0005534: Testlink web interface stops working - Size of a request header field exceeds server limit.

(0018351)
fman (administrator)
2013-02-19 20:16

1. additional info:
there is no indication about number of items you were trying to manage, may be there a limit exists

2. in any case you need to upgrade to 1.9.5 and retest

(0018358)
Lukas Mencl (reporter)
2013-02-20 08:36

Ad 1.
I am not sure what exactly you mean but there are some numbers:
- We have got 8 testlink test projects which have 5635 test cases in total.
- I was working (=testing regarding test plan) in test project which has 3862 test cases in total on test plan with 34 test cases when the issue appeared.
- I was not modifying any steps, I simply wanted to move to next test case when the issue appeared.

Ad 2.
We will try to install 1.9.5 and retest when we find time for it. I will let you know about the result.

(0018365)
moormanm (reporter)
2013-02-23 16:28
edited on: 2013-02-23 17:54

It appears that the large header data stems from the save/restore tree state logic in tcTree.js. Is there a more efficient way to store the collapse/expand tree state?

EDIT: to reproduce this behavior, click 'expand all' on a test case tree panel that has a large number of nested testcases and then click on one of them. Examine the resultant header data in the user request.

(0018366)
fman (administrator)
2013-02-23 17:58

>> Is there a more efficient way to store the collapse/expand tree state?
do not know.
we have taken example from EXT-JS version 3.x years ago.
Some research on ext-js site is needed,
right now think only workaround will be trying to increase cookie size if possible

(0018367)
fman (administrator)
2013-02-23 18:15

The silly workaround to apply is to clean up cookies on browser.
Do not think that loosing tree status (meaning what was opened on my last navigation) will be a great drawback

(0018375)
Julian (reporter)
2013-02-27 07:17

Duplicate of 0003714

(0021513)
elizk (reporter)
2014-08-27 13:23

we just faced this issue leading to server crashing - increasing the cookie size is a poor workaround.

(0021523)
fman (administrator)
2014-08-31 17:24

>> we just faced this issue leading to server crashing -
>> increasing the cookie size is a poor workaround.
Ready to receive any kind of implementation that will solve the issue.

(0022717)
modir (reporter)
2015-02-16 09:35

The problem still exists in 1.9.13. (We have about 7000 test cases.)

I can not give the code right now but the solutions would be like this:

Instead of saving the data in the cookie you would save it in a javascript variable. The variable could then be sent as a JSON object with a POST request to the server where it can be saved for later use. When a new page is loaded the values can be fetched again as a JSON object.

I could help and provide some code but I would need more background information on why this is saved at all, etc.

(0022718)
fman (administrator)
2015-02-16 09:52

Ideas are great, working solutions are better.

>> I could help and provide some code but I would need more background
>> information on why this is saved at all
hmm may be to restore tree status ?

(0022743)
modir (reporter)
2015-02-19 10:42

> hmm may be to restore tree status ?

I need more information. Do you do anything on the server side with this information? Or is it only used client side?

(0022746)
fman (administrator)
2015-02-19 10:51

>> I need more information. Do you do anything on the server side with this
>> information?
What do you mean?
how this can change your approach?
You need to read the code, and understand this, I can not explain any step done
on code on several notes.

(0023462)
modir (reporter)
2015-06-10 12:32

> What do you mean?

Does any PHP code do anything with those information in the cookie?

> how this can change your approach?

If the PHP code on the server is doing nothing with the cookie information then I am a lot more flexible on choosing the best solution on the client side. And I don't have to send any information to the server.

> You need to read the code, and understand this, I can not explain any step done on code on several notes.

I hoped that there would be some code documentation like e.g. PHPdoc that would help me to learn the code faster. Since there is nothing like this I ask you to help me getting faster a understanding of the code. At least of the code which I need to solve this problem. I will read through the code in this case.

(0023463)
fman (administrator)
2015-06-10 13:50

cookie is used ONLY to maintain tree status.

(0023575)
nitramf (reporter)
2015-07-02 15:48

I have the same problems in 1.9.13

Standard testlink, happens sometimes if I switch between Testprojects.

My guess is that the new tree data gets appended, but it should replace the old path.

The path name gets bigger and bigger

(0023636)
nitramf (reporter)
2015-07-20 13:57
edited on: 2015-07-20 15:03

The cookie size exceeds 4kb. So how to solve this? Is it possible to work without cookies? (exept the login cookie)

(0023637)
nitramf (reporter)
2015-07-20 16:10
edited on: 2015-07-20 16:10

I have uploaded a solution.
The path of the file is testlink/third_party/ext-js/

All I did is to do a substring from 0 to 3000 to limit the cookie size to 3kb.
http://mantis.testlink.org/file_download.php?file_id=4252&type=bug [^]

(0023638)
modir (reporter)
2015-07-20 16:14

@nitramf: I don't think that limiting the size like this is a good solution. Better would be to save it in the localStorage instead of the cookie.

Here is more information about localStorage:
http://diveintohtml5.info/storage.html [^]

I wanted to implement this myself but I need to wait till my client approves the money for it.

(0024759)
mwbeatty (reporter)
2016-04-27 18:18

Is there any progress on this issue? We are having issues daily and it is impeding our project.

(0024760)
fman (administrator)
2016-04-28 13:31

no progress, too many ideas but no one has done other than offer and idea, and sit waiting for resolution.
workaround => stop saving tree state in cookie.
any help is welcomed

(0028681)
imsa (reporter)
2019-04-08 07:50
edited on: 2019-04-08 07:52

Hello.
We are working with Testlink for our projects and some getting very big, we did not meet this BadRequest problem.
The cookies were getting too big.
Also, I developed a workaround: I delete cookies when changing projects.
If you want, I can post the details of this solution to you, if it suits you, that you integrate it in your future version.

I said that I had unsuccessfully tested the nitramf solution on the modified .js

Best regards
Y.A for IMSA

(0028685)
fman (administrator)
2019-04-08 15:07

@imsa:
please provide your code

regards

(0028686)
imsa (reporter)
2019-04-09 05:14
edited on: 2019-04-09 06:48

Hi.
My code is simple :

/*
* Fixbug BadRequest
* I could have made a single loop
* to delete cookies but I preferred
* to separate them
*/
    if (is_array($_COOKIE)) {
        foreach ($_COOKIE as $key => $value) {
        // Cookies for test case
            if (preg_match (" /^ys-edit_tc_tproject_id_/i ", $key)) {
            // Delete cookie
                setcookie($key, '', time() - 36000000, '/');
            }
        // Cookies for requirement
            if (preg_match (" /^ys-req_specification_tproject_id_/i ", $key)) {
            // Delete cookie
                setcookie($key, '', time() - 36000000, '/');
            }
        // Cookies for testplan
            if (preg_match (" /^ys-testplan_tplan_id_/i ", $key)) {
            // Delete cookie
                setcookie($key, '', time() - 36000000, '/');
            }
        // Cookies for execution
            if (preg_match (" /^ys-test_exec_build_id_/i ", $key)) {
            // Delete cookie
                setcookie($key, '', time() - 36000000, '/');
            }
        // Cookies for tcplan
            if (preg_match (" /^ys-add_remove_tc_tplan_id_/i ", $key)) {
            // Delete cookie
                setcookie($key, '', time() - 36000000, '/');
            }
        }
    }

I had this in this file :
lib/general/frmWorkArea.php

We have 2.500 users, and i have no bad feedback for this modification.
And this bad request error is missing.

The code is tested on testlink 1.9.14 and 1.9.17

Best regards

(0028705)
fman (administrator)
2019-04-11 21:09

@imsa 2500 users is a big number, do you never though about making a donation to support test link development?

(0028709)
imsa (reporter)
2019-04-12 05:40

Hi.
I'm just recipient. But I send your request to the head of department
Best regards

(0028712)
fman (administrator)
2019-04-12 09:48

thanks

Notes
(0018351) fman (administrator) 2013-02-19 20:16	1. additional info: there is no indication about number of items you were trying to manage, may be there a limit exists 2. in any case you need to upgrade to 1.9.5 and retest

(0018358) Lukas Mencl (reporter) 2013-02-20 08:36	Ad 1. I am not sure what exactly you mean but there are some numbers: - We have got 8 testlink test projects which have 5635 test cases in total. - I was working (=testing regarding test plan) in test project which has 3862 test cases in total on test plan with 34 test cases when the issue appeared. - I was not modifying any steps, I simply wanted to move to next test case when the issue appeared. Ad 2. We will try to install 1.9.5 and retest when we find time for it. I will let you know about the result.

(0018365) moormanm (reporter) 2013-02-23 16:28 edited on: 2013-02-23 17:54	It appears that the large header data stems from the save/restore tree state logic in tcTree.js. Is there a more efficient way to store the collapse/expand tree state? EDIT: to reproduce this behavior, click 'expand all' on a test case tree panel that has a large number of nested testcases and then click on one of them. Examine the resultant header data in the user request.

(0018366) fman (administrator) 2013-02-23 17:58	>> Is there a more efficient way to store the collapse/expand tree state? do not know. we have taken example from EXT-JS version 3.x years ago. Some research on ext-js site is needed, right now think only workaround will be trying to increase cookie size if possible

(0018367) fman (administrator) 2013-02-23 18:15	The silly workaround to apply is to clean up cookies on browser. Do not think that loosing tree status (meaning what was opened on my last navigation) will be a great drawback

(0018375) Julian (reporter) 2013-02-27 07:17	Duplicate of 0003714

(0021513) elizk (reporter) 2014-08-27 13:23	we just faced this issue leading to server crashing - increasing the cookie size is a poor workaround.

(0021523) fman (administrator) 2014-08-31 17:24	>> we just faced this issue leading to server crashing - >> increasing the cookie size is a poor workaround. Ready to receive any kind of implementation that will solve the issue.

(0022717) modir (reporter) 2015-02-16 09:35	The problem still exists in 1.9.13. (We have about 7000 test cases.) I can not give the code right now but the solutions would be like this: Instead of saving the data in the cookie you would save it in a javascript variable. The variable could then be sent as a JSON object with a POST request to the server where it can be saved for later use. When a new page is loaded the values can be fetched again as a JSON object. I could help and provide some code but I would need more background information on why this is saved at all, etc.

(0022718) fman (administrator) 2015-02-16 09:52	Ideas are great, working solutions are better. >> I could help and provide some code but I would need more background >> information on why this is saved at all hmm may be to restore tree status ?

(0022743) modir (reporter) 2015-02-19 10:42	> hmm may be to restore tree status ? I need more information. Do you do anything on the server side with this information? Or is it only used client side?

(0022746) fman (administrator) 2015-02-19 10:51	>> I need more information. Do you do anything on the server side with this >> information? What do you mean? how this can change your approach? You need to read the code, and understand this, I can not explain any step done on code on several notes.

(0023462) modir (reporter) 2015-06-10 12:32	> What do you mean? Does any PHP code do anything with those information in the cookie? > how this can change your approach? If the PHP code on the server is doing nothing with the cookie information then I am a lot more flexible on choosing the best solution on the client side. And I don't have to send any information to the server. > You need to read the code, and understand this, I can not explain any step done on code on several notes. I hoped that there would be some code documentation like e.g. PHPdoc that would help me to learn the code faster. Since there is nothing like this I ask you to help me getting faster a understanding of the code. At least of the code which I need to solve this problem. I will read through the code in this case.

(0023463) fman (administrator) 2015-06-10 13:50	cookie is used ONLY to maintain tree status.

(0023575) nitramf (reporter) 2015-07-02 15:48	I have the same problems in 1.9.13 Standard testlink, happens sometimes if I switch between Testprojects. My guess is that the new tree data gets appended, but it should replace the old path. The path name gets bigger and bigger

(0023636) nitramf (reporter) 2015-07-20 13:57 edited on: 2015-07-20 15:03	The cookie size exceeds 4kb. So how to solve this? Is it possible to work without cookies? (exept the login cookie)

(0023637) nitramf (reporter) 2015-07-20 16:10 edited on: 2015-07-20 16:10	I have uploaded a solution. The path of the file is testlink/third_party/ext-js/ All I did is to do a substring from 0 to 3000 to limit the cookie size to 3kb. http://mantis.testlink.org/file_download.php?file_id=4252&type=bug [^]

(0023638) modir (reporter) 2015-07-20 16:14	@nitramf: I don't think that limiting the size like this is a good solution. Better would be to save it in the localStorage instead of the cookie. Here is more information about localStorage: http://diveintohtml5.info/storage.html [^] I wanted to implement this myself but I need to wait till my client approves the money for it.

(0024759) mwbeatty (reporter) 2016-04-27 18:18	Is there any progress on this issue? We are having issues daily and it is impeding our project.

(0024760) fman (administrator) 2016-04-28 13:31	no progress, too many ideas but no one has done other than offer and idea, and sit waiting for resolution. workaround => stop saving tree state in cookie. any help is welcomed

(0028681) imsa (reporter) 2019-04-08 07:50 edited on: 2019-04-08 07:52	Hello. We are working with Testlink for our projects and some getting very big, we did not meet this BadRequest problem. The cookies were getting too big. Also, I developed a workaround: I delete cookies when changing projects. If you want, I can post the details of this solution to you, if it suits you, that you integrate it in your future version. I said that I had unsuccessfully tested the nitramf solution on the modified .js Best regards Y.A for IMSA

(0028685) fman (administrator) 2019-04-08 15:07	@imsa: please provide your code regards

(0028686) imsa (reporter) 2019-04-09 05:14 edited on: 2019-04-09 06:48	Hi. My code is simple : /* * Fixbug BadRequest * I could have made a single loop * to delete cookies but I preferred * to separate them */ if (is_array($_COOKIE)) { foreach ($_COOKIE as $key => $value) { // Cookies for test case if (preg_match (" /^ys-edit_tc_tproject_id_/i ", $key)) { // Delete cookie setcookie($key, '', time() - 36000000, '/'); } // Cookies for requirement if (preg_match (" /^ys-req_specification_tproject_id_/i ", $key)) { // Delete cookie setcookie($key, '', time() - 36000000, '/'); } // Cookies for testplan if (preg_match (" /^ys-testplan_tplan_id_/i ", $key)) { // Delete cookie setcookie($key, '', time() - 36000000, '/'); } // Cookies for execution if (preg_match (" /^ys-test_exec_build_id_/i ", $key)) { // Delete cookie setcookie($key, '', time() - 36000000, '/'); } // Cookies for tcplan if (preg_match (" /^ys-add_remove_tc_tplan_id_/i ", $key)) { // Delete cookie setcookie($key, '', time() - 36000000, '/'); } } } I had this in this file : lib/general/frmWorkArea.php We have 2.500 users, and i have no bad feedback for this modification. And this bad request error is missing. The code is tested on testlink 1.9.14 and 1.9.17 Best regards

(0028705) fman (administrator) 2019-04-11 21:09	@imsa 2500 users is a big number, do you never though about making a donation to support test link development?

(0028709) imsa (reporter) 2019-04-12 05:40	Hi. I'm just recipient. But I send your request to the head of department Best regards

(0028712) fman (administrator) 2019-04-12 09:48	thanks

Issue History
Date Modified	Username	Field	Change
2013-02-19 12:20	Lukas Mencl	New Issue
2013-02-19 12:20	Lukas Mencl	File Added: testlink_bad_request.png
2013-02-19 20:16	fman	Note Added: 0018351
2013-02-19 20:17	fman	Status	new => feedback
2013-02-20 08:36	Lukas Mencl	Note Added: 0018358
2013-02-20 08:36	Lukas Mencl	Status	feedback => new
2013-02-23 16:28	moormanm	Note Added: 0018365
2013-02-23 17:53	moormanm	Note Edited: 0018365	View Revisions
2013-02-23 17:54	moormanm	Note Edited: 0018365	View Revisions
2013-02-23 17:58	fman	Note Added: 0018366
2013-02-23 18:15	fman	Note Added: 0018367
2013-02-27 07:17	Julian	Note Added: 0018375
2013-03-17 16:31	fman	Relationship added	duplicate of 0003714
2014-08-27 13:23	elizk	Note Added: 0021513
2014-08-31 17:24	fman	Note Added: 0021523
2015-02-16 09:35	modir	Note Added: 0022717
2015-02-16 09:52	fman	Note Added: 0022718
2015-02-19 10:42	modir	Note Added: 0022743
2015-02-19 10:51	fman	Note Added: 0022746
2015-06-10 12:32	modir	Note Added: 0023462
2015-06-10 13:50	fman	Note Added: 0023463
2015-06-29 18:30	fman	Relationship added	related to 0007180
2015-07-02 15:48	nitramf	Note Added: 0023575
2015-07-20 13:57	nitramf	Note Added: 0023636
2015-07-20 15:03	nitramf	Note Edited: 0023636	View Revisions
2015-07-20 16:08	nitramf	File Added: ext-all.js
2015-07-20 16:10	nitramf	Note Added: 0023637
2015-07-20 16:10	nitramf	Note Edited: 0023637	View Revisions
2015-07-20 16:14	modir	Note Added: 0023638
2016-04-27 18:18	mwbeatty	Note Added: 0024759
2016-04-28 13:31	fman	Note Added: 0024760
2019-04-08 07:50	imsa	Note Added: 0028681
2019-04-08 07:52	imsa	Note Edited: 0028681	View Revisions
2019-04-08 15:07	fman	Note Added: 0028685
2019-04-09 05:14	imsa	Note Added: 0028686
2019-04-09 06:46	imsa	Note Edited: 0028686	View Revisions
2019-04-09 06:48	imsa	Note Edited: 0028686	View Revisions
2019-04-11 21:09	fman	Note Added: 0028705
2019-04-12 05:40	imsa	Note Added: 0028709
2019-04-12 09:48	fman	Note Added: 0028712