Cluster Admins - Call for experiences/issues encountered
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 This is specifically to galaxy administrators and developers who have worked with galaxy and cluster deployment. If you have successfully deployed a galaxy instance backed by one (or more) clusters, I'm looking to hear about your experience; - - the configuration - - any pitfalls encountered - - any extra tools that needed to be installed - - any ugly hacks you had to implement The current cluster documentation is somewhat lacking, and I'm willing to compile everyone's experiences deploying to clusters if you're willing to send them in. I'll update the cluster page on the wiki as I receive replies and can allocate time. By way of example, here's the summary of what I would reply to this email with my experiences deploying to an HTCondor cluster - - HTcondor ran fine, accepted jobs and worked for all users - - Connecting galaxy to it would result in some jobs running/some jobs failing. It was determined that those jobs which ran, were only on localhost. - - This was a result of the fact that galaxy's home directory/data was not exported - - Galaxy was made an LDAP user and a folder created in the correct location on all computers - - This still failed, as galaxy does not transfer files by default (as far as I can tell), even though that's available in HTCondor - - Finally galaxy's home directory was NFS exported and things seemed to work. Cheers, Eric - -- Eric Rasche Programmer II Center for Phage Technology Texas A&M University College Station, TX 77843 404-692-2048 esr@tamu.edu rasche.eric@yandex.ru -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.17 (GNU/Linux) iQIcBAEBAgAGBQJThL4PAAoJEMqDXdrsMcpVJ0wP/jPFrTxRYnWcavbUxhht8tWE VDhid2pst7o+qkkLAUylPyNzJz31movsU1Gd2FJLP4+RHb3lhy2GWm8mR9G4JuMM g7fHAwIRj1qTT5tMXAEN6EoCtrValZH0EkIQPaFUPIiN1wzGTTDP56V+L6f3xQNg zglk2GTO+hM+AunYGuDugDvdKTPyKgR1K2vaCNdc2kZlvQZe7TlOHZEbXaoDGTY4 vwHcNR0+8fvf6Hs1liCuCuh8nlTBABZLxMl5LgxOaZqbaAVQWsqXNdh1tF/rlERL +PlNv5HRBWz26iGE04jbAD1pKmxg33Knl3l4MXrUnB5ZJ2dIDJL0YHOhXM3mjI4T Hf8mv/iVr21xx3zrrAbAlRIt3Npw4DtaGSqBFdKvnxQZygjKEF7ZKTGfA4xLNRHh 4dPElUpv8b3jUe2GqhXyrmHYuBdWqBC5Y/ii0k4c/bHjoXW6yg0mPs7po7i5+5JS wnO3mfoFhnyEALd45EBprkQY+kx05zjSg4ntpeywl8fIM2uGT0GyANA0/prKhcmB 5PiYUejsaDTY4GigiiZqz0tnC5z7VDyjByfyl7ufNaaCWNRXlTNIzH0ePy0ZjLrU Sj8XyI1XoOSdtyxxvNNe8OPDxxffnJxi0rres4Q3CC/EbZK4vBF7chsisHd28Q29 M1e+lNdO8FBIoy5+MR3H =6Jbk -----END PGP SIGNATURE-----
Hi Eric, thanks for caring about that issue. Do you know about Galaxy Community Hubs? https://wiki.galaxyproject.org/Community/ The community is encouraged to share experience under that page. Maybe you can add yours to that page as well? Cheers, Bjoern Am 27.05.2014 18:32, schrieb Eric Rasche:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
This is specifically to galaxy administrators and developers who have worked with galaxy and cluster deployment.
If you have successfully deployed a galaxy instance backed by one (or more) clusters, I'm looking to hear about your experience;
- - the configuration - - any pitfalls encountered - - any extra tools that needed to be installed - - any ugly hacks you had to implement
The current cluster documentation is somewhat lacking, and I'm willing to compile everyone's experiences deploying to clusters if you're willing to send them in. I'll update the cluster page on the wiki as I receive replies and can allocate time.
By way of example, here's the summary of what I would reply to this email with my experiences deploying to an HTCondor cluster
- - HTcondor ran fine, accepted jobs and worked for all users - - Connecting galaxy to it would result in some jobs running/some jobs failing. It was determined that those jobs which ran, were only on localhost. - - This was a result of the fact that galaxy's home directory/data was not exported - - Galaxy was made an LDAP user and a folder created in the correct location on all computers - - This still failed, as galaxy does not transfer files by default (as far as I can tell), even though that's available in HTCondor - - Finally galaxy's home directory was NFS exported and things seemed to work.
Cheers, Eric
- -- Eric Rasche Programmer II Center for Phage Technology Texas A&M University College Station, TX 77843 404-692-2048 esr@tamu.edu rasche.eric@yandex.ru -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.17 (GNU/Linux)
iQIcBAEBAgAGBQJThL4PAAoJEMqDXdrsMcpVJ0wP/jPFrTxRYnWcavbUxhht8tWE VDhid2pst7o+qkkLAUylPyNzJz31movsU1Gd2FJLP4+RHb3lhy2GWm8mR9G4JuMM g7fHAwIRj1qTT5tMXAEN6EoCtrValZH0EkIQPaFUPIiN1wzGTTDP56V+L6f3xQNg zglk2GTO+hM+AunYGuDugDvdKTPyKgR1K2vaCNdc2kZlvQZe7TlOHZEbXaoDGTY4 vwHcNR0+8fvf6Hs1liCuCuh8nlTBABZLxMl5LgxOaZqbaAVQWsqXNdh1tF/rlERL +PlNv5HRBWz26iGE04jbAD1pKmxg33Knl3l4MXrUnB5ZJ2dIDJL0YHOhXM3mjI4T Hf8mv/iVr21xx3zrrAbAlRIt3Npw4DtaGSqBFdKvnxQZygjKEF7ZKTGfA4xLNRHh 4dPElUpv8b3jUe2GqhXyrmHYuBdWqBC5Y/ii0k4c/bHjoXW6yg0mPs7po7i5+5JS wnO3mfoFhnyEALd45EBprkQY+kx05zjSg4ntpeywl8fIM2uGT0GyANA0/prKhcmB 5PiYUejsaDTY4GigiiZqz0tnC5z7VDyjByfyl7ufNaaCWNRXlTNIzH0ePy0ZjLrU Sj8XyI1XoOSdtyxxvNNe8OPDxxffnJxi0rres4Q3CC/EbZK4vBF7chsisHd28Q29 M1e+lNdO8FBIoy5+MR3H =6Jbk -----END PGP SIGNATURE----- ___________________________________________________________ Please keep all replies on the list by using "reply all" in your mail client. To manage your subscriptions to this and other Galaxy lists, please use the interface at: http://lists.bx.psu.edu/
To search Galaxy mailing lists use the unified search at: http://galaxyproject.org/search/mailinglists/
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi Bjoern, Yes, I am aware of such. I had intended to update the Cluster docs as it's the more appropriate page, perhaps linking to it from the community log page. I was hoping that people might send in their experiences so I could aggregate them and write the entries for them. There's always a lot of apathy towards documentation, and given that this is a feature used by a select few (even if many organisations use it, only a handful of people are actually doing the deployment), I figured that they might be more easily convinced to simply summarise their issues/experiences in a dozen bulletpoints. Additionally by aggregation we could easily identify common problems and highlight those for others to see. That said...no one has replied to this yet. I fully intend to at least include mine, however I was holding out for some other replies. Cheers, Eric On 06/02/2014 08:55 AM, Björn Grüning wrote:
Hi Eric,
thanks for caring about that issue.
Do you know about Galaxy Community Hubs?
https://wiki.galaxyproject.org/Community/
The community is encouraged to share experience under that page. Maybe you can add yours to that page as well?
Cheers, Bjoern
Am 27.05.2014 18:32, schrieb Eric Rasche: This is specifically to galaxy administrators and developers who have worked with galaxy and cluster deployment.
If you have successfully deployed a galaxy instance backed by one (or more) clusters, I'm looking to hear about your experience;
- the configuration - any pitfalls encountered - any extra tools that needed to be installed - any ugly hacks you had to implement
The current cluster documentation is somewhat lacking, and I'm willing to compile everyone's experiences deploying to clusters if you're willing to send them in. I'll update the cluster page on the wiki as I receive replies and can allocate time.
By way of example, here's the summary of what I would reply to this email with my experiences deploying to an HTCondor cluster
- HTcondor ran fine, accepted jobs and worked for all users - Connecting galaxy to it would result in some jobs running/some jobs failing. It was determined that those jobs which ran, were only on localhost. - This was a result of the fact that galaxy's home directory/data was not exported - Galaxy was made an LDAP user and a folder created in the correct location on all computers - This still failed, as galaxy does not transfer files by default (as far as I can tell), even though that's available in HTCondor - Finally galaxy's home directory was NFS exported and things seemed to work.
Cheers, Eric
___________________________________________________________ Please keep all replies on the list by using "reply all" in your mail client. To manage your subscriptions to this and other Galaxy lists, please use the interface at: http://lists.bx.psu.edu/
To search Galaxy mailing lists use the unified search at: http://galaxyproject.org/search/mailinglists/
- -- Eric Rasche Programmer II Center for Phage Technology Texas A&M University College Station, TX 77843 -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.11 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/ iQIcBAEBAgAGBQJTjI4NAAoJEMqDXdrsMcpVk5YP+waHzbx5MS3g31a3l7hC+84M g0jakKQM2t6A/516q3YTm3FouGFpitcvxXYaesDDr/AbiRSWY8KGX6nqn8xk4LNH yxc2Unewfwuzo8hB+PGHUncRnYAIc+BSDr57BDQukY7cuSBV6dymKkipXyim3v3Z buLFKkby17SAvCmBxHN6cfGKnRs6ZLCh0qAyZaTWwmTmrqg3gY3WUh7jBsIS0qUV bMEomgQ4vmqsOoEhi2UsA3ph09y8XcmyQCOkKLcR8nszvRcH/3s89YdMScxs9ghU nGdVaO6B3zXG9QnWc8XgokISCldTcsP5GkKUPcMayz5n9tx2+Mf+BZQdQpGU7qQe P5hGX0sJstklN1TRoJgvR/e6GBaEfT3yPgHlEi+1L7vlrb57vPYXi2Empph/KYo/ fUKGS0q2joxcg3/z2MNzrHc5+y+H9wR0dYVZrUEn3VBY+bhvaIdwpA4BBjjSjf99 BcvUk7FIKi6eyhCFX8RzU8CF4mXNqtrSXzd26igL8datxW0y7bhYcfiied0RMFQv zz6Q4fQyV8hz5ufXKyf7Z1fJ+0PqIiGotN1/wD/l09mMFoKqJhvs4510LhCv7WW+ K7/KBJWJ2O8GrcuW4z7vydF+FxRLpkfGERzKI9IeMmxJiAs8OSYmIROsQzxg7alD MEHRwJf6505DZSTeFDTn =qHdv -----END PGP SIGNATURE-----
Hi,
Yes, I am aware of such. I had intended to update the Cluster docs as it's the more appropriate page, perhaps linking to it from the community log page.
I was hoping that people might send in their experiences so I could aggregate them and write the entries for them. There's always a lot of apathy towards documentation, and given that this is a feature used by a select few (even if many organisations use it, only a handful of people are actually doing the deployment), I figured that they might be more easily convinced to simply summarise their issues/experiences in a dozen bulletpoints. Additionally by aggregation we could easily identify common problems and highlight those for others to see.
That is always good, yes, but hard to motivate people. My problem is that I do not remember every detail of every problem, it's to far away. Galaxy is running for us now. But if I encounter any new problem I will let you and the wiki know.
That said...no one has replied to this yet. I fully intend to at least include mine, however I was holding out for some other replies.
Fingers crossed! Bjoern
Cheers, Eric
On 06/02/2014 08:55 AM, Björn Grüning wrote:
Hi Eric,
thanks for caring about that issue.
Do you know about Galaxy Community Hubs?
https://wiki.galaxyproject.org/Community/
The community is encouraged to share experience under that page. Maybe you can add yours to that page as well?
Cheers, Bjoern
Am 27.05.2014 18:32, schrieb Eric Rasche: This is specifically to galaxy administrators and developers who have worked with galaxy and cluster deployment.
If you have successfully deployed a galaxy instance backed by one (or more) clusters, I'm looking to hear about your experience;
- the configuration - any pitfalls encountered - any extra tools that needed to be installed - any ugly hacks you had to implement
The current cluster documentation is somewhat lacking, and I'm willing to compile everyone's experiences deploying to clusters if you're willing to send them in. I'll update the cluster page on the wiki as I receive replies and can allocate time.
By way of example, here's the summary of what I would reply to this email with my experiences deploying to an HTCondor cluster
- HTcondor ran fine, accepted jobs and worked for all users - Connecting galaxy to it would result in some jobs running/some jobs failing. It was determined that those jobs which ran, were only on localhost. - This was a result of the fact that galaxy's home directory/data was not exported - Galaxy was made an LDAP user and a folder created in the correct location on all computers - This still failed, as galaxy does not transfer files by default (as far as I can tell), even though that's available in HTCondor - Finally galaxy's home directory was NFS exported and things seemed to work.
Cheers, Eric
___________________________________________________________ Please keep all replies on the list by using "reply all" in your mail client. To manage your subscriptions to this and other Galaxy lists, please use the interface at: http://lists.bx.psu.edu/
To search Galaxy mailing lists use the unified search at: http://galaxyproject.org/search/mailinglists/
- -- Eric Rasche Programmer II Center for Phage Technology Texas A&M University College Station, TX 77843 -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.11 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/
iQIcBAEBAgAGBQJTjI4NAAoJEMqDXdrsMcpVk5YP+waHzbx5MS3g31a3l7hC+84M g0jakKQM2t6A/516q3YTm3FouGFpitcvxXYaesDDr/AbiRSWY8KGX6nqn8xk4LNH yxc2Unewfwuzo8hB+PGHUncRnYAIc+BSDr57BDQukY7cuSBV6dymKkipXyim3v3Z buLFKkby17SAvCmBxHN6cfGKnRs6ZLCh0qAyZaTWwmTmrqg3gY3WUh7jBsIS0qUV bMEomgQ4vmqsOoEhi2UsA3ph09y8XcmyQCOkKLcR8nszvRcH/3s89YdMScxs9ghU nGdVaO6B3zXG9QnWc8XgokISCldTcsP5GkKUPcMayz5n9tx2+Mf+BZQdQpGU7qQe P5hGX0sJstklN1TRoJgvR/e6GBaEfT3yPgHlEi+1L7vlrb57vPYXi2Empph/KYo/ fUKGS0q2joxcg3/z2MNzrHc5+y+H9wR0dYVZrUEn3VBY+bhvaIdwpA4BBjjSjf99 BcvUk7FIKi6eyhCFX8RzU8CF4mXNqtrSXzd26igL8datxW0y7bhYcfiied0RMFQv zz6Q4fQyV8hz5ufXKyf7Z1fJ+0PqIiGotN1/wD/l09mMFoKqJhvs4510LhCv7WW+ K7/KBJWJ2O8GrcuW4z7vydF+FxRLpkfGERzKI9IeMmxJiAs8OSYmIROsQzxg7alD MEHRwJf6505DZSTeFDTn =qHdv -----END PGP SIGNATURE-----
participants (2)
-
Björn Grüning -
Eric Rasche