Refresh all collections harvested from other Dataverse instances/resume automated regular harvesting

[landreev]

Since an unexpected breakthrough (for the lack of a better word *) was made in IQSS/dataverse#11828 it has become possible to refresh large harvested collections. This issue is to track checking the entire list of other Dataverses that our prod. instance is configured to harvest from and making sure all are refreshed and updated regularly going forward.

Updated	No failures	Name	Collection	Notes
		Borealis	borealis_harvested	See note 1) below
		Odum	odum	See note 1) below
		QDR	qdr
		UVA	uva_harvested	See note 1) below
		Dataverse-NL	dataverseNL_harvested
		UBC Abacus	ubc_harvested
		TDR	tdr_harvested
		Dataverse-NO	dataverseNO_harvested
		Madroño	madrono_harvested
		ICRAF	icraf_harvested	some failures; see note 1) below
		CIMMYT-software	cimmyt_harvested	1 failure; see note 1) below
		CIMMYT-iwyp	cimmyt_harvested
		CIMMYT	cimmyt_harvested	A few invalid records; todo: review the log
		heiDATA	heidata_harvested	some failures; see note 1) below
		CIP	cip_harvested	This dataverse instance appears to have been down for a while
		ICRISAT	icrisat-harvested	1 failure; see note 1) below
		CIFOR	cifor_harvested
		ICARDA-CGIAR	icarda_harvested
		PUCP	pucp
		IDSC	idsc_harvested
		Recherche Data Gouv	recherchedatagouv	See note 2) below
		Historic Shapefiles, U of Oxford	oxfordmiddleeastcentreshapefiles
		DFC Ghana	DFC_Ghana

*) "breakthrough" may be too fancy a word - it turned out that the observed performance degradation was on account of something very stupid we were doing on the database side and, as a result, it was very easy to address without needing any code changes.

Notes:

1. There appears to be a minor bug affecting re-harvests of datasets from which some files have been dropped since the last harvest. Dev. issue: Harvesting: a bug in dataset update (UpdateHarvestedDatasetCommand) dataverse#11933
2. There appears to be a bug in import, where a ddi with the xml attribute xml:lang in the codeBook element cannot be imported. It can be described as minor, on a large scale of things (in terms of the number of datasets affected - all of them in Recherche Gouv - vs the total number of harvested datasets). However, it makes most of the datasets in that specific collection unimportable. Dev. issue: Harvesting: DDI import fails when there's a xml:lang attribute present in the codeBook element dataverse#11932

[jggautier]

I'm going to close all GitHub issues I've opened related to refreshing the large harvested collections you've listed in this GitHub issues and in those issues mention this one.

[jggautier]

Just noting here that @landreev's resumed automated regular harvesting. @landreev mentioned it after we had a chance to chat about the recent harvesting successes!