Here's the script:
https://svn.lal.in2p3.fr/LCG/QWG/LAL/DPM/Tools/drain.sh
Note that in its current form, you will have to set DPNS_HOST and
DPM_HOST manually in the script.
Cheers,
Greig
On 21/02/08 09:51, Santanu Das wrote:
> It's dpm-drain is stopping pretending a successful exit and as if drain
> is complete - but actually not, too many times here. Now it's almost
> once after ah hr. of run.
> Is it like that by nature or a bug? Or there is something wrong with our
> DPM? Nothing significant in dpm or dpns logs, so far.
>
> cheers,
> Santanu
>
>
>
>
> Greig Alan Cowan wrote:
>> Hi Phil,
>>
>> On 20/02/08 18:21, Phil Roffe wrote:
>>> However it resulted in the following two errors...
>>>
>>> [root@gallows dpm-gsiftp]# dpm-drain --pool dpmPart1
>>> se01.dur.scotgrid.ac.uk:/gridstore0/atlas/2008-02-01/fileb2199f4c-29e4-43e2-b8da-b026637a093f.758587.0
>>> is being populated
>>> se01.dur.scotgrid.ac.uk:/gridstore0/atlas/2008-02-01/file3faa18a6-44df-4538-aa0c-09e31f33419e.758718.0
>>> is being populated
>>
>> This is caused by the file having a status in the Cns_file_replicas
>> table which is == 'P', meaning that the file is populated. I'm trying
>> to find out what this actually means.
>>
>> if (lp->status != '-') { /* file is being populated/deleted */
>> printf ("%s is being %s\n",
>> lp->sfn,(lp->status == 'P') ? "populated" : "deleted");
>> fflush (stdout);
>> rc = 1;
>> continue;
>> }
>>
>>
>> Greig
|