This randomized phase II trial is studying the side effects and how well giving bevacizumab together with irinotecan or temozolomide works in treating patients with recurrent or refractory glioblastoma multiforme or gliosarcoma. Monoclonal antibodies, such as bevacizumab, can block tumor growth in different ways. Some block the ability of tumor cells to grow and spread. Others find tumor cells and help kill them or carry tumor-killing substances to them. Bevacizumab may also stop the growth of tumor cells by blocking blood flow to the tumor. Drugs used in chemotherapy, such as irinotecan and temozolomide, work in different ways to stop the growth of tumor cells, either by killing the cells or by stopping them from dividing. Giving bevacizumab together with irinotecan or temozolomide may kill more tumor cells.

PRIMARY OBJECTIVES: I. Determine the efficacy of bevacizumab and irinotecan hydrochloride, in terms of 6-month progression-free survival rate, in patients with recurrent or refractory intracranial glioblastoma multiforme or gliosarcoma. II. Determine the adverse event profile and tolerability of bevacizumab and temozolomide in these patients. SECONDARY OBJECTIVES: I. Determine the efficacy of bevacizumab and temozolomide, in terms of 6-month progression-free survival rate, in patients previously treated with temozolomide. II. Determine the efficacy of bevacizumab and irinotecan hydrochloride, in terms of objective response, in patients with measurable disease. III. Determine the efficacy of bevacizumab and temozolomide, in terms of objective response, in patients with measurable disease who were previously treated with temozolomide. IV. Determine the toxicity profile and tolerability of bevacizumab and irinotecan hydrochloride in these patients. TERTIARY OBJECTIVES: I. Assess the potential role of perfusion MRI and magnetic resonance spectroscopy imaging as an early indicator of response to therapy after 2 weeks of treatment with bevacizumab. II. Assess the potential role of perfusion MRI and magnetic resonance spectroscopy imaging as a prognostic indicator based on images taken at baseline, at 2 weeks, and after 2 courses of study treatment. OUTLINE: This is a randomized, multicenter study. Patients are stratified according to age (\< 50 vs \>= 50 years of age) and Karnofsky performance status (70-80% vs 90-100%). Patients are randomized to 1 of 2 treatment arms with a 2:1 ratio (arm I:arm II). ARM I: Patients receive bevacizumab IV over 30-90 minutes on days 1 and 15 and oral temozolomide once daily on days 1-21. ARM II: Patients receive bevacizumab IV as in Arm I followed by irinotecan hydrochloride IV over 90 minutes on days 1 and 15. In both arms, treatment repeats every 28 days for up to 24 courses in the absence of disease progression or unacceptable toxicity. All patients undergo MRI at baseline and at every 2 courses (no 2-week MRI) per standard of care until progression or discontinuation of treatment to assess areas of breakdown of the blood-brain barrier. Patients undergo an additional MRI after study therapy. Consenting patients also undergo diffusion and perfusion MRI and magnetic resonance spectroscopic imaging for correlative studies. After completion of study therapy, patients are followed up for at least 1 month.