LISTSERV mailing list manager LISTSERV 16.5

Help for XROOTD-L Archives


XROOTD-L Archives

XROOTD-L Archives


XROOTD-L@LISTSERV.SLAC.STANFORD.EDU


View:

Message:

[

First

|

Previous

|

Next

|

Last

]

By Topic:

[

First

|

Previous

|

Next

|

Last

]

By Author:

[

First

|

Previous

|

Next

|

Last

]

Font:

Proportional Font

LISTSERV Archives

LISTSERV Archives

XROOTD-L Home

XROOTD-L Home

XROOTD-L  December 2019

XROOTD-L December 2019

Subject:

Re: Clustering (filecache) servers

From:

Andrew Hanushevsky <[log in to unmask]>

Reply-To:

Support use of xrootd by HEP experiments <[log in to unmask]>

Date:

Wed, 4 Dec 2019 15:07:59 -0800

Content-Type:

TEXT/PLAIN

Parts/Attachments:

Parts/Attachments

TEXT/PLAIN (391 lines)

Hi Sam,

Can I connect to it or is i behind a firewall? If I can connect do I need 
any credentials at least for he redirector? If it's not redirecting then 
it means it's in suspend status. Here are the reasons a redirector can 
suspend itself:

1) The number of nodes connected to the redirector has fallen below the 
mnimum required (configurable). The default is good for lage clusters but 
not for small ones. See, "delay servers" in:
https://xrootd.slac.stanford.edu/doc/dev410/cms_config.htm#_Toc8247276

2) All nodes connected to the redirector have suspended themselves. This 
is usually caused by the xrootd being unable to connect to it's twin cmsd. 
The xrootd log will show why. You can figure this out by greping the 
redirector log for "suspend" using '-i' (case insenstive).

You can also do something as follows to ascertain the status of your 
cluster:

xrdfs <redirector> 'locate *'

which will tell you the state of things at the redirector. Might 
interesting to see what it says.

Andy

On Wed, 4 Dec 2019, Sam Skipsey wrote:

> Hi Andy,
>
> The config is literally copied from the boilerplate example in the
> Disk Caching Proxy Cluster section in the documentation :) - I did the
> replication because it's how the example does it.
>
> It is entirely correct that there is just one server for now - I was
> building up the system from a single server once it actually worked.
> (If you read up, you will see that in this case, it's not even
> redirecting when there's just 1 server to consider.)
>
> Sam
>
> On Wed, 4 Dec 2019 at 22:38, Andrew Hanushevsky <[log in to unmask]> wrote:
>>
>> Hi Sam,
>>
>> Well, xrdmapc shows a redirector (cephs03) with one server (cephc01). I
>> sthat not correct? I assume you expected more nodes, yes? The config file
>> looks reasonable, though i doesn't have to repeat so much information.
>> Also, for consistency oss.localroot should be the sameverywhere (though
>> as specified it's essentially the same).
>>
>> Andy
>>
>> On Wed, 4 Dec 2019, Sam Skipsey wrote:
>>
>>> Hi Matevz,
>>>
>>> So:
>>>
>>> [root@cephs03 ~]# xrdmapc cephs03:1094
>>> 0**** cephs03.beowulf.cluster:1094
>>>      Srv cephc01.beowulf.cluster:1094
>>>
>>>
>>> I agree, though, that it looks like the cmsd and xrootd aren't
>>> properly connecting - I see none of the expected stuff in the xrootd
>>> log on the redirector.
>>>
>>> The config file (shared between the redirector and the server) is:
>>>
>>> #Clustered cache config (redirector also does authentication with lcmaps)
>>>
>>> all.manager cephs03.beowulf.cluster:1213
>>> all.export /xrootd:/ stage r/o
>>> all.export /root:/ stage r/o
>>> all.export * stage r/o
>>>
>>> if cephs03+
>>>
>>> all.role manager
>>> all.export /xrootd:/ stage r/o
>>> all.export /root:/ stage r/o
>>> all.export * stage r/o
>>>
>>> xrootd.trace emsg login stall redirect
>>> xrd.trace conn
>>> cms.trace defer files redirect stage
>>>
>>>
>>>
>>>
>>> #the cmsd on the cache node
>>> else if exec cmsd
>>>
>>> all.role server
>>> all.export /xrootd:/ stage
>>> all.export /root:/ stage
>>> all.export * stage
>>> oss.localroot /cache
>>>
>>> #the xrootd on the cache node
>>> else
>>>
>>> all.role server
>>>
>>> ofs.osslib    libXrdPss.so
>>> pss.cachelib  libXrdFileCache.so
>>> pfc.ram      16g
>>> pfc.trace     info
>>> pfc.diskusage 0.90 0.95
>>> oss.localroot /cache/
>>> cms.trace defer redirect stage
>>>
>>> all.export /xroot:/ r/w
>>> all.export /root:/ r/w
>>> all.export * r/w
>>>
>>> pss.origin localhost:1095
>>>
>>> xrd.allow host *.beowulf.cluster
>>>
>>> #inbound security protocol, for authenticating to the xrootd-ceph
>>> #sec.protocol
>>> xrootd.seclib /opt/xrootd/lib64/libXrdSec.so
>>> sec.protocol sss -s /etc/gateway/xrootd/sss.keytab.grp -c
>>> /etc/gateway/xrootd/sss.keytab.grp
>>> sec.protbind localhost:1095 only sss
>>>
>>> xrd.report 127.0.0.1:9527 every 5s all
>>>
>>>
>>> #try changing some pss options to tune this connection
>>>
>>> pss.setopt DebugLevel 3
>>> pss.setopt ParStreamsPerPhyConn 15
>>> pfc.blocksize 16M
>>>
>>>
>>> fi
>>>
>>>
>>> [Note that the shared secret is the same for all parts of the cluster,
>>> at the moment]
>>> At various points, I've messed with bits of the file (removing some of
>>> the stages, or even entire exports, and a bunch of other things), but
>>> this is the current config - using the default setup derived from the
>>> Proxy config documentation pg 38, and adding your debugging lines.
>>>
>>> Sam
>>>
>>>
>>>
>>> On Tue, 3 Dec 2019 at 20:37, Matevz Tadel <[log in to unmask]> wrote:
>>>>
>>>> On 2019-12-03 06:38, Sam Skipsey wrote:
>>>>> Okay, so what I have, including logs from all the relevant bits is below.
>>>>>
>>>>> I restarted all of the services (after adding the traces as requested)
>>>>> at around 14:30, so all of the servers come up.
>>>>> About 2 minutes later, I tried (from cephc01, using xrdcp) to copy a
>>>>> file via the redirector xrootd (on cephs03). This only seems to show
>>>>> up on the redirector xrootd itself - there's no record of anything in
>>>>> the other logs as far as I can see.
>>>>
>>>> Hmmh, this is stange, cmsd on redirector never seems to ask the servers.
>>>>
>>>> redirector cmsd:
>>>> 191203 11:57:35 51827 Select seeking /store/data/Run2018D/MET/MINIAOD/PromptReco-v2/000/322/431/00000/4AD1B758-3BB8-E811-92B8-FA163EA98227.root
>>>> 191203 11:57:35 51827 redirector.51097:17@xrootd XrdLink: Setting ref to 2+-1 post=0
>>>> 191203 11:57:35 51475 Dispatch server.6875:24@xcache-05:1094 for have dlen=107
>>>> 191203 11:57:35 51475 server.6875:24@xcache-05 XrdLink: Setting ref to 1+1 post=0
>>>> 191203 11:57:35 51341 XrdSched: running server inq=0
>>>> 191203 11:57:35 51341 server.6875:24@xcache-05:1094 do_Have: /store/data/Run2018D/MET/MINIAOD/PromptReco-v2/000/322/431/00000/4AD1B758-3BB8-E811-92B8-FA163EA98227.root
>>>>
>>>> redirector xrootd:
>>>> 191203 11:57:35 51514 nobody.351:[log in to unmask] XrootdProtocol: 0100 req=open dlen=106
>>>> 191203 11:57:35 51514 nobody.351:[log in to unmask] XrootdProtocol: 0100 open r /store/data/Run2018D/MET/MINIAOD/PromptReco-v2/000/322/431/00000/4AD1B758-3BB8-E811-92B8-FA163EA98227.root
>>>> 191203 11:57:35 51514 nobody.351:[log in to unmask] ofs_open: 0-660 fn=/store/data/Run2018D/MET/MINIAOD/PromptReco-v2/000/322/431/00000/4AD1B758-3BB8-E811-92B8-FA163EA98227.root
>>>> 191203 11:57:35 51109 Receive xrootd 26 bytes on 60327927
>>>> 191203 11:57:35 51109 Decode xrootd redirects nobody.351:[log in to unmask] to xcache-05.t2.ucsd.edu:1094 /store/data/Run2018D/MET/MINIAOD/PromptReco-v2/000/322/431/00000/4AD1B758-3BB8-E811-92B8-FA163EA98227.root
>>>> 191203 11:57:35 51514 nobody.351:[log in to unmask] XrootdProtocol: 0100 redirecting to xcache-05.t2.ucsd.edu:1094
>>>> 191203 11:57:35 51514 nobody.351:[log in to unmask] XrootdResponse: 0100 sending 25 data bytes; status=4004
>>>> 191203 11:57:35 51514 nobody.351:[log in to unmask] ofs_close: use=0 fn=dummy
>>>> 191203 11:57:35 51668 XrdSched: running cuser3.270:1838@cabinet-5-5-5 inq=0
>>>>
>>>> So, I'm suspecting redirector cmsd and xrootd never get connected, do you see something like this in the redirector xrootd log:
>>>> #### Initialization
>>>> ------ File system manager initialization completed.
>>>> 191203 11:14:07 51097 XrootdAioReq: Max aio/req=8; aio/srv=4096; Quantum=131072
>>>> 191203 11:14:07 51097 XrootdAioReq: Adding 18 aioreq objects.
>>>> 191203 11:14:07 51097 XrootdAio: Adding 18 aio objects; 4096 pending.
>>>> 191203 11:14:07 51097 XrdSched: scheduling xrootd protocol anchor in 3600 seconds
>>>> 191203 11:14:07 51097 XrdSched: scheduling transit protocol anchor in 3600 seconds
>>>> Config warning: 'xrootd.prepare logdir' not specified; prepare tracking disabled.
>>>> 191203 11:14:07 51111 XrdXeq: Admin traffic thread started
>>>> 191203 11:14:07 51109 XrdInet: Connected to xrootd.t2.ucsd.edu:2041        <-------
>>>> 191203 11:14:07 51109 cms_ClientMan: Connected to xrootd.t2.ucsd.edu v 3   <-------
>>>> 191203 11:14:07 51109 Hookup xrootd.t2.ucsd.edu qt=178ms rw=2
>>>> ------ xrootd protocol initialization completed.
>>>>
>>>> What does 'xrdmapc redirector:1094' give you?
>>>>
>>>> Would you mind sharing your redirector config?
>>>>
>>>> Andy, do you have a better idea what to look for / try?
>>>>
>>>> Matevz
>>>>
>>>>
>>>>
>>>>
>>>>> Sam
>>>>>
>>>>>
>>>>>
>>>>> redirector cmsd
>>>>>
>>>>> 191203 14:30:14 18957 Protocol: Primary server.202124:22@cephc01:1094 logged in.
>>>>> 191203 14:30:14 18957 Protocol: server.202124:22@cephc01:1094 system
>>>>> ID: [log in to unmask] 1213cephs03.beowulf.cluster
>>>>> =====> Routing for 10.1.50.11: local pub4 prv4
>>>>> =====> Route all4: 10.1.50.11 Dest=[::10.1.50.11]:1094
>>>>> 191203 14:31:16 18939 Config: manager service enabled.
>>>>> 191203 14:31:16 18953 State: Status changed to active + staging
>>>>>
>>>>>
>>>>> ----
>>>>> redirector xrootd
>>>>>
>>>>> ------ xrootd [log in to unmask]:1094 initialization completed.
>>>>> 191203 14:32:26 18970 XrootdXeq: root.218779:20@cephc01 pvt IPv4 login
>>>>> 191203 14:32:26 18970 root.218779:20@cephc01 XrootdResponse: sending
>>>>> err 3011: No servers have read access to the file
>>>>> 191203 14:32:26 18970 XrootdXeq: root.218779:20@cephc01 disc 0:00:00
>>>>>
>>>>>
>>>>> -----
>>>>> server cmsd
>>>>>
>>>>> ------ cmsd [log in to unmask]:46427 initialization completed.
>>>>> 191203 14:30:14 218768 do_Login:: Primary server 218748 logged in;
>>>>> data port is 1094
>>>>> Config Connecting to 1 manager and 1 site.
>>>>> 191203 14:30:14 218729 Config: server service enabled.
>>>>> 191203 14:30:14 218770 State: Status changed to active + staging
>>>>> 191203 14:30:14 218740 ManTree: Now connected to 1 root node(s)
>>>>> 191203 14:30:14 218740 Protocol: Logged into cephs03
>>>>>
>>>>>
>>>>> -----
>>>>> server xrootd
>>>>>
>>>>> 191203 14:30:14 218765 cms_Finder: Connected to cmsd via
>>>>> /tmp/cache/.olb/olbd.admin
>>>>> ------ xrootd protocol initialization completed.
>>>>> ------ xrootd [log in to unmask]:1094 initialization completed.
>>>>> 191203 14:30:15 218764 XrdFileCache_Manager: info Cache::Purge() Started.
>>>>> 191203 14:30:15 218764 XrdFileCache_Manager: info Cache::Purge()
>>>>> Finished, removed 0 data files, total size 0, bytes to remove at end:
>>>>> 0
>>>>> 191203 14:35:15 218764 XrdFileCache_Manager: info Cache::Purge() Started.
>>>>> 191203 14:35:15 218764 XrdFileCache_Manager: info Cache::Purge()
>>>>> Finished, removed 0 data files, total size 0, bytes to remove at end:
>>>>> 0
>>>>>
>>>>> On Mon, 2 Dec 2019 at 19:33, Matevz Tadel <[log in to unmask]> wrote:
>>>>>>
>>>>>> I'd try this:
>>>>>>
>>>>>> redirector:
>>>>>> xrootd.trace emsg login stall redirect
>>>>>> xrd.trace conn
>>>>>> cms.trace defer files redirect stage
>>>>>>
>>>>>> server:
>>>>>> # For debug, to see files being searched
>>>>>> # cms.trace    defer files redirect stage
>>>>>> cms.trace    defer redirect stage
>>>>>>
>>>>>>
>>>>>> You say xrdmapc shows the configured servers, right?
>>>>>>
>>>>>> We had some trouble with ipv4/6 at ucsd lately, clients will be redirected to
>>>>>> ipv6 servers only if they come in via ipv6 to the redirector.
>>>>>>
>>>>>> Can you restart redirector cmsd and then (after 30sec) look at:
>>>>>>
>>>>>> [1131] root@xrootd /var/log/xrootd/xcacheucsd# grep Routing cmsd.log | sort
>>>>>>
>>>>>> =====> Routing for bcache-1.t2.ucsd.edu: local pub4 prv4 pub6 prv6
>>>>>> =====> Routing for bcache-1.t2.ucsd.edu: local pub4 prv4 pub6 prv6
>>>>>> =====> Routing for xcache-00.t2.ucsd.edu: local pub4 prv4 pub6 prv6
>>>>>> =====> Routing for xcache-01.t2.ucsd.edu: local pub4 prv4 pub6 prv6
>>>>>> =====> Routing for xcache-02.t2.ucsd.edu: local pub4 prv4 pub6 prv6
>>>>>> =====> Routing for xcache-03.t2.ucsd.edu: local pub4 prv4 pub6 prv6
>>>>>> =====> Routing for xcache-04.t2.ucsd.edu: local pub4 prv4 pub6 prv6
>>>>>> =====> Routing for xcache-05.t2.ucsd.edu: local pub4 prv4 pub6 prv6
>>>>>> =====> Routing for xcache-06.t2.ucsd.edu: local pub4 prv4 pub6 prv6
>>>>>> =====> Routing for xcache-07.t2.ucsd.edu: local pub4 prv4 pub6 prv6
>>>>>> =====> Routing for xcache-08.t2.ucsd.edu: local pub4 prv4 pub6 prv6
>>>>>> =====> Routing for xcache-09.t2.ucsd.edu: local pub4 prv4 pub6 prv6
>>>>>> =====> Routing for xcache-10.t2.ucsd.edu: local pub4 prv4
>>>>>> =====> Routing for xcache-11.t2.ucsd.edu: local pub4 prv4 pub6 prv6
>>>>>> =====> Routing for xrd-cache-1.ultralight.org: local pub4 prv4 pub6 prv6
>>>>>> =====> Routing for xrd-cache-2.ultralight.org: local pub4 prv4 pub6 prv6
>>>>>>
>>>>>> Matevz
>>>>>>
>>>>>> On 2019-12-02 11:24, Sam Skipsey wrote:
>>>>>>> No, I explicitly did that. (As I noted, there's a typo for that in the
>>>>>>> example, as it uses "rw" not "r/w" , which doesn't work).
>>>>>>>
>>>>>>> I've tried basically every variation of stage/nostage/ r/w / r/o at
>>>>>>> different parts of the network, but the manager cmsd never seems to
>>>>>>> actually consider the servers (even when I've already pre-staged the
>>>>>>> file it's looking for by directly talking to the server xrootd service
>>>>>>> and getting it to cache).
>>>>>>>
>>>>>>> Sam
>>>>>>>
>>>>>>> On Mon, 2 Dec 2019 at 19:09, Matevz Tadel <[log in to unmask]> wrote:
>>>>>>>>
>>>>>>>> Hi,
>>>>>>>>
>>>>>>>> Before I go looking at what's wrong on the web page, have a look at this, page 21:
>>>>>>>>
>>>>>>>> https://indico.cern.ch/event/727208/contributions/3444604/
>>>>>>>>
>>>>>>>> Maybe you're missing the r/w for xrootd, stage r/o for cmsd trick?
>>>>>>>>
>>>>>>>> Cheers,
>>>>>>>> Matevz
>>>>>>>>
>>>>>>>> On 2019-11-28 08:22, Sam Skipsey wrote:
>>>>>>>>> Hello everyone,
>>>>>>>>>
>>>>>>>>> So, I have another question, working entirely from the documentation
>>>>>>>>> on xrootd.org
>>>>>>>>>
>>>>>>>>> In the documentation for cache configuration, there's an example of
>>>>>>>>> how to set up a cluster of disk caching proxies:
>>>>>>>>>
>>>>>>>>> https://xrootd.slac.stanford.edu/doc/dev410/pss_config.pdf [page 38,
>>>>>>>>> you can't copy it because weirdly its an image]
>>>>>>>>>
>>>>>>>>> I'm following that exactly (except for fixing the typo where the
>>>>>>>>> example has an export using "rw" and not "r/w" as an option), and,
>>>>>>>>> well, it just doesn't seem to work.
>>>>>>>>>
>>>>>>>>> If I talk directly to the server that the proxies talk to: I can get a file.
>>>>>>>>> If I talk to an individual proxy: I can also get a file (and it is cached)
>>>>>>>>> If I talk to the *redirector*, I get, with debugging on "Open has
>>>>>>>>> returned with status [ERROR] Server responded with an error: [3011] No
>>>>>>>>> servers have read access to the file"
>>>>>>>>>
>>>>>>>>> The redirector logs show that the cmsd on the proxy logs in (and is
>>>>>>>>> listed as a "server" in its list of servers), and the proxy cmsd logs
>>>>>>>>> also show that it happily registers to the redirector.
>>>>>>>>>
>>>>>>>>> How do I debug this?
>>>>>>>>>
>>>>>>>>> I've already tried adding and removing options to the various exports,
>>>>>>>>> making sure that all the relevant ports are open, etc.
>>>>>>>>>
>>>>>>>>> Sam
>>>>>>>>>
>>>>>>>>> ########################################################################
>>>>>>>>> Use REPLY-ALL to reply to list
>>>>>>>>>
>>>>>>>>> To unsubscribe from the XROOTD-L list, click the following link:
>>>>>>>>> https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=XROOTD-L&A=1
>>>>>>>>>
>>>>>>>>
>>>>>>
>>>>
>>>
>>> ########################################################################
>>> Use REPLY-ALL to reply to list
>>>
>>> To unsubscribe from the XROOTD-L list, click the following link:
>>> https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=XROOTD-L&A=1
>>>
>

########################################################################
Use REPLY-ALL to reply to list

To unsubscribe from the XROOTD-L list, click the following link:
https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=XROOTD-L&A=1

Top of Message | Previous Page | Permalink

Advanced Options


Options

Log In

Log In

Get Password

Get Password


Search Archives

Search Archives


Subscribe or Unsubscribe

Subscribe or Unsubscribe


Archives

March 2024
February 2024
January 2024
December 2023
November 2023
October 2023
September 2023
August 2023
July 2023
June 2023
May 2023
April 2023
March 2023
February 2023
January 2023
December 2022
October 2022
September 2022
August 2022
July 2022
June 2022
May 2022
April 2022
March 2022
February 2022
January 2022
December 2021
November 2021
October 2021
September 2021
August 2021
July 2021
June 2021
May 2021
April 2021
March 2021
February 2021
January 2021
December 2020
November 2020
October 2020
September 2020
August 2020
July 2020
June 2020
May 2020
April 2020
March 2020
February 2020
January 2020
December 2019
November 2019
October 2019
September 2019
August 2019
July 2019
June 2019
May 2019
April 2019
March 2019
February 2019
January 2019
December 2018
November 2018
October 2018
September 2018
August 2018
July 2018
June 2018
May 2018
April 2018
March 2018
February 2018
January 2018
December 2017
November 2017
October 2017
September 2017
August 2017
July 2017
June 2017
May 2017
April 2017
March 2017
February 2017
January 2017
December 2016
November 2016
October 2016
September 2016
August 2016
July 2016
May 2016
April 2016
March 2016
February 2016
January 2016
December 2015
November 2015
October 2015
September 2015
August 2015
July 2015
June 2015
May 2015
April 2015
March 2015
February 2015
January 2015
December 2014
November 2014
October 2014
September 2014
August 2014
July 2014
June 2014
May 2014
April 2014
March 2014
February 2014
January 2014
December 2013
November 2013
October 2013
September 2013
August 2013
July 2013
June 2013
May 2013
April 2013
March 2013
February 2013
January 2013
December 2012
November 2012
October 2012
September 2012
August 2012
July 2012
June 2012
May 2012
April 2012
March 2012
February 2012
January 2012
December 2011
November 2011
October 2011
September 2011
August 2011
July 2011
June 2011
May 2011
April 2011
March 2011
February 2011
January 2011
December 2010
October 2010
September 2010
August 2010
July 2010
June 2010
May 2010
April 2010
March 2010
February 2010
January 2010
December 2009
November 2009
October 2009
September 2009
July 2009
June 2009
May 2009
April 2009
March 2009
January 2009
December 2008
September 2008
August 2008
July 2008
June 2008
May 2008
April 2008
March 2008
February 2008
January 2008
December 2007
November 2007
October 2007
September 2007
August 2007
July 2007
June 2007
May 2007
April 2007
March 2007
February 2007
January 2007
December 2006
November 2006
October 2006
September 2006
August 2006
July 2006
June 2006
May 2006
April 2006
March 2006
February 2006
January 2006
December 2005
November 2005
October 2005
September 2005
August 2005
July 2005
June 2005
May 2005
April 2005
March 2005
February 2005
January 2005
December 2004
November 2004
October 2004
September 2004
August 2004

ATOM RSS1 RSS2



LISTSERV.SLAC.STANFORD.EDU

Secured by F-Secure Anti-Virus CataList Email List Search Powered by the LISTSERV Email List Manager

Privacy Notice, Security Notice and Terms of Use