[Dxspider-support] user.asc size ?

[Tim Tuck]

Hi Dirk,

Running perl against any of my user_asc.o files cause 100% CPU and this 
is from tiny user_asc files of about 2k.

It rapidly fills memory and swap until it barfs lots of stuff into 
syslog with lines like this...

Dec  1 20:41:26 dx kernel: [173016.802928] Out of memory: Killed process 
40685 (perl) total-vm:1367748kB, anon-rss:1352540kB, file-rss:1076kB, 
shmem-rss:0kB, UID:1002 pgtables:2724kB

Full syslog barf message available on request.

Likewise running export_user causes 100% CPU and the cluster to stall 
with the resultant user_asc file rapidly growing into the multi GB range 
and the only way to stop it is to kill -9 it all!

I suspect there is something terribly wrong inside the users.v3 file, is 
there some other way to reset it ?

This is a brand new node still in testing so destructive options can be 
entertained.

regards

Tim


On 2/12/21 7:16 am, Dirk Koopman via Dxspider-support wrote:
> Firstly, use user_asc.o (that's what it's there for).
>
> But it is text and you can 'less' and I guarantee that it will text 
> and blobs of binary.
>
> You can edit it and remove the rubbish. But perl user_asc.o will be 
> close enough.
>
> Dirk
>
> On 01/12/2021 19:52, Tim Tuck via Dxspider-support wrote:
>>
>> Hi Kin,
>>
>> OK, doing that caused a crash and this...
>>
>> Dec  1 19:49:13 dx kernel: [169883.949427] Out of memory: Killed 
>> process 38305 (perl) total-vm:1372232kB, anon-rss:1356840kB, 
>> file-rss:2616kB, shmem-rss:0kB, UID:1002 pgtables:2732kB oom_score_adj:0
>>
>> Is there any other way  to deal with it ?
>>
>> thanks
>>
>> Tim
>>
>>
>> On 1/12/21 6:36 pm, Joaquin . via Dxspider-support wrote:
>>> Hi Tim,
>>>
>>> It can't be, that file is wrong.
>>>
>>> 1. stop the cluster
>>> 2. cd /spider/data
>>> 3. perl user_asc
>>> 4. restart the cluster
>>>
>>> You should get a file with a few megabytes
>>> Mine after more than a year is < 6 M
>>>
>>> 73 de Kin
>>> EA3CV
>>>
>>>
>>>
>>> El mié, 1 dic 2021 a las 3:56, Tim Tuck via Dxspider-support 
>>> (<dxspider-support at tobit.co.uk 
>>> <mailto:dxspider-support at tobit.co.uk>>) escribió:
>>>
>>>     Hi all,
>>>
>>>     I'm monitoring our new cluster and I'm a bit concerned about the
>>>     size of
>>>     user.asc, Its currently 33G.
>>>
>>>     Is this normal ? or has something gone amiss ?
>>>
>>>     thanks
>>>
>>>     Tim
>>>
>>>     -- 
>>>     VK2XAX : QF56if : ITU59 : CQ30 : BMARC : WIA
>>>
>>>
>>>     _______________________________________________
>>>     Dxspider-support mailing list
>>>     Dxspider-support at tobit.co.uk <mailto:Dxspider-support at tobit.co.uk>
>>>     https://mailman.tobit.co.uk/mailman/listinfo/dxspider-support
>>>     <https://mailman.tobit.co.uk/mailman/listinfo/dxspider-support>
>>>
>>>
>>> _______________________________________________
>>> Dxspider-support mailing list
>>> Dxspider-support at tobit.co.uk
>>> https://mailman.tobit.co.uk/mailman/listinfo/dxspider-support
>> -- 
>> VK2XAX : QF56if : ITU59 : CQ30 : BMARC : WIA
>>
>> _______________________________________________
>> Dxspider-support mailing list
>> Dxspider-support at tobit.co.uk
>> https://mailman.tobit.co.uk/mailman/listinfo/dxspider-support
>
>
> _______________________________________________
> Dxspider-support mailing list
> Dxspider-support at tobit.co.uk
> https://mailman.tobit.co.uk/mailman/listinfo/dxspider-support

-- 
VK2XAX : QF56if : ITU59 : CQ30 : BMARC : WIA

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.tobit.co.uk/pipermail/dxspider-support/attachments/20211202/d4b76af8/attachment.htm>