Title | Affected Services | Affected VOs | Description | Associated Tickets | |
---|---|---|---|---|---|
Latest LB upgrade breaks WMS | WMS | All | The glite-lb-common upgrade breaks the WMS installation | GGUS:106143![]() |
|
Job submission fails for VOs supported by VOMS server with SHA-512 host certificate | CREAM, Argus, UI, WN. | All | This issue is not blocking the current operation but it's blocking the adoption of the VOMS server with SHA2 certificates. | GGUS:104768![]() |
|
Problem with opening files from EOS using Brazilian proxies | EOS | LHCb | Some users have issues accessing EOS with proxies of certs issued by a Brazilian CA | GGUS:106434![]() |
|
dCache upgrade side-effect | dCache | All | File protocol left enabled is undesirable | GGUS:106368![]() |
|
APEL fails to parse accounting records | APEL | All | The Apel version 1.2.1 ( part of Apel Publisher 4.2.0) released in EMI 3 the 7/08 has a problem when parsing compressed accounting data files. | n/a | |
dCache issue with a Brazilian CA Certificate | dCache | LHCb | dCache fails to authenticate users with a proxy coming from a Brazilian CA. The issue is related to the fact that the specific Brazilian CA uses a CA-root certificate with a UTF-8 character. | GGUS:107906![]() |
|
Broken symlink after EMI upgrade | CREAM | All | The symlink to the lb-client is broken on CREAM | GGUS:106121![]() |
|
Latest CREAM update breaks submissions from WMS | CREAM | All | WMS submission to latest cream is broken. It fails to parse the JDL coming from the WMS. | GGUS:106134![]() |
|
installation of several grid product is broken | CREAM, WMS, L&B, WMS , WN | All | classADS package was declared orphaned in EPEL, and retired from all EPEL repositories -> many grid services cannot be installed anymore | GGUS:108878![]() |
|
gridftp logging too verbose on DPM 1.8.9 | DPM | All | The latest version of DPM has an issue with gridftp logging which is always at teh maximum level even when disabled. | n/a | |
Storm and ARGUS integration issues | Storm | All | memory leak when contacting ARGUS | GGUS:104885![]() |
|
CREAM GAHP component crashes | CREAM, HTCondor | ATLAS | CREAM GAHP component used by Condor to submit to CREAM is crashing repeatedly since last week, both on pilot factories and in SAM pre-prod. | HTC:4406![]() |
|
RFC proxy support in HTCondor | HTCondor | All | HTCondor is using a very old version of the CREAM client and the issue can be fixed repackaging with a recent version. | n/a | |
Argus and CANL OCSP support | Argus | CMS | Argus for some CAs use the OCSP protocol handling (through CANL ) causing performance issues and instability | GGUS:105666![]() ![]() |
|
ARGUS PAP-admin client and SSLv3 | ARGUS | All | With the latest version of openJDK the support for sslv3 has been disabled. It turned out that the ARGUS - PAP-client forces the usage of sslv3 by default , therefore after the java upgrade the client fails to workt | GGUS:111505![]() |
|
RHEL6.6 kernel fuse bug affecting CVMFS NFS installations | CVMFS | All | l It was found that the new Red Hat kernel 2.6.32-504 that comes with RHEL6.6 and as a security patch for older RHEL6 installations has a bug wrt. fuse. Due to the bug, cvmfs clients that are exported via NFS immediately provoke a kernel panic. The problem is not cvmfs specific but all fuse modules that are exported via NFS are affected. Site's that use the NFS exported cvmfs client must not update to this kernel. The common cvmfs deployment mode as a fuse module on the worker nodes seems not to be affected. | BUGZILLA 1162782![]() |
|
Storm and new GlueValidator check | Storm | All | There is an issue in the info-provider for Storm which has been discovered when running a check introduced in the new glue-validator version | GGUS:108556![]() |
|
xroot package deployed with ROOT 6 breaks access to dCache storage | dCache/Xrootd | LHCb, CMS | latest xrootd version negotiates a "vector read" access which dCache doesn't know how to handle. Subsequently , if run within ROOT, the application crashes. | ROOT-6639![]() |
|
No checksum stored on DPM and no returned by srm-ls | DPM | LHCb | LHCb cannot retrieve checksum via srm-ls after transfer via FTS is performed | GGUS:111403![]() |
|
Slapd process on TopBDII crashes after upgrade to CentOS/SLC6.7 | BDII | All | The slapd process under high load condition ( Top BDII collecting all site bdii info) crashes. this problem started to appear with openldap-servers-2.4.40-5.el6 | https://bugzilla.redhat.com/show_bug.cgi?id=1257543![]() |
|
HTCondor 8.4.0 dependency issues with UMD WN | HTCondor, WN | All | The latest version of HTCondor cannot be installed on a WN installed from UMD due to incompatible dependency, ( HTCondor depends now on condor-classads, while WN depends on glite-lb-common which depends on classads). | GGUS:116274![]() |
|
DPM-Xrootd wrongly sends the XrdCmsClient Remove notices on ENOENT | DPM | CMS | This condition can be triggered if a remote client includes a "tried" list that indicates the site has been tried before. Having the cmsd then send a removed notice can cause problems, as reported by AAA (cms) xrootd federation. | https://its.cern.ch/jira/browse/LCGDM-1821![]() |
|
dCache possible data loss | dCache | All | dCache pools configured to use Berkeley DB as metadata backend, which can lead to data loss. The problem affects dCache >2.12. The problem has been fixed by developers, please find details on the release notes of each release or from the Wiki: https://twiki.cern.ch/twiki/pub/LCG/WLCGBaselineVersions/dcache-bug.txt | n/a | |
Missing Key Usage extension in delegated proxy | CREAM | ATLAS | ATLAS jobs cannot use all Rucio functionality due to the WN proxy being defective | GGUS:105188![]() |
ETA ~ Oct |
gridsite/canl crashes in multithread env | FTS3 | All | gridsite/canl lib wrongly call openssl lib initialization for each operation, which is against what suggested by the openssl lib manual. for this reason FTS3 usage of the lib in a multithreaded env lead to a server crash | GGUS:106533![]() GGUS:106668 ![]() |
It seems that the fix requires a major refactoring of the lib. FTS3 has been patched with a workaround to disable multithreading on delegation |
issue with LSF and PBS and CREAM-CE job cancel/purge | CREAM-BLAH | All | There is a bug in how BUpdaterLSF manage removed jobs. The output of bjobs command for a killed job should contain "Exited with exit status" string to be considered by the bupdaterlsf, while some jobs including the latest one has only "Exited". So this problem does not trigger the execution of bhist and so the job in not considered removed until next bhist triggered by a different job | GGUS:128151![]() |
Fix for LSF under release in UMD 4.5, issue with PBS still under investigation |
Storm and ARGUS integration issues | StoRM | All | Missing gridftp banning via ARGUS in Storm | GGUS:96374![]() |
the gridftp banning is under implementation. |
CREAM JOB CANCEL does not work with HTCondor | CREAM-BLAH | All | CREAM JOB CANCEL does not work with HTCondor | GGUS:124034![]() |
Fix to be released in UMD 4.5 |
StoRM srm space not updated when data deleted via WebDav | StoRM | All | When deleting files via WebDav the related spacetoken is not updated | GGUS:126896![]() |
Workaround documented by devs on the GGUS ticket |