Re: [Evolution] death by "too many open files"?/emacs key bindings?



Mmmh
I am still using the original MDK 9.1 evolution (1.2.4). I am connected
to a Samba server. I have never been able to run the evo mbox files on
the Samba Server. I am actually running an archive (large file, mbox,
230MB). I have never been able to archive to the samba server after the
file reached a certain size. I am storing the file on the local machine
and then copy it over to the samba server in order to get backed up
during the nightly samba backup.
May be this problem has a different cause.

Tom



On Wed, 2003-10-01 at 13:59, Phil Sackinger wrote:
On Wed, 2003-10-01 at 14:20, Jeffrey Stedfast wrote:
On Wed, 2003-10-01 at 16:03, Phil Sackinger wrote:
On Wed, 2003-10-01 at 13:33, Jeffrey Stedfast wrote:
depends on what files are open I guess

/home/me/evolution/local/myfoldername/mbox.ibex.index (nfsserver:/u/me)

one open file exceeds the limit? :-)

Sorry I wasn't more clear. Definitely more than one. Overall, right now,

lsof|grep evolution|wc
7385

Some of the open files are what I expect:

evolution 30842 pasacki  mem    REG        8,2    14824   1368843
/usr/lib/bonob
o/monikers/libmoniker_std_2.so

and there's plenty (1531)  of those.

There are also lots (122) of open files that are:
evolution 30842 pasacki   16r  FIFO        0,5           15219446 pipe
evolution 30842 pasacki   17w  FIFO        0,5           15219446 pipe
evolution 30842 pasacki   18r  FIFO        0,5           15219447 pipe
.
.
.

But most strange are *man*y (5600) of these:

evolution 30842 pasacki   35u  sock        0,0           17110030 can't
identify
 protocol

and, lastly, about 179 of the mbox.ibex.index.data files open on my
nfsserver.



At one point I heard folklore  that Solaris NFS servers and Linux NFS
clients [Solaris 8; Linux = RH7.3] do not always reliably interact.

Might that have something to do with the problem?

I have no idea as I have heard no such rumour nor have I any idea was to
what that could be in reference to.
I was afraid that evolution might be locking and unlocking files for
mail that were not getting closed properly due to some hanging NFS
connection. I'm speculating about subjects which I don't know, so I'm
probably way off base here.




On Wed, 2003-10-01 at 12:03, Phil Sackinger wrote:
I've noticed Evo 1.4.4 dying after, oh, about a week or so because of
"too many open files". 

An "lsof | grep evolution | wc" gives me 6748 lines.

Lots of entries say "can't identify protocol".

Is this:
      an evo problem, 
      me with too many folders, or 
      a system limitation that I should fix (ulimit -n)?

On an unrelated matter, I'd like to get emacs style keybindings that I
enjoyed under Evo 1.2 and can't find in 1.4.



_______________________________________________
evolution maillist  -  evolution lists ximian com
http://lists.ximian.com/mailman/listinfo/evolution





[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]