Re: directory storage is not responding
Toby Dickenson <[email protected]>
| Newsgroups | gmane.comp.web.zope.zodb.dirstorage |
|---|---|
| Message-ID | <[email protected]> |
On Wednesday 14 July 2004 17:39, Stefano Maffulli wrote: > On Fri, 2004-07-09 at 18:39, Toby Dickenson wrote: > > > According to strace, it hangs after the first fork(). I can send you > > > the strace output if you want. > > > > Yes, thanks > > it is here (strace -f -o strace.out zopectl start altoforno): > http://mediatria.com/~maffulli/strace.out.gz I think you meant http://mediatria.com/~stefano/strace.out.gz pid 6656 is zopectl. On line 9992, the supervisor zope process, pid 6657, calls waitpid on the zope worker process, pid 6658. Lines up to 41733 are the initialisation of that worker process. Part way through the initgroups mechanism waitpid in 6657 returns with ECHILD, and we see no more action from pid 6658. This is *inside* a call to setgid32.... I guess 6658 has terminated somehow. From line 41735 to 492866, pid 6657 is spinning calling waitpid in a loop. Im confused..... pid 6657 sets SIGCHLD to SIG_DFL in line 1024 So we have three points of interest: 1. A zope bug that causes its supervisor process to spin calling waitpid, rather than detect that the worker process has terminated. 2. A bug that causes something bad to happen inside initgroups. I guess it terminates that process, and it certainly triggers the bad behaviour in the supervisor process. 3. Some strange connection that means it happens when using DirectoryStorage, but not for FileStorage. Any clues in syslog or dmesg? What linux kernel are you using? Anything unusual? If not, Im not sure where else to suggest taking this next...... ------------------------------------------------------- This SF.Net email is sponsored by BEA Weblogic Workshop FREE Java Enterprise J2EE developer tools! Get your free copy of BEA WebLogic Workshop 8.1 today. http://ads.osdn.com/?ad_idG21&alloc_id040&op=click