date:20170322

Re: [Gluster-devel] glusterd regression failure on centos

2017-03-22 Thread Ravishankar N


On 03/22/2017 11:54 PM, Atin Mukherjee wrote:
Please file a bug in project-infra in gluster asking for a centos 
slave machine to debug the issue further and Nigel should be able to 
assist you on that.


On Wed, 22 Mar 2017 at 13:55, Gaurav Yadav > wrote:


Hi All,

glusterd regression is getting failed while executing
"tests/basic/afr/arbiter-mount.t "test case.

Test Summary Report
*14:16:27* ---
*14:16:27* ./tests/basic/afr/arbiter-mount.t (Wstat: 0 Tests: 22 Failed: 4)
*14:16:27*Failed tests:  7, 17, 21-22
*14:16:27* Files=1, Tests=22, 71 wallclock secs ( 0.03 usr  0.01 sys +  
1.54 cusr  2.42 csys =  4.00 CPU)
*14:16:27* Result: FAIL
*14:16:27* End of test ./tests/basic/afr/arbiter-mount.t

Here is the link of logs generated by jenkins :-
https://build.gluster.org/job/centos6-regression/3732/consoleFull


"EXPECT_WITHIN $NFS_EXPORT_TIMEOUT "1" is_nfs_export_available"  is failing.
Looks like rpcbind was not running on the slave. From the regression 
log, the `cleanup` which is called before the TESTs in the .t are run is 
spewing out some errors:


*4:14:04* [14:14:04] Running tests in file ./tests/basic/afr/arbiter-mount.t
*14:14:47* rm: cannot remove `/mnt/glusterfs/0/xy_zzy': Transport endpoint is 
not connected
*14:14:47* mount.nfs: rpc.statd is not running but is required for remote 
locking.
*14:14:47* mount.nfs: Either use '-o nolock' to keep locks local, or start 
statd.
*14:14:47* mount.nfs: an incorrect mount option was specified
*14:15:10* mount.nfs: rpc.statd is not running but is required for remote 
locking.
*14:15:10* mount.nfs: Either use '-o nolock' to keep locks local, or start 
statd.
*14:15:10* mount.nfs: an incorrect mount option was specified


-Ravi


I executed tests/basic/afr/arbiter-mount.t script explicitly, but this 
test-case passed for me.

prove tests/basic/afr/arbiter-mount.t
tests/basic/afr/arbiter-mount.t .. 9/22 rm: cannot remove 
'/mnt/glusterfs/0/xy_zzy': Transport endpoint is not connected
tests/basic/afr/arbiter-mount.t .. 10/22 mount.nfs: Remote I/O error
tests/basic/afr/arbiter-mount.t .. ok
All tests successful.
Files=1, Tests=22, 55 wallclock secs ( 0.03 usr  0.00 sys +  0.69 cusr  
0.55 csys =  1.27 CPU)
Result: PASS


Thanks
Gaurav

___
Gluster-devel mailing list
Gluster-devel@gluster.org 
http://lists.gluster.org/mailman/listinfo/gluster-devel

--
- Atin (atinm)


___
Gluster-devel mailing list
Gluster-devel@gluster.org
http://lists.gluster.org/mailman/listinfo/gluster-devel



___
Gluster-devel mailing list
Gluster-devel@gluster.org
http://lists.gluster.org/mailman/listinfo/gluster-devel

Re: [Gluster-devel] glusterd regression failure on centos

2017-03-22 Thread Atin Mukherjee

Please file a bug in project-infra in gluster asking for a centos slave
machine to debug the issue further and Nigel should be able to assist you
on that.

On Wed, 22 Mar 2017 at 13:55, Gaurav Yadav  wrote:

> Hi All,
>
> glusterd regression is getting failed while executing
> "tests/basic/afr/arbiter-mount.t "test case.
>
>  Test Summary Report*14:16:27* ---*14:16:27* 
> ./tests/basic/afr/arbiter-mount.t (Wstat: 0 Tests: 22 Failed: 4)*14:16:27*   
> Failed tests:  7, 17, 21-22*14:16:27* Files=1, Tests=22, 71 wallclock secs ( 
> 0.03 usr  0.01 sys +  1.54 cusr  2.42 csys =  4.00 CPU)*14:16:27* Result: 
> FAIL*14:16:27* End of test ./tests/basic/afr/arbiter-mount.t
>
> Here is the link of logs generated by jenkins :-
> https://build.gluster.org/job/centos6-regression/3732/consoleFull
>
>
> I executed tests/basic/afr/arbiter-mount.t script explicitly, but this 
> test-case passed for me.
>
> prove tests/basic/afr/arbiter-mount.t
> tests/basic/afr/arbiter-mount.t .. 9/22 rm: cannot remove 
> '/mnt/glusterfs/0/xy_zzy': Transport endpoint is not connected
> tests/basic/afr/arbiter-mount.t .. 10/22 mount.nfs: Remote I/O error
> tests/basic/afr/arbiter-mount.t .. ok
> All tests successful.
> Files=1, Tests=22, 55 wallclock secs ( 0.03 usr  0.00 sys +  0.69 cusr  0.55 
> csys =  1.27 CPU)
> Result: PASS
>
>
> Thanks
> Gaurav
>
> ___
> Gluster-devel mailing list
> Gluster-devel@gluster.org
> http://lists.gluster.org/mailman/listinfo/gluster-devel

-- 
- Atin (atinm)
___
Gluster-devel mailing list
Gluster-devel@gluster.org
http://lists.gluster.org/mailman/listinfo/gluster-devel

[Gluster-devel] glusterd regression failure on centos

2017-03-22 Thread Gaurav Yadav

Hi All,

glusterd regression is getting failed while executing
"tests/basic/afr/arbiter-mount.t "test case.

 Test Summary Report*14:16:27* ---*14:16:27*
./tests/basic/afr/arbiter-mount.t (Wstat: 0 Tests: 22 Failed:
4)*14:16:27*   Failed tests:  7, 17, 21-22*14:16:27* Files=1,
Tests=22, 71 wallclock secs ( 0.03 usr  0.01 sys +  1.54 cusr  2.42
csys =  4.00 CPU)*14:16:27* Result: FAIL*14:16:27* End of test
./tests/basic/afr/arbiter-mount.t

Here is the link of logs generated by jenkins :-
https://build.gluster.org/job/centos6-regression/3732/consoleFull


I executed tests/basic/afr/arbiter-mount.t script explicitly, but this
test-case passed for me.

prove tests/basic/afr/arbiter-mount.t
tests/basic/afr/arbiter-mount.t .. 9/22 rm: cannot remove
'/mnt/glusterfs/0/xy_zzy': Transport endpoint is not connected
tests/basic/afr/arbiter-mount.t .. 10/22 mount.nfs: Remote I/O error
tests/basic/afr/arbiter-mount.t .. ok
All tests successful.
Files=1, Tests=22, 55 wallclock secs ( 0.03 usr  0.00 sys +  0.69 cusr
 0.55 csys =  1.27 CPU)
Result: PASS


Thanks
Gaurav
___
Gluster-devel mailing list
Gluster-devel@gluster.org
http://lists.gluster.org/mailman/listinfo/gluster-devel

Re: [Gluster-devel] Gluster volume snapshot - Invitation to edit

2017-03-22 Thread Mark Ferrell

I currently have 2 branches related to snapshots on GlusterFS.

https://github.com/major0/glusterfs/tree/lvm-snapshot-cleanup
Based code from Sriram that I found on GitHub, though I removed the ZFS
code from Sriram's patches and created this LVM-only branch. Subsequent
changes to this branch where done i response to changes necessary to
implement support for btrfs snapshots (and cleanup a memory leak in the
missed-snapshot code)...


https://github.com/major0/glusterfs/tree/btrfs-snapshots
This code is branched from my lvm-snapshot-cleanup branch and is directly
dependant on the changes contained there-in.

Both of these branches are currently in-flux, and I have been performing
regular rebasing against the GlusterFS master branch.

Btrfs snapshots currently work and have been tested on Ubuntu 16.04, though
the process for configuring btrfs for gluster snapshots has not yet been
documented.

Required interfaces to support hard-coded snapshot support can be found:
https://github.com/major0/glusterfs/blob/lvm-snapshot-cleanup/xlators/mgmt/glusterd/src/snapshot/glusterd-lvm-snapshot.h

It should be relatively straight-forward to package up these interfaces
into a structure of function pointers and allow glusterd to itterate the
array firing off the glusterd_is__brick() function at the path to
figure out which subsystem to use.
___
Gluster-devel mailing list
Gluster-devel@gluster.org
http://lists.gluster.org/mailman/listinfo/gluster-devel

Re: [Gluster-devel] Gluster volume snapshot - Plugin architecture proposal

2017-03-22 Thread sriram

Hi Amar, 



On Wed, Mar 22, 2017, at 04:26 AM, Amar Tumballi wrote:

> Hi Sriram,

> Thanks for sharing this. Just one comment below.

> 

> On Tue, Mar 21, 2017 at 10:12 AM,  wrote:

>> __

>> Hi Raghavendra, 

>> 

>> My name is Sriram I'd been working with Rajesh on creating a plugin
>> structure for snapshot functionality. Below is the document which
>> Rajesh'd created and I've edited the same with ideas and problems.
>> Could you have a look and review so that we could take it forward?
>> 

>> https://docs.google.com/document/d/1dHij_oy8V8CF2I7WfdYqKXFTGw0SKGzGlngDBpSwSGc/edit?usp=sharing
>> 

> 

> I am not sure if any 'code' has been written already for this. If not,
> great, because we want any improvements in 'glusterd' space to come
> through approval of new GlusterD[1] design phase, so we can move away
> with the current glusterd.
> But, looking at the design, the plugin can be pretty much be
> independent of the glusterd architecture, even then better to run it
> through everyone involved first.


There was a series of patches which I'd sent initially but, then Rajesh
and Avra proposed that we revisit the design once. So we'd stopped
writing any code.


https://review.gluster.org/#/c/16138/ This was the series I'd posted.  I'm not 
really sure who'd be helping on this activity to get them involved in the mail 
loop. I'll have a look at the GlusterD design phase approval wiki. Let me know 
as how we proceed on this. 


Sriram





> 

> If already significant effort has been made here, then we are fine
> with continuing the effort to take these to completion.
> 

> Regards,

> Amar

>  

>> 

>> Sriram

>> 

>>> 

>>> On Thu, Mar 16, 2017, at 11:02 AM, Rajesh Joseph wrote:

 Hi Sriram,

 I already took a look at the document and added some minor
 comments. As you are aware Raghavendra Bhatt is the new maintainer
 so he will be more involved in this. I may not be able to spend as
 much time as I would like. But I will try to help as much as
 possible.
 Regards,

 Rajesh

 

 On Thu, Mar 16, 2017 at 10:41 AM,  wrote:

> __

> Hi Rajesh, 

> 

> You'd a look at the document? Would you be involved in this
> activity, going forward?
> 

> Sriram

> 

> 

> On Fri, Mar 10, 2017, at 05:49 PM, sri...@marirs.net.in wrote:

>> Hi Rajesh, 

>> 

>> Thank you. I've replied on this mail and added gluster-devel. Let
>> me know your thoughts and if you could share the same idea with
>> Raghavendra Bhat who is the new maintainer it'd be great for me
>> to take this forward with him.
>> 

>> Sriram

>> 

>> 

>> On Fri, Mar 10, 2017, at 07:48 AM, Rajesh Joseph wrote:

>>> Hi Sriram,

>>> Sorry could not look into this as I was busy. Also I am leaving
>>> Red Hat so may not be able to spend much time on this. I think
>>> it would be great if you include gluster-devel in this
>>> discussion. Meanwhile I will try to review the doc in a day or
>>> two.
>>> 

>>> Thanks & Regards,

>>> Rajesh

>>> 

>>> On Mon, Mar 6, 2017 at 11:27 AM,  wrote:

 __

 Hi Rajesh, 

 

 Could you have a look at the below doc and let me know about
 it?
 

 Sriram

 

 

 On Mon, Feb 27, 2017, at 12:44 PM, sri...@marirs.net.in wrote:
> Hi Rajesh, 

> 

> Could you have a look at the below link and let me know your
> comments?
> 

> Sriram

> 

> 

> On Mon, Feb 20, 2017, at 11:31 AM, sri...@marirs.net.in wrote:
>> Hi Rajesh/Avra,

>> 

>> https://docs.google.com/document/d/1dHij_oy8V8CF2I7WfdYqKXFTGw0SKGzGlngDBpSwSGc/edit?usp=sharing
>> 

>> I've updated the doc with details on how we could design.
>> Sorry about the delay, was held up at work. Could you have a
>> look at let me know the comments?
>> 

>> Sriram

>> 

>> 

>> On Mon, Feb 6, 2017, at 12:24 PM, Rajesh Joseph wrote:

>>> Done.

>>> 

>>> On Mon, Feb 6, 2017 at 12:08 PM, 
>>> wrote:
 __

 Hi Rajesh, 

 

 Was able to check-in last week without a logon, but now
 unable to. Could you permit "sriram@gmail.com" this is
 my gmail account to use the doc shared?
 

 Sriram

 

 

 On Wed, Jan 18, 2017, at 03:42 PM, Rajesh Joseph (via
 Google Docs) wrote:
> 

> 

> Rajesh Joseph[2] has invited you to *edit* the following
> document:
> 

> 

>  

> Gluster volume snapshot[3]

> 

>

Re: [Gluster-devel] Nodeid changed due to write-behind option chagned online will lead to un-expected umount by kernel

2017-03-22 Thread Lian, George (Nokia - CN/Hangzhou)

Hi, Csaba,

Could you please give some comments for this issues?

Thanks & Best Regards,
George

-Original Message-
From: Raghavendra Gowdappa [mailto:rgowd...@redhat.com] 
Sent: Monday, March 20, 2017 9:35 PM
To: Lian, George (Nokia - CN/Hangzhou) 
Cc: Zhang, Bingxuan (Nokia - CN/Hangzhou) ; 
Gluster-devel@gluster.org; Venetjoki, Tero (Nokia - FI/Espoo) 
; Zhou, Cynthia (Nokia - CN/Hangzhou) 
; Csaba Henk 
Subject: Re: [Gluster-devel] Nodeid changed due to write-behind option chagned 
online will lead to un-expected umount by kernel

+csaba

- Original Message -
> From: "Raghavendra Gowdappa" 
> To: "George Lian (Nokia - CN/Hangzhou)" 
> Cc: "Bingxuan Zhang (Nokia - CN/Hangzhou)" , 
> Gluster-devel@gluster.org, "Tero Venetjoki
> (Nokia - FI/Espoo)" , "Cynthia Zhou (Nokia - 
> CN/Hangzhou)" 
> Sent: Monday, March 20, 2017 7:02:50 PM
> Subject: Re: [Gluster-devel] Nodeid changed due to write-behind option 
> chagned online will lead to un-expected umount
> by kernel
> 
> 
> 
> - Original Message -
> > From: "George Lian (Nokia - CN/Hangzhou)" 
> > To: "Raghavendra Gowdappa" , Gluster-devel@gluster.org
> > Cc: "Bingxuan Zhang (Nokia - CN/Hangzhou)" ,
> > "Cynthia Zhou (Nokia - CN/Hangzhou)"
> > , "Tero Venetjoki (Nokia - FI/Espoo)"
> > 
> > Sent: Monday, March 20, 2017 8:44:30 AM
> > Subject: Nodeid changed due to write-behind option chagned online will lead
> > to un-expected umount by kernel
> > 
> > 
> > Hi, GlusterFS expert,
> > 
> > In our latest test, We found an issue potentially related to glusterfs.
> > When
> > I execute “gluster volume set  performance.write-behind on/off”,
> > some bind mounts will get lost, this issue is permanent!
> > Test steps:
> > a)  mkdir -p /mnt/log/node1/test1/test2; mkdir -p /mnt/log/node2/test1/test2
> > b)  mkdir -p /mnt/test1/test2
> > c)  mount –bind /mnt/log/node1/test1 /mnt/test1
> > d)  mount –bind /mnt/log/node2/test1/test2 /mnt/test1/test2
> > e)  mount | cut -d " " -f 3|xargs stat
> > f)  gluster volume set log performance.write-behind on/off    check
> > /mnt/test1/test2 , could found this bind mount get lost
> > 
> > We’ve consulted the linux kernel side guys, their explanations for this
> > “bind
> > mount lost” issue is when kernel side do lookup or stat(), it will find the
> > nodeid has changed and trigger fuse_dentry_revalidate, and
> > fuse_dentry_relivalidate() fails for some dentry and invalidates it which
> > leads to unmounting.
> > 
> >88.912262 |   0)stat-437|   |
> >fuse_dentry_revalidate [fuse]() {
> > <...>
> >88.912264 |   0)stat-437|   |
> >fuse_simple_request [fuse]() {
> > <...>
> >88.921383 |   0)stat-437| # 9119.255 us |  } /*
> >fuse_simple_request [fuse] */
> >88.921383 |   0)stat-437|   0.093 us|
> >dput();
> >88.921384 |   0)stat-437|   |
> >fuse_queue_forget [fuse]() {
> > <...>
> >88.921427 |   0)stat-437| + 42.737 us   |  }
> >88.921427 |   0)stat-437| # 9164.967 us |} /*
> >fuse_dentry_revalidate [fuse] */
> >88.921427 |   0)stat-437|   |
> >d_invalidate() {
> >88.921427 |   0)stat-437|   0.048 us|
> >_raw_spin_lock();
> >88.921428 |   0)stat-437|   |
> >d_walk() {
> >88.921428 |   0)stat-437|   0.040 us|
> >_raw_spin_lock();
> >88.921428 |   0)stat-437|   0.046 us|
> >detach_and_collect();
> >88.921429 |   0)stat-437|   0.723 us|  }
> >88.921429 |   0)stat-437|   |
> >__detach_mounts() {
> > 
> > fuse_queue_forget() got called inside fuse_dentry_revalidate(), so the
> > failure is clearly caused by  (outarg.nodeid != get_node_id(inode)) check:
> > 
> > 7078187a795f8 (Miklos Szeredi 2014-12-12 09:49:05 +0100  230)
> > if (!ret) {
> > 6314efee3cfee (Miklos Szeredi 2013-10-01 16:41:22 +0200  231)
> > fi = get_fuse_inode(inode);
> > 9e6268db496a2 (Miklos Szeredi 2005-09-09 13:10:29 -0700  232)
> > if (outarg.nodeid != get_node_id(inode)) {
> > 07e77dca8a1f1 (Miklos Szeredi 2010-12-07 20:16:56 +0100  233)
> > fuse_queue_forget(fc, forget, outarg.nodeid,
> > 1);
> > e2a6b95236eba (Miklos Szeredi 2013-09-05 11:44:43 +0200  234)
> > goto invalid;
> > 9e6268db496a2 (Miklos Szeredi 2005-09-09 13:10:29 -0700  235)
> > }
> > 
> > The latest  kernel is doing the following in fs/fuse/dir.c:
> > 
> > if (outarg.nodeid != get_node_id(inode)) {
> >

Re: [Gluster-devel] About inode table: client, server and inconsistency

2017-03-22 Thread Tahereh Fattahi

a university project!

On Sun, Mar 19, 2017 at 11:31 PM, Nithya Balachandran 
wrote:

>
>
> On 18 March 2017 at 21:42, Tahereh Fattahi  wrote:
>
>> Thank you very much.
>> Is it possible to change something in server inode table during a fop
>> from client? (I want to change the dht_layout of a directory when create a
>> file in that directory, but I dont know how send the changed layout to
>> servers)
>>
>
> Why do you want  to change the layout when files are created?
>
> Regards,
> Nithya
>
>>
>> On Sat, Mar 18, 2017 at 6:36 PM, Amar Tumballi 
>> wrote:
>>
>>>
>>>
>>> On Thu, Mar 16, 2017 at 10:30 PM, Tahereh Fattahi >> > wrote:
>>>
 Hi
 Is it correct that each brick has one inode table for itself and each
 client has one inode table that stores anything that is stored in bricks
 inode table?

 For a given inode, the contents on client side and server side would be
>>> very much different between how the volume graph is.
>>>
>>>

 Does all inode tables store in RAM all the time?

>>>
>>> Client (mainly fuse) inode table will be in memory all the time, until
>>> kernel sends a FORGET. Brick side we have limited number of inodes in
>>> memory. (There is an option called 'lru-limit').
>>>
>>>


 When and how client's inode table update (how solve the inconsistency
 problem between clinet and brick inode table that is because of rebalance
 or other client fops) ?


>>> All the translators are designed to handle the consistency check in
>>> their 'lookup()' code, and they should send a response up with error saying
>>> its a stale inode (ESTALE), upon receiving which, the client inode table
>>> refreshes its inode, and does a fresh lookup again. This allows us to keep
>>> the inode table in consistency.
>>>
>>> Hope that answers the question.
>>>
>>> -Amar
>>>
>>>
>>>
 ___
 Gluster-devel mailing list
 Gluster-devel@gluster.org
 http://lists.gluster.org/mailman/listinfo/gluster-devel

>>>
>>>
>>>
>>> --
>>> Amar Tumballi (amarts)
>>>
>>
>>
>> ___
>> Gluster-devel mailing list
>> Gluster-devel@gluster.org
>> http://lists.gluster.org/mailman/listinfo/gluster-devel
>>
>
>
___
Gluster-devel mailing list
Gluster-devel@gluster.org
http://lists.gluster.org/mailman/listinfo/gluster-devel

Re: [Gluster-devel] glusterd regression failure on centos

Re: [Gluster-devel] glusterd regression failure on centos

[Gluster-devel] glusterd regression failure on centos

Re: [Gluster-devel] Gluster volume snapshot - Invitation to edit

Re: [Gluster-devel] Gluster volume snapshot - Plugin architecture proposal

Re: [Gluster-devel] Nodeid changed due to write-behind option chagned online will lead to un-expected umount by kernel

Re: [Gluster-devel] About inode table: client, server and inconsistency

7 matches

Site Navigation

Mail list logo

Footer information