-
Notifications
You must be signed in to change notification settings - Fork 31
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
After Upgrade to 5.13 HTTP/HTTPS Farms not working anymore. #123
Comments
Its a Hyper-V environment. Thanks, Michael |
Same issue here, I upgraded and all my backends seem to run into a Timeout most of the times.
It feels super weird. Can I go back to 5.12? |
Just for fun i installed version 5.13 manually on an fresh debian 11 (bullseye). Here the supportsave from this version. Thanks, Michael |
I cannot access the admin gui to create a supportfile. asap when im on the premises tomorrow I will create one. |
Hi, supportsave can be generated via the command line too. /usr/local/zevenet/bin/supportsaveIt will generate the file in /tmp directory. Best regards. |
Here u go :) |
Some update? |
@Sysadminfromhell |
Hi, If the the other services experience timeouts they are caused by the backend communication Mar 1 15:11:26 HOSTNAME zproxy: timeout request for client YYYY:65486 All the config files seems to be well-configured. Could you test the connection to the backends? Could you enable the farm logs to view more info in /var/log/syslog and test a connection to the service farm? Could you paste the state of the farm after the test service farm? curl --unix-socket /tmp/FARMNAME_proxy.socket http://localhost/listener/0 Thanks in advance. |
Telnet works: The curl doesnt work:, because my farm isnt existing. And the log says:
Different access, different backend:
|
Hi, We will keep you updated. Thanks! |
Hi @Sysadminfromhell could you please try to increase the timeouts? In addition, which kind of application is running in the backend side? Thanks! |
Hi again @Sysadminfromhell Feb 28 03:42:01 HMLRPX02 farmguardian[744]: (INFO) lslb :: Deleting established sessions to a backend 0 from farm LoadBalancer in service Grafana-Dashboard |
Hi again @Sysadminfromhell Also, some "error reading from client" detected. Are those legitimated clients? Feb 28 05:19:11 HMLRPX02 zproxy: error reading from client 134.209.254.251:25978 (Connection reset by peer) |
@nevola Best regards. |
We need to figure out what is happening with the web GUI,so please do the
following:
Press 12 the development console will be loaded
Go to network tab where all the requests will be shown
Reproduce the problem, a line in red colour will be shown or I guess.
Finally export all the data in the dev console to a HAR, here explained how
to export a HAR file.
https://support.google.com/admanager/answer/10358597?hl=en
Finally attach the file to this issue for our review
Thanks
El vie., 3 mar. 2023 15:50, Trylok ***@***.***> escribió:
… @nevola <https://github.com/nevola>
Just to be clear. As soon as you start an http/https the farmlist is empty
and the dashboard chrashes, as I have already described above. Even if no
backends are configured.
l4xnat farms to not show this beavior.
So, it has nothing to do with the backends, Changing the timing makes no
difference.
Not even a simple url redirect works without crashing the farm.
Best regards.
—
Reply to this email directly, view it on GitHub
<#123 (comment)>, or
unsubscribe
<https://github.com/notifications/unsubscribe-auth/AFBQEPBQRE2JGM3MEQJLZG3W2IAMLANCNFSM6AAAAAAVKR3R5Y>
.
You are receiving this because you are subscribed to this thread.Message
ID: ***@***.***>
|
Here you go. This is from the dashboard/home section, where i get the endless spinning cycle after starting the farm. The farms section looks like this before i start the farm. |
I tried, this fixed the "error" partial.
not legit, but nothing really to be worried about, Happen all the time.
The guardian check on this backend is only "ping" this should work all the time. Farm explenation: pve => Proxmox VE Backend Update: The pve and pbs works sometimes, the grafana and Guacamole one is always just white and empty. Nothing comes up not even a error page. |
It seems like the request:
https://192.168.0.32:444/zapi/v4.0/zapi.cgi/farms
is generating in the load balancer a 500 Error.
Please send us again a fresh supportsave of the home server 192.168.0.32
for our review?
Thanks!
i/v4.0/zapi.cgi/farms
El vie, 3 mar 2023 a las 16:13, Trylok ***@***.***>)
escribió:
… Here you go. This is from the dashboard/home section.
192.168.0.32.zip
<https://github.com/zevenet/zlb/files/10883189/192.168.0.32.zip>
This is my test installation with just one farm configured.
The farms section looks like this before i start the farm.
[image: image]
<https://user-images.githubusercontent.com/50360497/222755287-780b4566-20d8-4c01-a6a0-ebc92c58d375.png>
And this after starting it.
[image: image]
<https://user-images.githubusercontent.com/50360497/222754874-6a1b87a3-f6fe-44b1-8177-d4f5d03b57bf.png>
—
Reply to this email directly, view it on GitHub
<#123 (comment)>, or
unsubscribe
<https://github.com/notifications/unsubscribe-auth/AFBQEPH4NKHFZDF5MG7K22DW2IDBJANCNFSM6AAAAAAVKR3R5Y>
.
You are receiving this because you commented.Message ID:
***@***.***>
--
Emilio CamposZEVENET Teamwww.zevenet.com
Review ZEVENET Product at Gartner Insights <https://gtnr.io/LfXtqnsSr>
<https://www.linkedin.com/company/zevenet> <https://twitter.com/zevenet>
<https://www.facebook.com/zevenet> <https://github.com/zevenet>
[image: ZEVENET] <https://www.zevenet.com/signature/>
DISCLAIMER: This message contains confidential information and is intended
only for the individual named. If you
are not the named addressee please notify the sender immediately by email
if you have received it by mistake and
delete it from your system, you should not disseminate, distribute or copy
this email in whole or in part.
|
Here you go. |
Hi Trylok, Please, try to install ZEVENET CE 5.13 in a Debian Buster. Thanks! |
@Rendone-zevenet |
Hi @Sysadminfromhell, it seems that it is related to network issues, please try to leave a ping from ZEVENET command line to your backends and check the lost packets. If you need us to review something else, please open a new issue.
|
I cannot really see that this should be a problem.
|
I can confirm that there are no losts packets to the backends. All of them perfectly available. |
For what it's worth I can confirm that since the upgrade to 5.13 my HTTP/HTTPS Farms are no longer working properly either. |
Dear @ct-utel could you please enable the logs in the affected farm, then reproduce the problem and finally, generate a support save file to be shared with us? Also, a wireshark pcap file would help to analyze what is happening. Thank you! |
Hi nevola, Please find a support save I generated earlier today. I'm about to head home for the day but I will try and generate the pcap file in the morning. |
Hi @ct-utel I'm seeing: Mar 5 13:34:38 svr-lb-01 zproxy: [f:][th:61c5f700] Malformed chunk, size was not found If you could export an HAR file from your browser with the malformed request would help. Please send the HAR file to [email protected] Thank you!
|
Quick Update. |
Good day, @Trylok, We're currently applying patches to fix this issue, and we seem to have narrowed down the issue and found a patch that appears to resolve it. To be sure we've narrowed it down correctly, can you tell me the properties of the request? Primarily I'd like to know if the Transfer-Encoding is chunked or not. Thank you. |
@naortega |
Okay. We have a patch ready for this issue which we hope to be releasing soon. Thank you for your patience and cooperation. |
I'm posting to notify that a new version of zproxy (0.9.2) has been release that should fix the issues with chunking and timeouts. |
@naortega thanks for the update. With the latest updates we are still having issues. We have HTTPS redirect for a TeamCity server, the GUI doesn't load at all via the load balancer but going to IP:port directly its fine. We have a Home Assistance instance that uses the same farm, the initial login page loads but times out trying to use the dashboard. Again going directly to it, loads fine. There is also a Project Send instance on the same farm. It loads fine internally but externally it's very slow to respond. I'd be happy to send another support save file it that would help? so you might be able to compare the before and after upgrade. |
A support save would be nice, but it would also be helpful to know better what problems you are experiencing. Regarding the first issue you experience (HTTPS redirect for TeamCity server): what exactly doesn't load? What's the request, what's the expected result, and what's the erroneous result you're getting? In regards to the second issue (Home Assistance): Are you sure this isn't due to a network issue or a malconfiguration of the Zevenet machine? Could you show me the logs pertaining to this problem? Thank you for your cooperation. |
@naortega |
@naortega I have just e-mailed a new support save to the existing e-mail chain, I believe you are in it. I'm pretty certain none of the issues are network related as the behavior started the moment we upgraded zevenet to 5.13, also getting to services directly without the load balance in the middle is fine. |
one question, do you use websocket in this service?
Thanks!
El lun, 20 mar 2023 a las 14:22, ct-utel ***@***.***>)
escribió:
… A support save would be nice, but it would also be helpful to know better
what problems you are experiencing. Regarding the first issue you
experience (HTTPS redirect for TeamCity server): what exactly doesn't load?
What's the request, what's the expected result, and what's the erroneous
result you're getting?
In regards to the second issue (Home Assistance): Are you sure this isn't
due to a network issue or a malconfiguration of the Zevenet machine? Could
you show me the logs pertaining to this problem?
Thank you for your cooperation.
@naortega <https://github.com/naortega> I have just e-mailed a new
support save to the existing e-mail chain, I believe you are in it.
The TeamCity server is getting "504 Gateway timeout - zproxy" just trying
to get to it's web gui.
[image: image]
<https://user-images.githubusercontent.com/127128354/226350446-4db4d74b-0d64-41da-8aad-a2978f63f7ad.png>
I'm pretty certain none of the issues are network related as the behavior
started the moment we upgraded zevenet to 5.13, also getting to services
directly without the load balance in the middle is fine.
—
Reply to this email directly, view it on GitHub
<#123 (comment)>, or
unsubscribe
<https://github.com/notifications/unsubscribe-auth/AFBQEPHHT5K2EGLQBY7Y6ADW5BKY3ANCNFSM6AAAAAAVKR3R5Y>
.
You are receiving this because you commented.Message ID:
***@***.***>
--
Emilio CamposZEVENET Teamwww.zevenet.com
Review ZEVENET Product at Gartner Insights <https://gtnr.io/LfXtqnsSr>
<https://www.linkedin.com/company/zevenet> <https://twitter.com/zevenet>
<https://www.facebook.com/zevenet> <https://github.com/zevenet>
[image: ZEVENET] <https://www.zevenet.com/signature/>
DISCLAIMER: This message contains confidential information and is intended
only for the individual named. If you
are not the named addressee please notify the sender immediately by email
if you have received it by mistake and
delete it from your system, you should not disseminate, distribute or copy
this email in whole or in part.
|
Thanks for the confirmation, I guess we have to take this information into
consideration in order to investigate further in this issue.
Regards!
El lun, 20 mar 2023 a las 14:40, ct-utel ***@***.***>)
escribió:
… It looks like TeamCity might but should be able to fallback if unavailable
and Home Assistant can for API calls but not the GUI itself.
[image: image]
<https://user-images.githubusercontent.com/127128354/226357110-c6cad877-6df8-4425-8306-b7bb26bf44ef.png>
—
Reply to this email directly, view it on GitHub
<#123 (comment)>, or
unsubscribe
<https://github.com/notifications/unsubscribe-auth/AFBQEPFZPQX7U6XH3YUHUVLW5BM4FANCNFSM6AAAAAAVKR3R5Y>
.
You are receiving this because you commented.Message ID:
***@***.***>
--
Emilio CamposZEVENET Teamwww.zevenet.com
Review ZEVENET Product at Gartner Insights <https://gtnr.io/LfXtqnsSr>
<https://www.linkedin.com/company/zevenet> <https://twitter.com/zevenet>
<https://www.facebook.com/zevenet> <https://github.com/zevenet>
[image: ZEVENET] <https://www.zevenet.com/signature/>
DISCLAIMER: This message contains confidential information and is intended
only for the individual named. If you
are not the named addressee please notify the sender immediately by email
if you have received it by mistake and
delete it from your system, you should not disseminate, distribute or copy
this email in whole or in part.
|
Hi, Kroki |
Hi @Trylok could you please increase Client timeout from 30 to 60, and see if there is any improvement? Thank you. |
@nevola |
Hi @Trylok , Good news. We have a new binary with several fixes that might resolve your issues. Please download it from here: https://drive.google.com/file/d/1s9iA28TsD07kcvr5OSIiHW5l8J562KsV/view?usp=share_link You've to upload to the appliance and rename it in the path /usr/local/zevenet/app/zproxy/bin/zproxy After that, please restart the farms and let us know how it goes, |
With the new Binary i didn't see any issues with chunked transfers any more. Kroki |
@Kroki0815 thank you for the confirmation. |
@nevola |
@nevola |
@nevola |
We'll have a look into the authorization errors (401 response code). The pending connections seems to be an issue with the counter, but it should not be an issue with actual connections pending. We'll try resolving these issues as soon as possible. |
@Trylok Would it be possible for you to send a HAR that displays the 401 error you get? Is it when connecting via websocket? Is it a normal HTTP POST request? |
@naortega |
Hi @Trylok could you please enable farm logs, then reproduce the problem and then generate a new support save to [email protected] (please do not send it via github). |
File has been send. |
@Trylok In the most recent release of Zproxy (0.9.4) we've resolved the issue of pending connections counting. Are you still having these issues? Also, in regards to the 401 responses, have these been resolved? |
@naortega |
Understood. Would you mind taking a look at the logs corresponding to these requests on the backends? Zproxy itself does not give 401 responses unless received from the backend. It may be possible that Zproxy is erroneously manipulating the requests such that your backend gives that response, in which case there should be some error log from your backend. |
After starting a HTTP farm the farms list ist empty.
l4xnat farms are working but not displayed anymore.
Only stopping the corresponding Virtual Interface and rebooting the system brings the List back.
For testing i did a fresh iso install, only creating a new HTTP farm. With the same result.
HTTP/HTTPS Farms do not work anymore, no matter on upgrade or fresh install.
Thanks, Michael
The text was updated successfully, but these errors were encountered: