Hi Andre,
We are trying to reproduce the issue inhouse, mean while you can try
increasing the "node_timeout" and "conn_timeout" values from
60 to anything
higher.
Kindly correct me If I am wrong, you are using the BETA1 or BETA2 version
of UFO. If yes try upgrading it to the newer version available at
http://download.gluster.com/pub/gluster/glusterfs/3.2/UFO/.
It would be helpful if you can open a bug and list down the steps to
reproduce it and the configuration you are running.
On Fri, Feb 24, 2012 at 11:45 PM, Andr? Cruz <andre.cruz at co.sapo.pt>
wrote:
> Hello.
>
> I've setup the UFO stack as described in the documentation but I'm
running
> into a problem. I have a script that inserts a large number of objects with
> size < 4MB, and after it has inserted a few thousand files the API
suddenly
> stops responding. PUTs fail after a 60s wait with either a 404 or 503
> result.
>
> This is what I see in the logs:
>
> Feb 24 16:35:08 vmdev-gluster03 account-server 127.0.0.1 - -
> [24/Feb/2012:16:35:08 +0000] "HEAD
>
/AUTH_8140ab2e72bd3fd9d3e89e00d2d2c620/0/AUTH_8140ab2e72bd3fd9d3e89e00d2d2c620"
> 204 - "txd1be8077fb3e4c1e870dcee493dd8681" "-"
"-" 0 .2608 ""****
> Feb 24 16:36:08 vmdev-gluster03 proxy-server ERROR with Container server
> 127.0.0.1:6011/AUTH_8140ab2e72bd3fd9d3e89e00d2d2c620 re: Trying to get
> container info for /AUTH_8140ab2e72bd3fd9d3e89e00d2d2c620/blocks10: Timeout
> (60s) (t xn: txd1be8077fb3e4c1e870dcee493dd8681)****
> Feb 24 16:36:08 vmdev-gluster03 proxy-server 10.135.7.165 10.135.7.165
> 24/Feb/2012/16/36/08 PUT
> /v1/AUTH_8140ab2e72bd3fd9d3e89e00d2d2c620/blocks10/treta HTTP/1.0 404 -
> curl/7.21.0%20%28x86_64-pc-linux-gnu%29%20libcurl/7.21.0%20O
> penSSL/0.9.8o%20zlib/1.2.3.4%20libidn/1.15%20libssh2/1.2.6
> TESTER%3Asapodisco%2CAUTH_tk5934e6d77edf42ebb388ee5bc4cfe788 - - -
> txd1be8077fb3e4c1e870dcee493dd8681 - 60.2707****
> Feb 24 16:36:08 vmdev-gluster03 proxy-server STDOUT: 10.135.7.165 - -
> [24/Feb/2012 16:36:08] code 400, message Bad request syntax ('Feb 22
> 06:32:57 vmdev-gluster02 mpt-statusd: detected non-optimal RAID
status')
> (txn: txd1be8077 fb3e4c1e870dcee493dd8681)****
> Feb 24 16:36:08 vmdev-gluster03 proxy-server STDOUT: 10.135.7.165 - -
> [24/Feb/2012 16:36:08] "Feb 22 06:32:57 vmdev-gluster02 mpt-statusd:
> detected non-optimal RAID status" 400 - (txn:
> txd1be8077fb3e4c1e870dcee493dd8681)****
> Feb 24 16:36:16 vmdev-gluster03 container-server 127.0.0.1 - -
> [24/Feb/2012:16:36:16 +0000] "HEAD
>
/AUTH_8140ab2e72bd3fd9d3e89e00d2d2c620/0/AUTH_8140ab2e72bd3fd9d3e89e00d2d2c620/blocks10"
> 204 - "txd1be8077fb3e4c1e870dcee493dd8681 " "-"
"-" 67.5196****
> ** **
> **
> **
> **So between my PUTs some process that runs locally executes a HEAD
> request which, since the container has a few thousand files, takes too long
> to respond and the request fails. Which process send this HEAD request?**
> **
> **
> **When the container-server receives this HEAD request it starts scanning
> the whole dir, as expected:**
> **
>
stat("/mnt/gluster-object/AUTH_8140ab2e72bd3fd9d3e89e00d2d2c620/blocks10/5/V/m/5VmTFnaQ6G5WL0klYlRjMizCBzZSme4_-VNrRnP35lhef3yzIxyvzjaG_NnerWeFrtlgHf80g4VnmP0DS2G32g",
> {st_mode=S_IFREG|0600, st_size=37768, ...}) = 0****
>
getxattr("/mnt/gluster-object/AUTH_8140ab2e72bd3fd9d3e89e00d2d2c620/blocks10/5/V/m/5VmTFnaQ6G5WL0klYlRjMizCBzZSme4_-VNrRnP35lhef3yzIxyvzjaG_NnerWeFrtlgHf80g4VnmP0DS2G32g",
> "user.swift.metadata", 0x0, 0) = 207****
>
getxattr("/mnt/gluster-object/AUTH_8140ab2e72bd3fd9d3e89e00d2d2c620/blocks10/5/V/m/5VmTFnaQ6G5WL0klYlRjMizCBzZSme4_-VNrRnP35lhef3yzIxyvzjaG_NnerWeFrtlgHf80g4VnmP0DS2G32g",
> "user.swift.metadata",
>
"\x80\x02}q\x01(U\x10X-Content-Lengthq\x02U\x0537768U\x0eX-Content-Typeq\x03U\x18application/octet-streamq\x04U\x0bX-Timestampq\x05U\x101330097164.20775q\x06U\x0dX-Object-Typeq\x07U\x04fileq\x08U\x06X-Typeq\x09U\x06Objectq\x0aU\x06X-ETagq\x0bU
> 837ed1a98765ac2d471f499f5c44f618q\x0cu.", 207) = 207****
>
getxattr("/mnt/gluster-object/AUTH_8140ab2e72bd3fd9d3e89e00d2d2c620/blocks10/5/V/m/5VmTFnaQ6G5WL0klYlRjMizCBzZSme4_-VNrRnP35lhef3yzIxyvzjaG_NnerWeFrtlgHf80g4VnmP0DS2G32g",
> "user.swift.metadata1", 0x0, 0) = -1 ENODATA (No data
available)****
>
stat("/mnt/gluster-object/AUTH_8140ab2e72bd3fd9d3e89e00d2d2c620/blocks10/5/V/m/5VmTFnaQ6G5WL0klYlRjMizCBzZSme4_-VNrRnP35lhef3yzIxyvzjaG_NnerWeFrtlgHf80g4VnmP0DS2G32g",
> {st_mode=S_IFREG|0600, st_size=37768, ...}) = 0****
>
lstat("/mnt/gluster-object/AUTH_8140ab2e72bd3fd9d3e89e00d2d2c620/blocks10/5/o",
> {st_mode=S_IFDIR|0777, st_size=44, ...}) = 0****
>
open("/mnt/gluster-object/AUTH_8140ab2e72bd3fd9d3e89e00d2d2c620/blocks10/5/o",
> O_RDONLY|O_NONBLOCK|O_DIRECTORY|O_CLOEXEC) = 11****
> getdents(11, /* 4 entries */, 32768) = 96****
> getdents(11, /* 0 entries */, 32768) = 0****
> close(11) = 0****
>
stat("/mnt/gluster-object/AUTH_8140ab2e72bd3fd9d3e89e00d2d2c620/blocks10/5/o/A",
> {st_mode=S_IFDIR|0777, st_size=105, ...}) = 0****
>
stat("/mnt/gluster-object/AUTH_8140ab2e72bd3fd9d3e89e00d2d2c620/blocks10/5/o/t",
> {st_mode=S_IFDIR|0777, st_size=198, ...}) = 0****
>
getxattr("/mnt/gluster-object/AUTH_8140ab2e72bd3fd9d3e89e00d2d2c620/blocks10/5/o/A",
> "user.swift.metadata", 0x0, 0) = 194****
>
getxattr("/mnt/gluster-object/AUTH_8140ab2e72bd3fd9d3e89e00d2d2c620/blocks10/5/o/A",
> "user.swift.metadata",
>
"\x80\x02}q\x01(U\x10X-Content-Lengthq\x02K\x00U\x0eX-Content-Typeq\x03U\x15application/directoryq\x04U\x0bX-Timestampq\x05U\x101330097187.00000U\x0dX-Object-Typeq\x06U\x03dirq\x07U\x06X-Typeq\x08U\x06Objectq\x09U\x06X-ETagq\x0aU
> d41d8cd98f00b204e9800998ecf8427eu.", 194) = 194****
>
getxattr("/mnt/gluster-object/AUTH_8140ab2e72bd3fd9d3e89e00d2d2c620/blocks10/5/o/A",
> "user.swift.metadata1", 0x0, 0) = -1 ENODATA (No data
available)****
>
getxattr("/mnt/gluster-object/AUTH_8140ab2e72bd3fd9d3e89e00d2d2c620/blocks10/5/o/t",
> "user.swift.metadata", 0x0, 0) = 194****
>
getxattr("/mnt/gluster-object/AUTH_8140ab2e72bd3fd9d3e89e00d2d2c620/blocks10/5/o/t",
> "user.swift.metadata",
>
"\x80\x02}q\x01(U\x10X-Content-Lengthq\x02K\x00U\x0eX-Content-Typeq\x03U\x15application/directoryq\x04U\x0bX-Timestampq\x05U\x101330097357.00000U\x0dX-Object-Typeq\x06U\x03dirq\x07U\x06X-Typeq\x08U\x06Objectq\x09U\x06X-ETagq\x0aU
> d41d8cd98f00b204e9800998ecf8427eu.", 194) = 194****
>
getxattr("/mnt/gluster-object/AUTH_8140ab2e72bd3fd9d3e89e00d2d2c620/blocks10/5/o/t",
> "user.swift.metadata1", 0x0, 0) = -1 ENODATA (No data
available)****
>
lstat("/mnt/gluster-object/AUTH_8140ab2e72bd3fd9d3e89e00d2d2c620/blocks10/5/o/A",
> {st_mode=S_IFDIR|0777, st_size=105, ...}) = 0****
>
open("/mnt/gluster-object/AUTH_8140ab2e72bd3fd9d3e89e00d2d2c620/blocks10/5/o/A",
> O_RDONLY|O_NONBLOCK|O_DIRECTORY|O_CLOEXEC) = 11****
> getdents(11, /* 2 entries */, 32768) = 48****
> getdents(11, /* 1 entries */, 32768) = 112****
> getdents(11, /* 0 entries */, 32768) = 0****
> close(11) = 0****
>
stat("/mnt/gluster-object/AUTH_8140ab2e72bd3fd9d3e89e00d2d2c620/blocks10/5/o/A/5oAs3UlU-Ah3ml--FMrmZfsx_G-QXDXeSh1ZJ6vnjLnJy6iJrLZO8gLYTnSuWh5hEsNlaIbqB-J6iHaEwQJQ8A",
> {st_mode=S_IFREG|0600, st_size=8189, ...}) = 0****
>
getxattr("/mnt/gluster-object/AUTH_8140ab2e72bd3fd9d3e89e00d2d2c620/blocks10/5/o/A/5oAs3UlU-Ah3ml--FMrmZfsx_G-QXDXeSh1ZJ6vnjLnJy6iJrLZO8gLYTnSuWh5hEsNlaIbqB-J6iHaEwQJQ8A",
> "user.swift.metadata", 0x0, 0) = 203****
>
getxattr("/mnt/gluster-object/AUTH_8140ab2e72bd3fd9d3e89e00d2d2c620/blocks10/5/o/A/5oAs3UlU-Ah3ml--FMrmZfsx_G-QXDXeSh1ZJ6vnjLnJy6iJrLZO8gLYTnSuWh5hEsNlaIbqB-J6iHaEwQJQ8A",
> "user.swift.metadata",
>
"\x80\x02}q\x01(U\x10X-Content-Lengthq\x02U\x010U\x0eX-Content-Typeq\x03U\x18application/octet-streamq\x04U\x0bX-Timestampq\x05U\x101330097187.91304q\x06U\x0dX-Object-Typeq\x07U\x04fileq\x08U\x06X-Typeq\x09U\x06Objectq\x0aU\x06X-ETagq\x0bU
> e185fccc1ed0dbd69b304726ff54d294q\x0cu.", 203) = 203****
>
getxattr("/mnt/gluster-object/AUTH_8140ab2e72bd3fd9d3e89e00d2d2c620/blocks10/5/o/A/5oAs3UlU-Ah3ml--FMrmZfsx_G-QXDXeSh1ZJ6vnjLnJy6iJrLZO8gLYTnSuWh5hEsNlaIbqB-J6iHaEwQJQ8A",
> "user.swift.metadata1", 0x0, 0) = -1 ENODATA (No data
available)****
>
stat("/mnt/gluster-object/AUTH_8140ab2e72bd3fd9d3e89e00d2d2c620/blocks10/5/o/A/5oAs3UlU-Ah3ml--FMrmZfsx_G-QXDXeSh1ZJ6vnjLnJy6iJrLZO8gLYTnSuWh5hEsNlaIbqB-J6iHaEwQJQ8A",
> {st_mode=S_IFREG|0600, st_size=8189, ...}) = 0****
>
lstat("/mnt/gluster-object/AUTH_8140ab2e72bd3fd9d3e89e00d2d2c620/blocks10/5/o/t",
> {st_mode=S_IFDIR|0777, st_size=198, ...}) = 0****
>
open("/mnt/gluster-object/AUTH_8140ab2e72bd3fd9d3e89e00d2d2c620/blocks10/5/o/t",
> O_RDONLY|O_NONBLOCK|O_DIRECTORY|O_CLOEXEC) = 11****
> getdents(11, /* 3 entries */, 32768) = 160****
> getdents(11, /* 0 entries */, 32768) = 0
> ...
>
>
> Is this the normal behavior for the UFO API?
>
> Best regards,
> Andr? Cruz
> **
>
> _______________________________________________
> Gluster-users mailing list
> Gluster-users at gluster.org
> http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://supercolony.gluster.org/pipermail/gluster-users/attachments/20120228/dcabc265/attachment.html>