>>> maintenance, which may or may not be a problem to implement this type
>>> of validation.
>>
>> I mean just adding a LOCALFS storage domain with an invalid path, like
>> this:
>>
https://picasaweb.google.com/lh/photo/FWNiou2Y12GZO3AjfCH6K7QAv8cs6edaj3f...
>>
>>
>> Currently we have this:
>>
https://picasaweb.google.com/lh/photo/Pof6Z8ohgQAkRTDpEJKG-LQAv8cs6edaj3f...
>>
>>
>> I sent a patch...
>> http://gerrit.ovirt.org/4857
>> ... in order to do this:
>>
https://picasaweb.google.com/lh/photo/PgzYrZHkkvm-WtFk_UFZLrQAv8cs6edaj3f...
>>
>>
>> Just like "New NFS Domain" currently does:
>>
https://picasaweb.google.com/lh/photo/Fd3zWegWE0T5C2tDo_tPZrQAv8cs6edaj3f...
>>
>>
>
> ok - sounds basic enough of a change. I'd just note in the patch you
> are re-using same REGEX used for the previous validation happening in
> the backend (and actually try to re-use it, to not maintain a change
> in too many places0
There is no validation in backend. I mean, the error "Error while
executing action AddLocalStorageDomain: Remote path is illegal" is
coming from VDSM:
https://gist.github.com/2762656
And is not caused by regex. VDSM is just comparing path /test/ with
os.path.abspath(), /test:
https://gist.github.com/2867874
sounds like a valid validation to add then
4553
days inactive
4567
days old
6 comments
4 participants
participants (4)
-
Amador Pahim -
Amador pahim
-
Gilad Chaplik -
Itamar Heim