Hi, I am working on an academic research project using nfs-ganesha. I am trying to do some debugging and believe that logging messages are going missing. For example, I am trying to trace when writes occur in FSAL/FSAL_VFS/file.c:vfs_write2.c. I put a four logging messages in using LogDebug, and only one that shows entry into the function gets logged - the others which surround the write and show function exiting do not appear in the log. I have put some code in that writes my own trace messages to a file so I can see that these other log messages should also be appearing as the flow of control passes through them, but they are not written. Is there some throttling mechanism that I am running into? Or something else that I might need to be aware of? Thanks
No, there's no throttling. If anything at a given log level / component is written, it all should be written. One thing to check is if they're all the same component, and which components have logging enabled. Daniel On 1/16/20 3:30 PM, clay@cs.georgetown.edu wrote:
Hi,
I am working on an academic research project using nfs-ganesha. I am trying to do some debugging and believe that logging messages are going missing.
For example, I am trying to trace when writes occur in FSAL/FSAL_VFS/file.c:vfs_write2.c. I put a four logging messages in using LogDebug, and only one that shows entry into the function gets logged - the others which surround the write and show function exiting do not appear in the log. I have put some code in that writes my own trace messages to a file so I can see that these other log messages should also be appearing as the flow of control passes through them, but they are not written.
Is there some throttling mechanism that I am running into? Or something else that I might need to be aware of?
Thanks _______________________________________________ Devel mailing list -- devel@lists.nfs-ganesha.org To unsubscribe send an email to devel-leave@lists.nfs-ganesha.org
Could it be because the LogDebug is between vfs_set_credentials() and vfs_restore_ganesha_credentials()? Between those, the thread is running as a different user and won't have write permissions to the log file. Thanks, Pradeep On Fri, Jan 17, 2020 at 6:41 AM Daniel Gryniewicz <dang@redhat.com> wrote:
No, there's no throttling. If anything at a given log level / component is written, it all should be written. One thing to check is if they're all the same component, and which components have logging enabled.
Daniel
Hi,
I am working on an academic research project using nfs-ganesha. I am
On 1/16/20 3:30 PM, clay@cs.georgetown.edu wrote: trying to do some debugging and believe that logging messages are going missing.
For example, I am trying to trace when writes occur in
FSAL/FSAL_VFS/file.c:vfs_write2.c. I put a four logging messages in using LogDebug, and only one that shows entry into the function gets logged - the others which surround the write and show function exiting do not appear in the log. I have put some code in that writes my own trace messages to a file so I can see that these other log messages should also be appearing as the flow of control passes through them, but they are not written.
Is there some throttling mechanism that I am running into? Or something
else that I might need to be aware of?
Thanks _______________________________________________ Devel mailing list -- devel@lists.nfs-ganesha.org To unsubscribe send an email to devel-leave@lists.nfs-ganesha.org
_______________________________________________ Devel mailing list -- devel@lists.nfs-ganesha.org To unsubscribe send an email to devel-leave@lists.nfs-ganesha.org
Pradeep, Good thought. You’re right, if you do a log inside a set credentials block, it won’t show up. Two fixes for development debug if you can’t move the log outside the set credentials block: Change the permissions of the log file to world writeable. Use syslog instead of logging to a file (I’m pretty sure with syslog we keep a syslogd connection open for the duration so it shouldn’t be affected by current process credentials, and may not be affected by fsuid/fsgid anyway which is what set credentials changes (it leaves euid/egid alone). Frank From: Pradeep [mailto:pradeepthomas@gmail.com] Sent: Friday, January 17, 2020 4:52 PM To: Daniel Gryniewicz <dang@redhat.com> Cc: Ganesha-devel <devel@lists.nfs-ganesha.org> Subject: [NFS-Ganesha-Devel] Re: Limits to logging Could it be because the LogDebug is between vfs_set_credentials() and vfs_restore_ganesha_credentials()? Between those, the thread is running as a different user and won't have write permissions to the log file. Thanks, Pradeep On Fri, Jan 17, 2020 at 6:41 AM Daniel Gryniewicz <dang@redhat.com <mailto:dang@redhat.com> > wrote: No, there's no throttling. If anything at a given log level / component is written, it all should be written. One thing to check is if they're all the same component, and which components have logging enabled. Daniel On 1/16/20 3:30 PM, clay@cs.georgetown.edu <mailto:clay@cs.georgetown.edu> wrote:
Hi,
I am working on an academic research project using nfs-ganesha. I am trying to do some debugging and believe that logging messages are going missing.
For example, I am trying to trace when writes occur in FSAL/FSAL_VFS/file.c:vfs_write2.c. I put a four logging messages in using LogDebug, and only one that shows entry into the function gets logged - the others which surround the write and show function exiting do not appear in the log. I have put some code in that writes my own trace messages to a file so I can see that these other log messages should also be appearing as the flow of control passes through them, but they are not written.
Is there some throttling mechanism that I am running into? Or something else that I might need to be aware of?
Thanks _______________________________________________ Devel mailing list -- devel@lists.nfs-ganesha.org <mailto:devel@lists.nfs-ganesha.org> To unsubscribe send an email to devel-leave@lists.nfs-ganesha.org <mailto:devel-leave@lists.nfs-ganesha.org>
Devel mailing list -- devel@lists.nfs-ganesha.org <mailto:devel@lists.nfs-ganesha.org> To unsubscribe send an email to devel-leave@lists.nfs-ganesha.org <mailto:devel-leave@lists.nfs-ganesha.org>
participants (4)
-
clay@cs.georgetown.edu -
Daniel Gryniewicz -
Frank Filz -
Pradeep