Re: [Moosefs-users] More nexenta bugs

[Michal B. <mic...@ge...>]

Hi Robert!

 

Yes, we'd like MooseFS to be fully compatible with Nexenta. So if you still
have some issues, please send them to us.

 

We are (almost) sure that hanging up of the chunkserver was not connected
with SIGPIPE signal. To the "csserv_serve" function we added a test
fragment:

 

//sigpipe test

        int fd[2];

        pipe(fd);

        close(fd[0]);

        if (write(fd[1],"",1)<0) {

                mfs_errlog(LOG_NOTICE,"sigppipe test");

        }

        close(fd[1]);        

//sigpipe test

 

and everything worked fine, the logs showed:

 

(...) nexenta mfschunkserver[14100]: [ID 801593 daemon.notice] sigppipe
test: EPIPE (Broken pipe)

(...)

 

 

Is your problem repeatable? Where did you get information about SIGPIPE
from? Have you correctly run gdb (with the correct core)? Can you run "info
threads" in gdb? Maybe it will show us a real cause of the problem.

 

Thank you!

 

Michał Borychowski 

MooseFS Support Manager

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

Gemius S.A.

ul. Wołoska 7, 02-672 Warszawa

Budynek MARS, klatka D

Tel.: +4822 874-41-00

Fax : +4822 874-41-01

 

 

 

From: Robert Dye [mailto:ro...@in...] 
Sent: Thursday, March 17, 2011 7:28 PM
To: moo...@li...
Subject: [Moosefs-users] More nexenta bugs

 

I guess I should ask if I should be submitting Nexenta bugs, if this is
something the project will support in the future?  Either way, looks like I
was able to get specific trace information (shown below).

 

Problem:  mfschunkserver segfaults after running for a few hours.

 

Output:  

[New LWP 5]

main server module: listen on *:9422

[New LWP 6]

[New LWP 7]

[New LWP 8]

[New LWP 9]

[New LWP 10]

[New LWP 11]

[New LWP 12]

[New LWP 13]

[New LWP 14]

[New LWP 15]

[New LWP 16]

[New LWP 17]

[New LWP 18]

[New LWP 19]

[New LWP 20]

[New LWP 21]

[New LWP 22]

[New LWP 23]

[New LWP 24]

[New LWP 25]

stats file has been loaded

mfschunkserver daemon initialized properly

 

 

 

Program received signal SIGPIPE, Broken pipe.

0xfeed57d5 in __write () from /lib/libc.so.1

 

(gdb) bt

#0  0xfeed57d5 in __write () from /lib/libc.so.1

#1  0xfeebf37e in write () from /lib/libc.so.1

#2  0x0804d684 in csserv_write (eptr=0x8a32bf0) at csserv.c:1631

#3  0x0804f604 in csserv_serve (pdesc=0x8033c90) at csserv.c:1900

#4  0x0806208c in mainloop () at ../mfscommon/main.c:348

#5  0x080626fb in main (argc=<value optimized out>, argv=0x40) at
../mfscommon/main.c:1101