HSM input/output error

[jaberle]

I have a very intermintent issue where when recalling some files from hsm, immediately gets an input or output error from the OS when trying to copy the file elsewhere. No errors are entered on the console, TSM console, or dsmerror.log. I can successfully, restore the file with dsmc then migrate and recall that file. It seems like an issue with the stub file not knowing where to go get the file - possibly corrupt stub file. This only happens on some files, not relative to size.



We are currently using TSM 4.2.2 on the client. We had this problem in 4.1 as well.



Has anyone seen a similar problem? thanks for the help,

 

[nyquesty]

Late replay, but here I go anyway.



I am having the same issue. When I do large dsmrecalls, the whole HSM FS will lockup. I was thinking this has something to do with my GPFS (which is my HSM FS). If I kill and restart GPFS, all is well. By any chance are you using GPFS?



I am TSM 5.2, HSM 5.2, GPFS 1.2. I recall 20 20GB files at a time. Maybe that is just too much?

 

[jaberle]

Not using GPFS, however by going to 5.1.6.15 I believe my issues were resolved with the following PTF:



IC37278 TSM HSM CLIENT MAY HANG DURING RECALL OF FILE.

 

[mtr]

Hallo,



JFS2 is supported at the end of 2004!!

 

[pns]

<TABLE BORDER=0 ALIGN=CENTER WIDTH=85%><TR><TD><font class="pn-sub">Quote:</font><HR></TD></TR><TR><TD><FONT class="pn-sub"><BLOCKQUOTE>Not using GPFS, however by going to 5.1.6.15 I believe my issues were resolved with the following PTF:



IC37278 TSM HSM CLIENT MAY HANG DURING RECALL OF FILE.</BLOCKQUOTE></FONT></TD></TR><TR><TD><HR></TD></TR></TABLE>



We have always experienced HSM lockups while recovering large (up to 8GB) files.

We are using VxFS on Solaris. Currently our HSM client is 5.1.5.6 - I would be interested if there had been any further feedback on whether this PTF fixes the problem - a previous poster on this thread mentioned problems when recalling 20GB files, but it looked as if he was already on HSM 5.2



We have also had many problems with slow access to files once they had been migrated and then recalled - after much woe and lost hair we've given up trying to use HSM in the way we originally intended, and have to copy recalled files elsewhere before using them - this fixes the speed problem! IBM did look into it, but beyond being able to replicate the problem, they had no solution. That was a year or so ago - hence my interest to see if there has been any recent progress.



Phil