Last modified 8 years ago Last modified on 06/17/14 11:45:04

Checking a record number

This may be required if Slony processes have stalled or if for some reason, JSOC records are offline. The JMD may stop trying to download them.

If the URL :;record=12697191-12697191

fails to download, then we need to use show_info to check on that record number and get a storage unit number (sunum) :

show_info -S ds="hmi.m_45s[?t_rec_index=12697191?]"

And then query JSOC about that sunum :

wget -O - ""

which might return something like :


Indicating that the sunum does not exist at JSOC (the NA's indicating that things like the path are not properly set). It could be that this means that you have to contact JSOC to resolve the problem. Even after the problem is resolved, it could be that the local JMD has (correctly) marked it as being offline and won't download it until the JMD's internal database is cleaned out. To see the records marked "OFLN" (offline) by the JMD, type

select * from drms.su_cache where state='OFLN';

The JMD is doing is refusing to download those records because it thinks they are offline. To clear it, do something like :

delete from drms.su_cache where state='OFLN';

Or, equivalently :

echo "delete from drms.su_cache where state='OFLN';" | /opt/JMD/bin/jmdij

The path /opt/JMD may be different on your system depending on where the JMD is installed.

You will not lose data since the JMD will now be able to try downloading those records again. It may be best to wait until the JSOC has confirmed they are available, however.