Re: [OpenAFS] Re: OpenAFS client crashes on RHEL 5.10 and RHEL 6.5
On Tue, Mar 18, 2014 at 3:04 PM, Andrew Deason adea...@sinenomine.net wrote: Due to the various bug reports filed about this issue, Red Hat has indicated that they are considering this issue a regression in Red Hat Enterprise Linux and will be providing a fix in at least RHEL 6.6 and an RHEL 6.5 update. Red Hat has not yet indicated that a fix will be available for RHEL 5, though they will probably be investigating a fix for that version. I haven't seen anyone mention it here, so I thought it may be worth pointing out that Red Hat has released a kernel update for RHEL 6.5 that includes the fix for this issue, as kernel 2.6.32-431.17.1. See advisory here: https://rhn.redhat.com/errata/RHSA-2014-0475.html And some technical details here: https://access.redhat.com/site/documentation/en-US/Red_Hat_Enterprise_Linux/6/html/6.5_Technical_Notes/kernel.html Marc ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
Re: [OpenAFS] Re: OpenAFS client crashes on RHEL 5.10 and RHEL 6.5
Thanks for that information, it is very useful to know, I had been wondering for a while which version would include the fix. Cheers, Stephen On 13 May 2014 20:35, Marc Dionne marc.c.dio...@gmail.com wrote: On Tue, Mar 18, 2014 at 3:04 PM, Andrew Deason adea...@sinenomine.net wrote: Due to the various bug reports filed about this issue, Red Hat has indicated that they are considering this issue a regression in Red Hat Enterprise Linux and will be providing a fix in at least RHEL 6.6 and an RHEL 6.5 update. Red Hat has not yet indicated that a fix will be available for RHEL 5, though they will probably be investigating a fix for that version. I haven't seen anyone mention it here, so I thought it may be worth pointing out that Red Hat has released a kernel update for RHEL 6.5 that includes the fix for this issue, as kernel 2.6.32-431.17.1. See advisory here: https://rhn.redhat.com/errata/RHSA-2014-0475.html And some technical details here: https://access.redhat.com/site/documentation/en-US/Red_Hat_Enterprise_Linux/6/html/6.5_Technical_Notes/kernel.html Marc ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
RE: [OpenAFS] Re: OpenAFS client crashes on RHEL 5.10 and RHEL 6.5
I've opened Case #01053566 with Red Hat today for USGS. - Ken SLAC has Red Hat case # 01051669 opened on this. I have made a request in our case to link to your case to add weight to the issue. Karl :�� T���j)b� b�өzpJ)ߢ�^��좸!��l��b��(���~�+Y���b�ا~�~ȧ~
RE: [OpenAFS] Re: OpenAFS client crashes on RHEL 5.10 and RHEL 6.5
Can an openafs developer (who is knowledgeable about the details of this issue) open a redhat bugzilla case on this (if not already done). SLAC's redhat TAM said this will be the best way to get traction on this. He wants to link all the Red Hat customer cases to a bugzilla case. SLAC's redhat TAM wrote: do you have the BZ number for the case filed by the OpenAFS folks? I can add all three cases to that and that would be the most effective approach. In theory our redhat TAM (or I, or anyone) could open a bugzilla, but the best description of the problem (and the most authoritative source) would be someone more familiar with the openafs code. ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
Re: [OpenAFS] Re: OpenAFS client crashes on RHEL 5.10 and RHEL 6.5
FYI From our open RH case for 5.x. Quote is from RH support: We have requested this regression be repaired in RHEL 5.11 under Bug 1080606, we have also requested that the fix be considered for backport into 5.10.z. ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
RE: [OpenAFS] Re: OpenAFS client crashes on RHEL 5.10 and RHEL 6.5
Hi Everyone, RedHat has a patched kernel for testing the fix on RHEL6. This is purely for testing. I don't have any RHEL6 boxes lying around to test this, and I don't have the time to set one up. If you would like to test the packages on RHEL6, please contact me off-list, and I'll give you a download link. I can also report your success or failure to RedHat as part of my support ticket. These packages are only for testing and should NOT be used on any critical or production system. No fix is available for RHEL5 at this time. Thanks, Jason --- Jason Edgecombe | Linux and Solaris Administrator UNC Charlotte | The William States Lee College of Engineering 9201 University City Blvd. | Charlotte, NC 28223-0001 Phone: 704-687-1943 jwedg...@uncc.edu | http://engr.uncc.edu | Facebook --- If you are not the intended recipient of this transmission or a person responsible for delivering it to the intended recipient, any disclosure, copying, distribution, or other use of any of the information in this transmission is strictly prohibited. If you have received this transmission in error, please notify me immediately by reply e-mail or by telephone at 704-687-1943. Thank you. -Original Message- From: openafs-info-ad...@openafs.org [mailto:openafs-info-ad...@openafs.org] On Behalf Of Rich Sudlow Sent: Wednesday, March 12, 2014 1:34 PM To: Ken Dreyer; ren...@slac.stanford.edu Cc: Jeffrey Altman; openafs-info@openafs.org; k...@slac.stanford.edu Subject: Re: [OpenAFS] Re: OpenAFS client crashes on RHEL 5.10 and RHEL 6.5 On 03/12/2014 01:24 PM, Ken Dreyer wrote: I've opened Case #01053566 with Red Hat today for USGS. - Ken Notre Dame also opened one a few days ago #01051867. Rich On Wed, Mar 12, 2014 at 11:12 AM, Renata Maria Dart ren...@slac.stanford.edu wrote: Hi Jeffmakes sense...we'll add more to our 2 cents with them. Thanks, Renata On Wed, 12 Mar 2014, Jeffrey Altman wrote: As with any situation in which there are products from two vendors involved it is critical that bug reports be opened by affected customers with both vendors. Only by receiving a significant number of bug reports can a vendor decide to prioritize a work around, fix or other change. A large vendor with hundreds of thousands or millions of customers is unlikely to prioritize a change if they determine that a small number of customers would benefit. This is especially true if the change would impact a larger number of customers in some adverse way. In this particular situation I am aware of several cases that have been opened with Red Hat. There can never be too many. Jeffrey Altman ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info -- Rich Sudlow University of Notre Dame Center for Research Computing - Union Station 506 W. South St South Bend, In 46601 (574) 631-7258 (office) (574) 807-1046 (cell) ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
RE: [OpenAFS] Re: OpenAFS client crashes on RHEL 5.10 and RHEL 6.5
From the knowledgebase article https://access.redhat.com/site/solutions/721353: Red Hat currently intend to also supply a fix for this issue in RHEL 5. Support issue #01056785 was filed with RedHat support for The University of North Carolina at Charlotte. Sorry that I'm late to the party. ;) --- Jason Edgecombe | Linux and Solaris Administrator UNC Charlotte | The William States Lee College of Engineering 9201 University City Blvd. | Charlotte, NC 28223-0001 Phone: 704-687-1943 jwedg...@uncc.edu | http://engr.uncc.edu | Facebook --- If you are not the intended recipient of this transmission or a person responsible for delivering it to the intended recipient, any disclosure, copying, distribution, or other use of any of the information in this transmission is strictly prohibited. If you have received this transmission in error, please notify me immediately by reply e-mail or by telephone at 704-687-1943. Thank you. -Original Message- From: openafs-info-ad...@openafs.org [mailto:openafs-info-ad...@openafs.org] On Behalf Of Andrew Deason Sent: Tuesday, March 18, 2014 3:04 PM To: openafs-annou...@openafs.org Cc: openafs-info@openafs.org Subject: [OpenAFS] Re: OpenAFS client crashes on RHEL 5.10 and RHEL 6.5 Due to the various bug reports filed about this issue, Red Hat has indicated that they are considering this issue a regression in Red Hat Enterprise Linux and will be providing a fix in at least RHEL 6.6 and an RHEL 6.5 update. Red Hat has not yet indicated that a fix will be available for RHEL 5, though they will probably be investigating a fix for that version. Customers with a Red Hat support contract are urged to follow up with Red Hat support for more details and timelines. Customers that concerned about this issue on RHEL 5 are strongly urgeed to file a bug with Red Hat to ensure that this issue is fixed in RHEL 5. When opening a case with Red Hat about this issue, it may be helpful to reference the following bug IDs, though you may not be able to view any information associated with them: - 1042731 - 1063201 - 1063198 - 1053566 - 1051867 In addition, Red Hat is maintaining the following page to provide information about this issue: https://access.redhat.com/site/solutions/721353. Thank you to everyone that contacted Red Hat about this issue. I would also like to thank Red Hat for being very cooperative and helpful when trying to resolve this issue. Also, note that while the fix that Red Hat is providing does appear to completely fix this latest regression, I'm not sure if the general issue of problems with OpenAFS' Linux mountpoint integration is completely solved. We're still working on further improvements in this area to make sure that problems like this cannot happen again in the future, as the relevant code in Linux continues to evolve. -- Andrew Deason adea...@sinenomine.net ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://urldefense.proofpoint.com/v1/url?u=https://lists.openafs.org/mailman/listinfo/openafs-infok=rPDsRmU4QOB12tMYRtqLtA%3D%3D%0Ar=ZX%2BVCzWU1kjcQl9dbGEC%2BfzUcVyHYrfkrueEUOFgZy4%3D%0Am=tojAh2ZU31qTQFPPMf8zDuizE3%2FcbiOVHAfAaXTBIHA%3D%0As=ca635db3c7a7bfa1d7d988c7e2ab74b04e3035938ca630b76029d5c910673542 ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
[OpenAFS] Re: OpenAFS client crashes on RHEL 5.10 and RHEL 6.5
Due to the various bug reports filed about this issue, Red Hat has indicated that they are considering this issue a regression in Red Hat Enterprise Linux and will be providing a fix in at least RHEL 6.6 and an RHEL 6.5 update. Red Hat has not yet indicated that a fix will be available for RHEL 5, though they will probably be investigating a fix for that version. Customers with a Red Hat support contract are urged to follow up with Red Hat support for more details and timelines. Customers that concerned about this issue on RHEL 5 are strongly urgeed to file a bug with Red Hat to ensure that this issue is fixed in RHEL 5. When opening a case with Red Hat about this issue, it may be helpful to reference the following bug IDs, though you may not be able to view any information associated with them: - 1042731 - 1063201 - 1063198 - 1053566 - 1051867 In addition, Red Hat is maintaining the following page to provide information about this issue: https://access.redhat.com/site/solutions/721353. Thank you to everyone that contacted Red Hat about this issue. I would also like to thank Red Hat for being very cooperative and helpful when trying to resolve this issue. Also, note that while the fix that Red Hat is providing does appear to completely fix this latest regression, I'm not sure if the general issue of problems with OpenAFS' Linux mountpoint integration is completely solved. We're still working on further improvements in this area to make sure that problems like this cannot happen again in the future, as the relevant code in Linux continues to evolve. -- Andrew Deason adea...@sinenomine.net ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
[OpenAFS] Re: OpenAFS client crashes on RHEL 5.10 and RHEL 6.5
On Fri, 7 Mar 2014 10:03:14 -0600 Andrew Deason adea...@sinenomine.net wrote: It has been discovered that the OpenAFS client interacts poorly with a change to the Linux kernel introduced in Red Hat Enterprise Linux versions 5.10 and 6.5, which can cause a kernel panic with certain AFS access patterns. Sites may want to exercise caution when considering upgrading RHEL systems that are running OpenAFS clients. Red Hat has indicated they will be providing an update to RHEL 6.5 to fix this issue, and it will be fixed in 6.6 (I can no longer reproduce our issue with their fix in place). Thanks to everyone that sent in bug reports. However, they have not given any specifics for anything in RHEL 5 (though they probably will do something about RHEL 5, but maybe not as urgently). If you care about this wrt RHEL 5, you need to send in a bug report specifically mentioning RHEL 5. I'll collect some more specifics and send them to -announce when I'm sure I'm allowed to send them; I just wanted to get this little bit of information out. This also means that I consider the redoing how we represent mountpoints work to be less urgent now (though it may still be required in the future). I'm going to send some stuff to gerrit shortly, but actually fixing all of the issues with it is a bit less urgent. -- Andrew Deason adea...@sinenomine.net ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
Re: [OpenAFS] Re: OpenAFS client crashes on RHEL 5.10 and RHEL 6.5
On 03/17/2014 12:15 PM, Andrew Deason wrote: On Fri, 7 Mar 2014 10:03:14 -0600 Andrew Deason adea...@sinenomine.net wrote: It has been discovered that the OpenAFS client interacts poorly with a change to the Linux kernel introduced in Red Hat Enterprise Linux versions 5.10 and 6.5, which can cause a kernel panic with certain AFS access patterns. Sites may want to exercise caution when considering upgrading RHEL systems that are running OpenAFS clients. Red Hat has indicated they will be providing an update to RHEL 6.5 to fix this issue, and it will be fixed in 6.6 (I can no longer reproduce our issue with their fix in place). Thanks to everyone that sent in bug reports. However, they have not given any specifics for anything in RHEL 5 (though they probably will do something about RHEL 5, but maybe not as urgently). If you care about this wrt RHEL 5, you need to send in a bug report specifically mentioning RHEL 5. I'll collect some more specifics and send them to -announce when I'm sure I'm allowed to send them; I just wanted to get this little bit of information out. This also means that I consider the redoing how we represent mountpoints work to be less urgent now (though it may still be required in the future). I'm going to send some stuff to gerrit shortly, but actually fixing all of the issues with it is a bit less urgent. Thanks Andrew I checked the case I opened previously (01051867) it was for 6.5 although I did ask for Red Hat 5 6 supported. Just opened another case specifically for RHELS 5.10 - case # 01055427 Rich -- Rich Sudlow University of Notre Dame Center for Research Computing - Union Station 506 W. South St South Bend, In 46601 (574) 631-7258 (office) (574) 807-1046 (cell) ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
[OpenAFS] Re: OpenAFS client crashes on RHEL 5.10 and RHEL 6.5
On Wed, 12 Mar 2014 08:19:22 -0700 (PDT) Renata Maria Dart ren...@slac.stanford.edu wrote: Hi Andrew, we opened a ticket with RedHat, but it is not clear how to proceedare you thinking that we should be asking them for a fix? If you're concerned about this, the way I would frame the discussion with them is that you're asking if Red Hat can provide a workaround in their kernel. One possible way they can maybe achieve that is to remove the referenced change, since I am not really sure what this fixes (or what its motivation was): [RHEL5] - [fs] vfs: stop d_splice_alias creating directory aliases (J. Bruce Fields) [785916] [RHEL6] - [fs] vfs: stop d_splice_alias creating directory aliases (J. Bruce Fields) [820446] Alternatively, they can engage us in conversation about how OpenAFS can more easily avoid this. J. Bruce Fields is already discussing this a little on openafs-devel, so you could just point them at that or the referenced central.org RT ticket. -- Andrew Deason adea...@sinenomine.net ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
Re: [OpenAFS] Re: OpenAFS client crashes on RHEL 5.10 and RHEL 6.5
Thanks Andrew...we'll see what we can find out. Renata On Wed, 12 Mar 2014, Andrew Deason wrote: On Wed, 12 Mar 2014 08:19:22 -0700 (PDT) Renata Maria Dart ren...@slac.stanford.edu wrote: Hi Andrew, we opened a ticket with RedHat, but it is not clear how to proceedare you thinking that we should be asking them for a fix? If you're concerned about this, the way I would frame the discussion with them is that you're asking if Red Hat can provide a workaround in their kernel. One possible way they can maybe achieve that is to remove the referenced change, since I am not really sure what this fixes (or what its motivation was): [RHEL5] - [fs] vfs: stop d_splice_alias creating directory aliases (J. Bruce Fields) [785916] [RHEL6] - [fs] vfs: stop d_splice_alias creating directory aliases (J. Bruce Fields) [820446] Alternatively, they can engage us in conversation about how OpenAFS can more easily avoid this. J. Bruce Fields is already discussing this a little on openafs-devel, so you could just point them at that or the referenced central.org RT ticket. -- Andrew Deason adea...@sinenomine.net ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
Re: [OpenAFS] Re: OpenAFS client crashes on RHEL 5.10 and RHEL 6.5
As with any situation in which there are products from two vendors involved it is critical that bug reports be opened by affected customers with both vendors. Only by receiving a significant number of bug reports can a vendor decide to prioritize a work around, fix or other change. A large vendor with hundreds of thousands or millions of customers is unlikely to prioritize a change if they determine that a small number of customers would benefit. This is especially true if the change would impact a larger number of customers in some adverse way. In this particular situation I am aware of several cases that have been opened with Red Hat. There can never be too many. Jeffrey Altman On 3/12/2014 12:49 PM, Renata Maria Dart wrote: Thanks Andrew...we'll see what we can find out. Renata On Wed, 12 Mar 2014, Andrew Deason wrote: On Wed, 12 Mar 2014 08:19:22 -0700 (PDT) Renata Maria Dart ren...@slac.stanford.edu wrote: Hi Andrew, we opened a ticket with RedHat, but it is not clear how to proceedare you thinking that we should be asking them for a fix? If you're concerned about this, the way I would frame the discussion with them is that you're asking if Red Hat can provide a workaround in their kernel. One possible way they can maybe achieve that is to remove the referenced change, since I am not really sure what this fixes (or what its motivation was): [RHEL5] - [fs] vfs: stop d_splice_alias creating directory aliases (J. Bruce Fields) [785916] [RHEL6] - [fs] vfs: stop d_splice_alias creating directory aliases (J. Bruce Fields) [820446] Alternatively, they can engage us in conversation about how OpenAFS can more easily avoid this. J. Bruce Fields is already discussing this a little on openafs-devel, so you could just point them at that or the referenced central.org RT ticket. -- Andrew Deason adea...@sinenomine.net ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info smime.p7s Description: S/MIME Cryptographic Signature
Re: [OpenAFS] Re: OpenAFS client crashes on RHEL 5.10 and RHEL 6.5
Hi Jeffmakes sense...we'll add more to our 2 cents with them. Thanks, Renata On Wed, 12 Mar 2014, Jeffrey Altman wrote: As with any situation in which there are products from two vendors involved it is critical that bug reports be opened by affected customers with both vendors. Only by receiving a significant number of bug reports can a vendor decide to prioritize a work around, fix or other change. A large vendor with hundreds of thousands or millions of customers is unlikely to prioritize a change if they determine that a small number of customers would benefit. This is especially true if the change would impact a larger number of customers in some adverse way. In this particular situation I am aware of several cases that have been opened with Red Hat. There can never be too many. Jeffrey Altman ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
Re: [OpenAFS] Re: OpenAFS client crashes on RHEL 5.10 and RHEL 6.5
I've opened Case #01053566 with Red Hat today for USGS. - Ken On Wed, Mar 12, 2014 at 11:12 AM, Renata Maria Dart ren...@slac.stanford.edu wrote: Hi Jeffmakes sense...we'll add more to our 2 cents with them. Thanks, Renata On Wed, 12 Mar 2014, Jeffrey Altman wrote: As with any situation in which there are products from two vendors involved it is critical that bug reports be opened by affected customers with both vendors. Only by receiving a significant number of bug reports can a vendor decide to prioritize a work around, fix or other change. A large vendor with hundreds of thousands or millions of customers is unlikely to prioritize a change if they determine that a small number of customers would benefit. This is especially true if the change would impact a larger number of customers in some adverse way. In this particular situation I am aware of several cases that have been opened with Red Hat. There can never be too many. Jeffrey Altman ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
Re: [OpenAFS] Re: OpenAFS client crashes on RHEL 5.10 and RHEL 6.5
On 03/12/2014 01:24 PM, Ken Dreyer wrote: I've opened Case #01053566 with Red Hat today for USGS. - Ken Notre Dame also opened one a few days ago #01051867. Rich On Wed, Mar 12, 2014 at 11:12 AM, Renata Maria Dart ren...@slac.stanford.edu wrote: Hi Jeffmakes sense...we'll add more to our 2 cents with them. Thanks, Renata On Wed, 12 Mar 2014, Jeffrey Altman wrote: As with any situation in which there are products from two vendors involved it is critical that bug reports be opened by affected customers with both vendors. Only by receiving a significant number of bug reports can a vendor decide to prioritize a work around, fix or other change. A large vendor with hundreds of thousands or millions of customers is unlikely to prioritize a change if they determine that a small number of customers would benefit. This is especially true if the change would impact a larger number of customers in some adverse way. In this particular situation I am aware of several cases that have been opened with Red Hat. There can never be too many. Jeffrey Altman ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info -- Rich Sudlow University of Notre Dame Center for Research Computing - Union Station 506 W. South St South Bend, In 46601 (574) 631-7258 (office) (574) 807-1046 (cell) ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info