[Gluster-devel] Mounting with file servers is failing very frequently (in every day)
krishna at zresearch.com
Thu Nov 1 13:27:41 UTC 2007
On 10/31/07, Joseph Job <joseph at spectrum.net.in> wrote:
> Hi Raghavendra,
> I have commented the transport-timeout, and remounted it again. After
> working for some time the mounting failed again...see the logs in the
> client side...any help in this regard would be highly appreciated...
> 2007-10-31 10:26:20 W [client-protocol.c:210:call_bail] client_213:
> ail-out. pending frames = 1. last sent = 2007-10-31 10:22:58. last
> received =
> 07-10-31 10:23:22 transport-timeout = 108
> 2007-10-31 10:26:20 C [client-protocol.c:218:call_bail] client_213:
> 2007-10-31 10:26:20 D [tcp.c:131:cont_hand] tcp: forcing
> poll/read/write to
> k on blocked socket (if any)
> Regarding the type of files, all files with average of 2 to 4 mb each.
> But there are around 2 to 3 lakhs in every folder...around 20 folders...
Currently if a directory has more than 20 to 30 thousand files, the AFR
selfheal verification code handles it inefficiently which causes the
delay you are seeing. I will be committing a fix soon (hopefully
in a week's time) to make it fast.
More information about the Gluster-devel