Skip to content
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

Client are often disconnected without any reason [CORE3344] #3710

Closed
firebird-automations opened this issue Feb 10, 2011 · 24 comments
Closed

Comments

@firebird-automations
Copy link
Collaborator

Submitted by: vander clock stephane (arkadia)

Attachments:
ATTACHMENTS.xls

we often receive this kind of error : INET/inet_error: read errno = 10054

i was thinking it's was probably in our application, but today during a
simple gback we also receive this :

gbak:5940000 records written
gbak: ERROR:Error reading data from the connection.
gbak: ERROR:gds_$receive failed
gbak:Exiting before completion due to errors
Backup error with Gbak!

and in the firebird.log their is only
INET/inet_error: read errno = 10054 (loss of the connection)

the network not look like down because their was also 50 other clients
connected of the firebird server at this time that are still connected and
work normally

i start a new thread for this bug, because i m not sure it's related to the
previous bug i submit (INET/inet_error: read errno = 10055)

stéphane

@firebird-automations
Copy link
Collaborator Author

Commented by: @dyemanov

What is the upper line (right before the one you quoted) in firebird.log? I mean the one which has the host name. Does it have any (Server) or (Client) prefix?

@firebird-automations
Copy link
Collaborator Author

Commented by: vander clock stephane (arkadia)

SERVER35 Thu Feb 10 02:42:56 2011
INET/inet_error: read errno = 10054

on the client side (where gback was running) nothing in the firebird.log

@firebird-automations
Copy link
Collaborator Author

Modified by: @dyemanov

assignee: Vlad Khorsun [ hvlad ]

@firebird-automations
Copy link
Collaborator Author

Commented by: vander clock stephane (arkadia)

today it's different, some client (6 exactly) are not disconnected, BUT firebird not answer them, and freeze their connection !
The firebird server is still working and still answering correctly to all the other client (and no CPU is a 100%)

the client that are "freezing" are some client that was mostly doing some update in the database.

actually i can see them (the freezed connection) in the monitoring table ! they have all an inactive (0) state

i wait you answer before to shut them down in case you want i do something special on the firebird server
to better understand what's happen. but please answer fastly, i can not leave the connection in this state
to much of time ...

thanks by advance
stéphane

@firebird-automations
Copy link
Collaborator Author

Commented by: @hvlad

If it is not too late : take full memory dump (using task manager, for example) and show me contents of mon$attachment table (i hope you know, that you shoud start new transaction before query mon$xxx tables to refresh monitoring snapshot)

@firebird-automations
Copy link
Collaborator Author

Commented by: @hvlad

As for issue itself : i don't think i can do something without a reproducible test case.
I currently working on some reasons for 10054 errors but can't say if it is related with this ticket.

@firebird-automations
Copy link
Collaborator Author

Commented by: vander clock stephane (arkadia)

sure, find attached the Attachements.xls. the memory dump is 291 Mega bytes where i can put it for you ? i can create a ftp account for you but i need a way to send you privately the login / password

@firebird-automations
Copy link
Collaborator Author

Modified by: vander clock stephane (arkadia)

Attachment: ATTACHMENTS.xls [ 11906 ]

@firebird-automations
Copy link
Collaborator Author

Commented by: @hvlad

Yes, ftp is OK. Just don't forget to compress the dump :)
My e-mail is known to you, so feel free to send me all necessary info.

@firebird-automations
Copy link
Collaborator Author

Commented by: vander clock stephane (arkadia)

thanks for all vlad.. i send you the stuff by email.
can i shunt down the frozen connection or you think you can still need them ?

stephane

@firebird-automations
Copy link
Collaborator Author

Commented by: @hvlad

> i send you the stuff by email.
received, i'll try to look at it at weekend

> can i shunt down the frozen connection or you think you can still need them ?
sure, you can

@firebird-automations
Copy link
Collaborator Author

Commented by: vander clock stephane (arkadia)

hello vlad,

did you check the problem ? yesterday i try to install the last snapshot of firebird, but today it's look like a disaster .. all the cpu are at 90% of utilisation and some simple select query never return or are very very slow (10 min to return instead of few ms). i will need to kiil to process, but before to kill it do you want i do something on it ?

:(

@firebird-automations
Copy link
Collaborator Author

Commented by: vander clock stephane (arkadia)

another remark, we have actually a gbak running but it's look like blocked on this step :

gbak: 1000000 reccords written

also gbak is the official release of gbak (2.5.064) runing on a different server, where fb_inet_server.exe is the last snapshot version of firebird.

i don't know if it's matter, but just to say

@firebird-automations
Copy link
Collaborator Author

Commented by: @hvlad

> did you check the problem ?

just downloaded memory dump...
What was exact firebird version and build number when you produced this dump ?

> yesterday i try to install the last snapshot of firebird, but today it's look like a disaster .. all the cpu are at 90% of utilisation and some simple select query never return or are very very slow (10 min to return instead of few ms). i will need to kiil to process, but before to kill it do you want i do something on it ?

ensure you have no sweep running.
Killing of firebird process is a very bad idea in general.

> another remark, we have actually a gbak running but it's look like blocked on this step :
> gbak: 1000000 reccords written

are you specified switch -g in gbak's command line ?

> also gbak is the official release of gbak (2.5.064) runing on a different server, where fb_inet_server.exe is the last snapshot version of firebird.
> i don't know if it's matter, but just to say

is doesn't matter

@firebird-automations
Copy link
Collaborator Author

Commented by: vander clock stephane (arkadia)

> just downloaded memory dump...
> What was exact firebird version and build number when you produced this dump ?

the official release of firebird 2.5 (not snapshot or beta)

> ensure you have no sweep running.

no sweep running, but as i m still investigating now, it's look like that a simple query is using different plan in the last snapshot version than in the official release... i m still on it now, will come back with my conclusion on this point

> Killing of firebird process is a very bad idea in general.

:( yes i know, i always try to first shunt down the fb_inet_server by the service, and only if it's not shuntdown i kill the process (but this is rarely)

>are you specified switch -g in gbak's command line ?

yes

@firebird-automations
Copy link
Collaborator Author

Modified by: Sean Leyne (seanleyne)

summary: Client are often disconnected without any reason => Regression is optimizer handling of SQL statement

@firebird-automations
Copy link
Collaborator Author

Commented by: @dyemanov

Restored the original ticket title, as the reported optimizer issue is a completely different one and affects only v2.5.1 snapshot builds, so it's to be addressed separately. I will also remove all the related comments from here, in order to avoid further confusions.

@firebird-automations
Copy link
Collaborator Author

Modified by: @dyemanov

summary: Regression is optimizer handling of SQL statement => Client are often disconnected without any reason

@firebird-automations
Copy link
Collaborator Author

Commented by: Shari Clark (binman)

I am having the same issue on my server with clients connecting, can you advise what the bug was here please?

@firebird-automations
Copy link
Collaborator Author

Commented by: @hvlad

Similar behaviour not means "same issue". This ticket is 9 years old and a lot of new Firebrid releases with bugs fixed was issued since then.

If you think you see a not fixed bug in Firebird - feel free to create new ticket and describe what you see with as much details as possible.
Don't forget to specify version of Firebird server and client, architecture, OS and add any additional details necessary.

@firebird-automations
Copy link
Collaborator Author

Commented by: @asfernandes

Maybe it's then time to close this ticket if there is no reproducible case?

@firebird-automations
Copy link
Collaborator Author

Commented by: Sean Leyne (seanleyne)

As Vlad has pointed out, this ticket is more than 9 years old, and no one has been able to provide a reproducible test case.

@firebird-automations
Copy link
Collaborator Author

Modified by: Sean Leyne (seanleyne)

status: Open [ 1 ] => Resolved [ 5 ]

resolution: Cannot Reproduce [ 5 ]

@firebird-automations
Copy link
Collaborator Author

Modified by: @pcisar

status: Resolved [ 5 ] => Closed [ 6 ]

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

No branches or pull requests

2 participants