On 30Dec2014 08:36, Neal Becker <ndbecker2@xxxxxxxxx> wrote:
On my server, after updating f20->f21, shell startup is _extremely_ slow.
Using strace, it appears that the problem is dbus
08:22:57 connect(5, {sa_family=AF_LOCAL,
sun_path="/var/run/dbus/system_bus_socket"}, 33) = 0
...
08:22:57 recvmsg(5, 0x7fff640c3970, MSG_CMSG_CLOEXEC) = -1 EAGAIN (Resource
temporarily unavailable)
08:22:57 poll([{fd=5, events=POLLIN}], 1, 25000) = 1 ([{fd=5, revents=POLLIN}])
08:23:22 recvmsg(5, {msg_name(0)=NULL,
msg_iov(1)=[{"l\3\1\0013\0\0\0\4\0\0\0m\0\0\0\6\1s\0\7\0\0\0:1.1155\0"...,
2048}], msg_controllen=0, msg_flags=MSG_CMSG_CLOEXEC}, MSG_CMSG_CLOEXEC) = 179
08:23:22 write(6, "\1\0\0\0\0\0\0\0", 8) = 8
Notice the 30second wait for a response from dbus?
Also in /var/log/messages, various dbus messages such as:
Dec 30 08:19:32 nbecker7 su: (to root) nbecker on pts/0
Dec 30 08:20:02 nbecker7 dbus[809]: [system] Connection has not authenticated
soon enough, closing it (auth_timeout=30000ms, elapsed: 30004ms)
Any hints?
Those log line times don't match the strace times. But I'll assume they're
typical, and associate with your shell startups timingwise.
A few questions:
- why does the shell use dbus at all? (just wondering, I'm presuming this is some bash feature new to me)
- can you strace dbus (pid 809 in the above log)? using strace's -f option if dbus forks.
- 30s is in my head as a standard DNS timeout, but it is also common to some
other things and your sample strace above shows only 25s. (Though I notice also
that the timeout in the poll() call is itself exactly 25s).
So you're not seeing a 25s wait before dbus replies, you're seeing a 25s
timeout on the client (your shell?)?
I'd be trying 2 things: see if dbus has a config option to turn up its
debugging messages, and strace dbus around a shell start. Eg:
strace -p 809 -f -o strace.dbus.out
Disclaimer: not running Fedora here, so this is all general advice.
Cheers,
Cameron Simpson <cs@xxxxxxxxxx>
They utter mere words; with empty oaths they make promises; so lawsuits
spring up like poisonous weeds in the furrows of the field.
Sounds like 20th-century USA, doesn't it? This is from the book of Hosea,
chapter 10, the 4th verse. It's only a few thousand years old...
- Dan Nitschke <elvis@xxxxxxxxxxxx>
--
users mailing list
users@xxxxxxxxxxxxxxxxxxxxxxx
To unsubscribe or change subscription options:
https://admin.fedoraproject.org/mailman/listinfo/users
Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Have a question? Ask away: http://ask.fedoraproject.org
