As I wrote on the topic about tkt 4126, also our CNAF-T1 (and INFN-CNAF)
sites are experiencing intermittent replica problem.
The italian BDII used by all the IT-ROC sites is responding with similar
times as the CERN one, but I suspect some intermittent internal network
problem we need to investigate, because the other IT-ROC sites seem not
to suffer this problem.
When we are full staff again, we'll try also to better monitor ldap
queries to BDIIs with Nagios...
cheers
Alessandro
Laurence Field wrote:
> Querying your BDII seems to take a long time.
>
> time ldapsearch -H ldap://bdii.prd.hp.com:2170 -x -b o=grid > /dev/null
>
> real 1m0.337s
> user 0m0.540s
> sys 0m0.010s
>
> Such a long query time could be a reason for your problems. Usually
> the query should respond within a few seconds.
>
> time ldapsearch -H ldap://lcg-bdii.cern.ch:2170 -x -b o=grid > /dev/null
>
> real 0m3.278s
> user 0m2.360s
> sys 0m0.020s
>
> You might want to investigate what is making your BDII run so slow.
>
> Laurence
>
>
> Sotomayor, Maniel wrote:
>
>> Hello all,
>>
>> We've been experiencing multiple incidents related to replica
>> management.
>> In multiple occasions various tests fail randomly, but all are
>> related to
>> the replica management functionality. Tests fail consecutively
>> approximately 5 or 6 times and then work OK. The fact is that,
>> configuration has not been changed at all to none of the site nodes.
>> Did
>> someone have encountered this problem on their site?
>>
>> Could it be a problem with the information system?
>> The error messages that appear repeatedly are:
>> "lcg_*: Invalid argument"
>> "lcg_*: No such file or directory"
>>
>> Any help for diagnosing these problems are greatly appreciated.
>>
>> Best Regards,
>> ./MS
>> -------------------------------------
>> Maniel Sotomayor
>> Software Designer
>> Hewlett-Packard Technology Center, PR
>> +1(787)819-7673
>>
>>
--
Alessandro Cavalli
INFN - CNAF
Viale Berti Pichat 6/2
40127 Bologna
Italy
tel: +39 051 6092849
fax: +39 051 6092746
ICQ: 12771368
|